threads(M6): getCurrentId, docs, and CI wiring — threading built

New thread_self=42 syscall backs runtime.Thread.getCurrentId (the calling
thread's kernel task id). thread-test gains an id mode: two workers read
getCurrentId and the main thread confirms all three ids are non-zero and
distinct. Per-thread threadlocal TLS is deferred by design (no consumer; it would
need context-switched fs.base for an unused feature), as are RwLock/WaitGroup.

Marks the threading feature built (M1-M6): threading.md + docs/README.md status
updated, all thread-* cases wired into qemu_test.py.

Gate thread-id PASS; full suite green: 21/21 (thread-spawn/join/futex/mutex/id,
aspace-refcount, + 15 guardrail cases), zig build clean, zig build test green.

.zig-version

@@ -1 +0,0 @@
-0.16.0

README.md

@@ -1,8 +1,8 @@
 # DanOS
-Codename: Shodan
-Version: 1
 
-A small resilient operating system, written from scratch in Zig.
+**Codename: Shodan**
+
+A very small resilient operating system.
 
 ## Zen of DanOS:
 
@@ -18,12 +18,12 @@ A small resilient operating system, written from scratch in Zig.
     - Useful during driver development.
     - Drivers can claim MMIO / ports
     - Driver resources (e.g. IRQ/Port/MMIO) claims are automatically cleaned up if the driver dies or is killed
-    - Drivers can also hook into the process lifecyle to clean up or reset hardware
+    - Drivers can also hook into the process lifecycle to clean up or reset hardware
 - No legacy to deal with
     - Zig code uses a clean coding style (Zen of Zig)
     - Favor reading code over writing code.
     - No magic numbers.
-    - No shortend names unless its for ABI compatibility or acronyms
+    - No shortened names unless its for ABI compatibility or acronyms
 - Inter-Process Communication (IPC)
   - Publish and subscribe to Asynchronous Messages
   - Talk to services and processes synchronously
@@ -54,6 +54,17 @@ the UEFI bootloader at `zig-out/EFI/BOOT/BOOTX64.efi`, the kernel at
 `zig-out/system/kernel`, init at `zig-out/system/services/init`, drivers under
 `zig-out/system/drivers/`, and the initial-ramdisk at `zig-out/boot/`.
 
+## Release media
+
+```sh
+zig build release-x86-64
+```
+
+Produces `zig-out/danos-x86-64.iso`, a hybrid ISO that boots flashed raw to a
+USB stick (balenaEtcher, dd) or burned to optical media — see
+[docs/release-iso.md](docs/release-iso.md). `zig build check-iso-image`
+validates it without booting.
+
 ## Run
 
 Boot it in QEMU with OVMF (opens a display window):

build.zig

@@ -54,6 +54,12 @@ fn timestamp(b: *std.Build) []const u8 {
 /// can't reach), linked against the `runtime` runtime library with the shared user
 /// link script. Pinned to LLVM + LLD so the script's PHDRS (segment permissions)
 /// are authoritative — the kernel's W^X user-ELF loader requires exact perms.
+///
+/// The compilation root is not the program's own file but the shared shim
+/// library/runtime/root.zig, which supplies the root declarations (`main`
+/// re-export, panic handler, `_start` pull) so a program only defines
+/// `pub fn main`. The program's file becomes the `program` module the shim
+/// imports; reach it through `programModule` to add per-binary imports.
 fn addUserBinary(
     b: *std.Build,
     target: std.Build.ResolvedTarget,
@@ -64,17 +70,40 @@ fn addUserBinary(
     name: []const u8,
     root: []const u8,
 ) *std.Build.Step.Compile {
-    const exe = b.addExecutable(.{
+    return addUserBinaryImpl(b, target, runtime_module, mmio_module, xkeyboard_config_module, acpi_ids_module, name, root, false);
-        .name = name,
+}
-        .root_module = b.createModule(.{
+
+/// As `addUserBinary`, but built multi-threaded (`single_threaded = false`) so real
+/// atomics/TLS work — required before a binary may call `runtime.Thread.spawn`
+/// (docs/threading.md). Threads are a deliberate per-binary opt-in.
+fn addThreadedUserBinary(
+    b: *std.Build,
+    target: std.Build.ResolvedTarget,
+    runtime_module: *std.Build.Module,
+    mmio_module: *std.Build.Module,
+    xkeyboard_config_module: *std.Build.Module,
+    acpi_ids_module: *std.Build.Module,
+    name: []const u8,
+    root: []const u8,
+) *std.Build.Step.Compile {
+    return addUserBinaryImpl(b, target, runtime_module, mmio_module, xkeyboard_config_module, acpi_ids_module, name, root, true);
+}
+
+fn addUserBinaryImpl(
+    b: *std.Build,
+    target: std.Build.ResolvedTarget,
+    runtime_module: *std.Build.Module,
+    mmio_module: *std.Build.Module,
+    xkeyboard_config_module: *std.Build.Module,
+    acpi_ids_module: *std.Build.Module,
+    name: []const u8,
+    root: []const u8,
+    threaded: bool,
+) *std.Build.Step.Compile {
+    // Settings (target, optimize, code model, ...) live on the root module only;
+    // the program and runtime modules leave theirs null and inherit them.
+    const program_module = b.createModule(.{
         .root_source_file = b.path(root),
-            .target = target,
-            .optimize = .ReleaseSmall,
-            .code_model = .large,
-            .single_threaded = true,
-            .sanitize_c = .off,
-            .stack_check = false,
-            .stack_protector = false,
         .imports = &.{
             .{ .name = "runtime", .module = runtime_module },
             // Typed volatile MMIO + memory barriers, for drivers. See library/mmio/.
@@ -86,6 +115,22 @@ fn addUserBinary(
             // (HardwareId.ps2_keyboard) instead of magic "_HID" strings.
             .{ .name = "acpi-ids", .module = acpi_ids_module },
         },
+    });
+    const exe = b.addExecutable(.{
+        .name = name,
+        .root_module = b.createModule(.{
+            .root_source_file = b.path("library/runtime/root.zig"),
+            .target = target,
+            .optimize = .ReleaseSmall,
+            .code_model = .large,
+            .single_threaded = !threaded, // a threaded binary needs real atomics/TLS
+            .sanitize_c = .off,
+            .stack_check = false,
+            .stack_protector = false,
+            .imports = &.{
+                .{ .name = "runtime", .module = runtime_module },
+                .{ .name = "program", .module = program_module },
+            },
         }),
     });
     exe.setLinkerScript(b.path("library/runtime/user.ld"));
@@ -96,6 +141,14 @@ fn addUserBinary(
     return exe;
 }
 
+/// The `program` module of a binary built by `addUserBinary` — the module rooted
+/// at the program's own source file. Per-binary imports (protocol modules, bus
+/// ABIs) go here, not on the root shim: module imports are not transitive, so an
+/// import added to the root would be invisible to the program's code.
+fn programModule(exe: *std.Build.Step.Compile) *std.Build.Module {
+    return exe.root_module.import_table.get("program").?;
+}
+
 /// The modules the kernel imports, gathered once so both kernel variants (the
 /// installed one and the serial-enabled one `run-x86-64` boots) are built from
 /// the same set. `build_options` is *not* here — it carries `serial`/`test_case`,
@@ -171,8 +224,9 @@ fn addKernel(
 /// Assemble the bootable FAT32 image (the in-repo Python builder) holding what
 /// the firmware and loader need off the ESP: the EFI stub, `kernel`, `init`, and
 /// the initial-ramdisk. Factored so the serial-enabled `run-x86-64` variant can
-/// bundle its own kernel while sharing the (serial-independent) loader, init, and
+/// bundle its own serial kernel while sharing the loader, init, and ramdisk — all
-/// ramdisk. Returns the image's LazyPath.
+/// built once per invocation (the loader's boot breadcrumbs and init's heartbeat
+/// both follow the top-level -Dserial). Returns the image's LazyPath.
 fn addBootImage(
     b: *std.Build,
     kernel_bin: std.Build.LazyPath,
@@ -440,6 +494,14 @@ pub fn build(b: *std.Build) void {
     // linked into the kernel's user region against the `runtime` runtime library, and
     // started in ring 3 by the kernel's user-ELF loader.
     const init_exe = addUserBinary(b, kernel_target, runtime_module, mmio_module, xkeyboard_config_module, acpi_ids_module, "init", "system/services/init/init.zig");
+    // init reads the same `serial` flag the kernel does: its liveness heartbeat is a
+    // serial/test-build diagnostic (the QEMU harness's init tests assert on it, and
+    // -Dserial images emit it), so a flashable image runs a purely event-driven PID 1
+    // that wakes only for real work. The test harness builds with -Dserial=true, so
+    // the heartbeat stays present under test.
+    const init_options = b.addOptions();
+    init_options.addOption(bool, "serial", serial);
+    programModule(init_exe).addImport("build_options", init_options.createModule());
     const init_install = b.addInstallArtifact(init_exe, .{ .dest_dir = .{ .override = .{ .custom = "system/services" } } });
     b.getInstallStep().dependOn(&init_install.step);
 
@@ -456,20 +518,20 @@ pub fn build(b: *std.Build) void {
     // The xHCI bus driver builds chapter-9 requests and decodes descriptors from
     // usb-abi, and reports each interface's (class,subclass,protocol) identity via
     // usb-ids.packTriple.
-    usb_xhci_bus_exe.root_module.addImport("usb-abi", usb_abi_module);
+    programModule(usb_xhci_bus_exe).addImport("usb-abi", usb_abi_module);
-    usb_xhci_bus_exe.root_module.addImport("usb-ids", usb_ids_module);
+    programModule(usb_xhci_bus_exe).addImport("usb-ids", usb_ids_module);
-    usb_xhci_bus_exe.root_module.addImport("usb-transfer-protocol", usb_transfer_protocol_module);
+    programModule(usb_xhci_bus_exe).addImport("usb-transfer-protocol", usb_transfer_protocol_module);
     // The USB HID class drivers: keyboard and mouse. They own no hardware — each
     // opens its device through runtime.usb (the transfer protocol) and publishes to
     // the input service. They build chapter-9 class requests from usb-abi.
     const usb_hid_keyboard_exe = addUserBinary(b, kernel_target, runtime_module, mmio_module, xkeyboard_config_module, acpi_ids_module, "usb-hid-keyboard", "system/drivers/usb-hid/keyboard.zig");
-    usb_hid_keyboard_exe.root_module.addImport("usb-abi", usb_abi_module);
+    programModule(usb_hid_keyboard_exe).addImport("usb-abi", usb_abi_module);
     const usb_hid_mouse_exe = addUserBinary(b, kernel_target, runtime_module, mmio_module, xkeyboard_config_module, acpi_ids_module, "usb-hid-mouse", "system/drivers/usb-hid/mouse.zig");
-    usb_hid_mouse_exe.root_module.addImport("usb-abi", usb_abi_module);
+    programModule(usb_hid_mouse_exe).addImport("usb-abi", usb_abi_module);
     // The USB mass-storage class driver: opens its device via runtime.usb, drives it
     // with Bulk-Only Transport + SCSI, and serves the block protocol under `.block`.
     const usb_storage_exe = addUserBinary(b, kernel_target, runtime_module, mmio_module, xkeyboard_config_module, acpi_ids_module, "usb-storage", "system/drivers/usb-storage/usb-storage.zig");
-    usb_storage_exe.root_module.addImport("block-protocol", block_protocol_module);
+    programModule(usb_storage_exe).addImport("block-protocol", block_protocol_module);
     // The FAT filesystem server: mounts the block device and serves it into the VFS
     // at /mnt/usb. Its engine (engine.zig / on-disk.zig) is imported relatively.
     const fat_exe = addUserBinary(b, kernel_target, runtime_module, mmio_module, xkeyboard_config_module, acpi_ids_module, "fat", "system/services/fat/fat.zig");
@@ -482,7 +544,7 @@ pub fn build(b: *std.Build) void {
     const pci_bus_exe = addUserBinary(b, kernel_target, runtime_module, mmio_module, xkeyboard_config_module, acpi_ids_module, "pci-bus", "system/drivers/pci-bus/pci-bus.zig");
     // The PCI bus driver decodes each function's class triple to human names in its
     // boot log (class/subclass/prog-IF), so pull in the shared pci-class reference.
-    pci_bus_exe.root_module.addImport("pci-class", pci_class_module);
+    programModule(pci_bus_exe).addImport("pci-class", pci_class_module);
     // A test fixture, not a real driver: hellos to the device manager, then faults —
     // what the driver-restart scenario drives the crash-loop cap with.
     const crash_test_exe = addUserBinary(b, kernel_target, runtime_module, mmio_module, xkeyboard_config_module, acpi_ids_module, "crash-test", "system/services/crash-test/crash-test.zig");
@@ -501,13 +563,13 @@ pub fn build(b: *std.Build) void {
         .fdt => "system/services/fdt/fdt.zig",
     };
     const discovery_exe = addUserBinary(b, kernel_target, runtime_module, mmio_module, xkeyboard_config_module, acpi_ids_module, "discovery", discovery_source);
-    if (discovery == .acpi) discovery_exe.root_module.addImport("aml", aml_module);
+    if (discovery == .acpi) programModule(discovery_exe).addImport("aml", aml_module);
     const device_manager_exe = addUserBinary(b, kernel_target, runtime_module, mmio_module, xkeyboard_config_module, acpi_ids_module, "device-manager", "system/services/device-manager/device-manager.zig");
     // Names the xHCI PCI class triple from the shared taxonomy instead of a bare 0x0C0330.
-    device_manager_exe.root_module.addImport("pci-class", pci_class_module);
+    programModule(device_manager_exe).addImport("pci-class", pci_class_module);
     // The manager matches reported USB interfaces by their (class,subclass,protocol)
     // triple (usbDriverForIdentity), built from the named usb-ids codes.
-    device_manager_exe.root_module.addImport("usb-ids", usb_ids_module);
+    programModule(device_manager_exe).addImport("usb-ids", usb_ids_module);
     // The input service and its exercisers: the fan-out server, a hardware-free synthetic
     // source, and a subscriber that doubles as the `input` test's oracle. See docs/input.md.
     const input_exe = addUserBinary(b, kernel_target, runtime_module, mmio_module, xkeyboard_config_module, acpi_ids_module, "input", "system/services/input/input.zig");
@@ -516,6 +578,9 @@ pub fn build(b: *std.Build) void {
     const args_echo_exe = addUserBinary(b, kernel_target, runtime_module, mmio_module, xkeyboard_config_module, acpi_ids_module, "args-echo", "system/services/args-echo/args-echo.zig");
     const process_test_exe = addUserBinary(b, kernel_target, runtime_module, mmio_module, xkeyboard_config_module, acpi_ids_module, "process-test", "system/services/process-test/process-test.zig");
     const log_flush_exe = addUserBinary(b, kernel_target, runtime_module, mmio_module, xkeyboard_config_module, acpi_ids_module, "log-flush", "system/services/log-flush/log-flush.zig");
+    // The first multi-threaded binary: exercises runtime.Thread over the thread ABI
+    // (docs/threading.md). Built threaded so its shared-memory poll is real.
+    const thread_test_exe = addThreadedUserBinary(b, kernel_target, runtime_module, mmio_module, xkeyboard_config_module, acpi_ids_module, "thread-test", "system/services/thread-test/thread-test.zig");
 
     // Pack the user binaries into the initial_ramdisk image with the host-side Python tool
     // (the container format is trivial, and Python sidesteps std API churn). Args:
@@ -559,6 +624,8 @@ pub fn build(b: *std.Build) void {
     mk_run.addFileArg(pci_bus_exe.getEmittedBin());
     mk_run.addArg("crash-test");
     mk_run.addFileArg(crash_test_exe.getEmittedBin());
+    mk_run.addArg("thread-test");
+    mk_run.addFileArg(thread_test_exe.getEmittedBin());
     mk_run.addArg("device-list");
     mk_run.addFileArg(device_list_exe.getEmittedBin());
     mk_run.addArg("discovery");
@@ -661,6 +728,31 @@ pub fn build(b: *std.Build) void {
     const check_fat_step = b.step("check-fat-image", "Verify the FAT32 USB image is valid and bootable");
     check_fat_step.dependOn(&check_fat.step);
 
+    // --- release-x86-64: danos-x86-64.iso, the flashable release image ---
+    // Wrap the FAT32 boot volume in a hybrid ISO (the in-repo Python builder
+    // again, no xorriso/isohybrid): an ISO9660 whose El Torito EFI boot entry
+    // and MBR ESP partition entry both point at the embedded FAT image. One
+    // file then boots every way release media is consumed — flashed raw to a
+    // USB stick with Etcher or dd, or burned to optical media — while
+    // danos-usb.img stays the raw superfloppy QEMU and the test harness boot.
+    const mk_iso = b.addSystemCommand(&.{"python3"});
+    mk_iso.addFileArg(b.path("tools/make-iso-image.py"));
+    const iso_image = mk_iso.addOutputFileArg("danos-x86-64.iso");
+    mk_iso.addFileArg(fat_image);
+    const iso_install = b.addInstallFile(iso_image, "danos-x86-64.iso");
+    const release_step = b.step("release-x86-64", "Build the flashable x86-64 release ISO (zig-out/danos-x86-64.iso; flash with Etcher or dd)");
+    release_step.dependOn(&iso_install.step);
+
+    // `zig build check-iso-image` — the ISO builder's own --verify (mirroring
+    // check-fat-image): the MBR partition, the El Torito catalog, and the
+    // embedded FAT32 image must all agree.
+    const check_iso = b.addSystemCommand(&.{"python3"});
+    check_iso.addFileArg(b.path("tools/make-iso-image.py"));
+    check_iso.addArg("--verify");
+    check_iso.addFileArg(iso_image);
+    const check_iso_step = b.step("check-iso-image", "Verify the release ISO is a valid hybrid (MBR ESP partition + El Torito EFI entry)");
+    check_iso_step.dependOn(&check_iso.step);
+
     // --- run-x86-64: boot the x86-64 kernel in QEMU via UEFI/OVMF ---
     // Firmware lives in different places per OS/distro, so probe the known
     // layouts (Architecture, Debian/Ubuntu, Fedora, macOS Homebrew) and use the first

docs/README.md

@@ -39,37 +39,42 @@ rather than restate it. Roughly in the order things happen at runtime:
     endpoints — the backbone the microkernel's isolated servers talk over.
 12. **[syscall.md](syscall.md) — system calls.** How ring 3 asks the kernel for
     something: the `syscall`/`sysret` fast path, the trap frame, and why the table is
-    deliberately tiny.
+    deliberately tiny. The numbers are a **private** ABI — [vdso.md](vdso.md) designs
-13. **[drivers.md](drivers.md) — writing a driver.** The payoff: a driver is an
+    the public boundary that will hide them.
+13. **[vfs-protocol.md](vfs-protocol.md) — the VFS wire protocol.** The language-neutral
+    byte-level spec of the file protocol spoken over IPC: request/reply headers,
+    the operation table, mount routing, and the append-only evolution rules — the
+    first IPC protocol documented as public ABI.
+14. **[drivers.md](drivers.md) — writing a driver.** The payoff: a driver is an
     ordinary ring-3 process that claims a device, maps its registers, and **sleeps
     until its hardware interrupts it**. The claim is the capability; `irq_ack` is the
     unmask.
-14. **[driver-model.md](driver-model.md) — buses, classes and host controllers.** How
+15. **[driver-model.md](driver-model.md) — buses, classes and host controllers.** How
     real driver stacks factor into three shapes and how families share code. The
     three primitives it proposed are long since built (M13 capability passing,
     M14 DMA + barriers, M15 MSI), and the driver *contract* on top of them —
     hello, supervision, restart — is built too (device-manager.md, M18).
-15. **[process-management.md](process-management.md) — process management.** The
+16. **[process-management.md](process-management.md) — process management.** The
     microkernel's `ps`/`kill`/SIGCHLD: enumerate as a table snapshot, the
     supervision link as the kill authority, and child-exit notifications over the
     same endpoints IRQs arrive on.
-16. **[process-lifecycle.md](process-lifecycle.md) — the process lifecycle.** Built
+17. **[process-lifecycle.md](process-lifecycle.md) — the process lifecycle.** Built
     (M17): signals over IPC as the one lifecycle vocabulary every process speaks — the
     POSIX.1-1990 words with message delivery instead of stack hijack, the stable
     `runtime.process` interface, exit reasons, published exit events any stateful
     service can subscribe to (the VFS releasing dead clients' handles), and the two
     iron rules (cleanup is the kernel's job; kill is not a signal).
-17. **[device-manager.md](device-manager.md) — the device manager.** Built (M18,
+18. **[device-manager.md](device-manager.md) — the device manager.** Built (M18,
     through the app surface): the
     tree, the matcher, and the supervisor. Tree structure lives in the manager,
     authority stays in the kernel; bus drivers report what they see; drivers are
     restarted through the lifecycle vocabulary — the plan that turns
     [resilience.md](resilience.md)'s restart goal into increments.
-18. **[input.md](input.md) — the input module.** Broadcasting input events (keyboard,
+19. **[input.md](input.md) — the input module.** Broadcasting input events (keyboard,
     mouse, joystick): why a synchronous rendezvous can't fan out to many listeners, the
     asynchronous `ipc_send` primitive built to fix it, and the per-device subscribe/publish
     service layered on top.
-19. **[display.md](display.md) — the display service.** The display half of the GUI
+20. **[display.md](display.md) — the display service.** The display half of the GUI
     track: a user-space compositor that owns the framebuffer, composes a layer stack into
     a double buffer, and presents it. Why GOP and the PCI display device are two views of
     one controller, the device-node + write-combining handoff, and what flicker-free buys
@@ -80,7 +85,7 @@ rather than restate it. Roughly in the order things happen at runtime:
     further out, two research snapshots survey what a *native* driver for real GPU silicon
     would take as another `.scanout` backend: [nvidia-gpus.md](nvidia-gpus.md) (RTX 3060 /
     Ampere) and [intel-igpu.md](intel-igpu.md) (Intel iGPU).
-20. **[halting.md](halting.md) — halting.** Why a kernel can't just "exit", and
+21. **[halting.md](halting.md) — halting.** Why a kernel can't just "exit", and
     how `while (true) hlt` parks the CPU safely once there's nothing left to do.
 
 Start with the north star:
@@ -99,6 +104,19 @@ Start with the north star:
   port to **one seam** (`std.os.danos`), so we build `runtime.os` (→ that seam) plus a
   thin `runtime.fs`, retire the `posix` shim, and follow a phased path to
   `zig build-exe hello.zig` running on danos — **not** Linux-ABI emulation.
+- **[threading.md](threading.md) — threads, the std-shaped way.** **Built** (M1–M6):
+  `runtime.Thread` mirrors `std.Thread`'s API (spawn/join/detach, Mutex/Condition/
+  Semaphore) over a **private** thread ABI — several tasks sharing one address space via
+  a `thread_spawn` syscall, futex-backed blocking, aspace refcounting. Why it's the
+  native type and not literal `std.Thread` (the [private ABI](syscall.md)), and why
+  threads stay a narrow opt-in against the [resilience](resilience.md) default. Build
+  plan + gates: [threading-plan.md](threading-plan.md).
+- **[vdso.md](vdso.md) — the vDSO, the public system-call boundary.** A design note
+  (not built yet) on keeping `abi.zig` genuinely private: a kernel-supplied, C-ABI
+  entry blob mapped into every process as the *only* way into the kernel — so the
+  syscall numbers can be renumbered or randomised at will, and Rust/C binaries get a
+  stable boundary without danos growing a dynamic linker. danos's public ABI = the
+  vDSO + the documented IPC wire protocols ([vfs-protocol.md](vfs-protocol.md) first).
 
 Cutting across all of these:
 
@@ -106,6 +124,11 @@ Cutting across all of these:
   hardware needed to run danos: minimum specs (UEFI x86-64, ACPI, PCIe ECAM,
   xHCI, ~128 MiB RAM) grounded in what the boot path actually assumes, plus a
   plain-language guide matching Intel/AMD CPU generations by name.
+- **[release-iso.md](release-iso.md) — the release ISO.** The flashable boot
+  media: `zig build release-x86-64` wraps the FAT32 boot volume in a hybrid ISO
+  (MBR ESP partition + El Torito EFI entry, one embedded image) that Etcher/dd
+  flash to USB or a burner writes to disc — built by an in-repo pure-Python
+  tool, like the FAT image itself.
 - **[arch.md](arch.md) — the architecture split.** How CPU-specific code is kept
   behind a build-time `arch` module so the generic kernel never names x86_64,
   leaving room for other systems (e.g. an AArch64 Raspberry Pi) later.
@@ -250,5 +273,5 @@ exception in [coding-standards.md](coding-standards.md) applies to that seam.
 | danos-native runtime (`runtime`): syscall wrappers, heap, IPC, device access, the file API (`fs`) — the stable application ABI | `library/runtime/` |
 | System services (init, the VFS server + `protocol`, the device-manager) | `system/services/` |
 | Device drivers, one sub-project each (`pci-bus`, `ps2-bus`, `usb-xhci-bus` bus drivers) | `system/drivers/` |
-| Build + `run-x86-64` (QEMU/OVMF) | `build.zig` |
+| Build + `run-x86-64` (QEMU/OVMF) + `release-x86-64` (the flashable ISO) | `build.zig` |
 | QEMU integration test harness | `test/qemu_test.py` |

docs/os-developer-guide/README.MD

@@ -0,0 +1,9 @@
+# OS Developer Guide
+
+This document is for those who need to understand the architectural decisions behind the OS.
+
+## Written in Zig?
+
+The os was initially written in zig because it has excellent support for EFI. With zig, we could forgo using a third party bootloader, reducing the time to boot up the kernel. Following the "Zen of Zig", helped to produce the most readable codebase for an operating system ever created. So those, new to OS development could quickly get up to speed.
+
+

docs/release-iso.md

@@ -0,0 +1,70 @@
+# The release ISO — flashable boot media
+
+`zig build release-x86-64` produces **`zig-out/danos-x86-64.iso`**, the file you
+hand to someone who wants to try danos on a real machine: point
+[balenaEtcher](https://etcher.balena.io) (or Raspberry Pi Imager, or plain `dd`)
+at it, flash a USB stick, and boot the stick. The same file also burns to
+optical media. `zig build check-iso-image` validates it without booting.
+
+```
+zig build release-x86-64
+# Etcher: select danos-x86-64.iso → select the stick → Flash
+# or: sudo dd if=zig-out/danos-x86-64.iso of=/dev/rdiskN bs=4m   (macOS; triple-check N)
+```
+
+## Why an ISO when danos-usb.img already boots
+
+`danos-usb.img` is a raw FAT32 **superfloppy** — a filesystem starting at
+sector 0, no partition table. UEFI firmware accepts that from a USB stick (it
+probes whole-disk FAT before giving up), which is why `dd`-ing the .img works
+and why QEMU and the test harness boot it directly. But it is a
+developer-shaped artifact: flashing apps expect an ISO, and a superfloppy
+can't be burned to a CD/DVD or carry a partition table for pickier firmware.
+
+The ISO wraps that same FAT image — bit-identical, built by the same
+`tools/make-fat-image.py` — in a container that boots everywhere release media
+gets consumed. One payload, two images: the .img stays the raw volume the QEMU
+harness mounts and boots, the .iso is what leaves the building.
+
+## How a hybrid ISO boots twice
+
+The trick (the same one Linux distribution ISOs use, usually via `xorriso
+-isohybrid…`) is that ISO9660 reserves its first 32 KiB as a **system area** it
+never touches — exactly where an MBR lives on a disk. So one file can carry two
+tables of contents, both pointing at the same embedded FAT image:
+
+* **Flashed to USB (Etcher, dd):** firmware sees a disk whose sector 0 is an
+  MBR with one partition of type `0xEF` (EFI System Partition) covering the
+  embedded FAT image. It mounts that ESP and runs `\EFI\BOOT\BOOTX64.efi` —
+  the standard removable-media path ([efi.md](efi.md)).
+* **Burned to optical media:** firmware reads the ISO9660 volume descriptors
+  at sector 16 and finds an **El Torito** boot record. Its catalog has one
+  entry, platform ID `0xEF` (EFI), whose start LBA is — again — the embedded
+  FAT image. The firmware exposes that image as a virtual disk and runs the
+  same `BOOTX64.efi` off it.
+
+Neither path involves the legacy BIOS boot-sector machinery: danos is
+UEFI-only ([system-requirements.md](system-requirements.md)), so the MBR holds
+no boot code, just the partition entry, and the El Torito entry is EFI-class,
+not floppy emulation.
+
+One El Torito wrinkle: the catalog's sector-count field is 16-bit (units of
+512 bytes), so it can name at most 32 MiB — less than the 64 MiB FAT image.
+That is fine in practice: firmware sizes the FAT filesystem from its own BPB,
+and the boot files sit in the first few MiB of the image (clusters are
+allocated from the front) either way. The USB path has no such cap.
+
+## The builder
+
+`tools/make-iso-image.py` follows the house rule of
+[make-fat-image.py](../tools/make-fat-image.py): pure Python 3 standard
+library, no external tools (no xorriso, mkisofs, or isohybrid), with a
+`--verify` mode the `check-iso-image` step runs — it checks that the MBR
+partition and the El Torito catalog agree on where the FAT image lives and
+that a FAT32 boot sector is actually there. Every timestamp field in the ISO
+is zeroed, so the build is reproducible byte-for-byte.
+
+The ISO9660 filesystem around the boot machinery is minimal but real: a root
+directory listing `BOOT.CAT` (the catalog) and `EFI.IMG` (the FAT image), so
+`file`, mount tools, and archive browsers can open the ISO and see what's in
+it.

docs/threading-plan.md

@@ -0,0 +1,290 @@
+# Threading — build plan (`runtime.Thread` over a private thread ABI)
+
+The ordered, checkpointable build-out for [threading.md](threading.md). Each milestone
+lands on its own and ends in a **verifiable gate** — shaped for a `/loop` run, like
+[display-v2-plan.md](display-v2-plan.md). Read threading.md first for the *why*.
+
+## Locked decisions (do not relitigate)
+
+- **`runtime.Thread` mirrors `std.Thread`'s API; the implementation is danos-native.**
+  Not literal `std.Thread` — that would break the [private ABI](syscall.md).
+- **Threads are a narrow, per-binary opt-in.** Default concurrency stays process + IPC
+  ([resilience.md](resilience.md)); only a service that asks is built
+  `single_threaded = false`.
+- **Blocking is futex-backed, never spin-backed** — waiters park in the kernel so an
+  idle core still halts ([halting.md](halting.md)).
+- **New syscalls are private**: extend [abi.zig](../system/abi.zig) `SystemCall` after
+  `shm_physical = 36` (`thread_spawn = 37`, `thread_exit = 38`, `current_core = 39`,
+  `futex_wait = 40`, `futex_wake = 41`) + a `library/runtime` wrapper; user code never names a number.
+- **Restart granularity stays the process** — a faulting thread kills its process; the
+  supervisor restarts the process, which respawns its threads.
+
+## Conventions
+
+Follow [coding-standards.md](coding-standards.md): spell out non-acronym abbreviations,
+kebab-case file names, no `Co-Authored-By` trailers. New user binaries go through
+`addUserBinary` (with the new `threaded` flag where a binary spawns threads) and get
+packed into the initial-ramdisk; new syscalls extend [abi.zig](../system/abi.zig)
+`SystemCall` + a `library/runtime` wrapper; test services live beside the code they
+exercise and register a `ServiceId` if they must be looked up.
+
+## How to verify along the way
+
+**Every gate is serial-checkable — no screenshots** (this plan runs unattended). A
+thread proves it ran by writing to **shared memory** the parent reads back, and proves
+parallelism by stamping the **core index** it ran on (like the `smp`/`affinity` cases).
+
+- `zig build test` — host unit tests (closure packing, mutex state machine, futex
+  wrapper encodings).
+- `python3 test/qemu_test.py <case>` — boots the kernel in QEMU; asserts on serial
+  markers. Thread cases set `smp: true` (real parallelism) and bump `mem` (they boot
+  the process/scheduler stack); each milestone **adds its case to `CASES`** so its gate
+  is runnable.
+- **Guardrail every milestone:** the concurrency-sensitive existing cases stay green —
+  `smoke`, `sched`, `priority`, `smp`, `affinity`, `process`, `process-kill`,
+  `supervision`, `fault-recovery`, `vfs-client-death`, `ipc`/`ipc-cap`,
+  `display-service`. A threading change that regresses those is rejected.
+
+## Unattended execution (the loop contract)
+
+This plan runs to completion **without human input**. Every design choice is already
+fixed in *Locked decisions*; the checkboxes are the only state. A loop iteration must:
+
+1. **Resume** at the first milestone that still has an unchecked `- [ ]`. (All earlier
+   milestones are done — do not revisit them.)
+2. **Work on a branch.** On the first iteration, branch off `main` (e.g. `threading`);
+   never commit threading work to `main`. All work stays local — **do not push**.
+3. **Implement** every unchecked item in that milestone, including adding its
+   `-Dtest-case` to `CASES` in [test/qemu_test.py](../test/qemu_test.py) (with
+   `smp: true` / a `mem` bump where noted) so the gate is runnable.
+4. **Run the gate**: `python3 test/qemu_test.py <case>`, then the full **guardrail
+   set**, then `zig build` (clean) and `zig build test` (green).
+5. **Decide, do not ask:**
+   - **Green** = the milestone's case prints its stated marker(s) and reports `PASS`,
+     the whole guardrail set passes, `zig build` is clean, and host tests are green.
+     → tick this milestone's boxes **and** its `**Gate:**`-referenced case, `git commit`
+     (`threads(M<n>): <summary>`, no `Co-Authored-By` trailer per
+     [coding-standards.md](coding-standards.md)), and continue to the next milestone in
+     the same iteration if budget remains; otherwise let the loop re-fire.
+   - **Red** = anything above fails. Diagnose from the captured serial log
+     (`zig-out/qemu-test/<case>-failed-serial.log`) and fix in place, then re-run — up to
+     **3 fix attempts** for that gate. A concurrency case that fails then passes on a
+     bare re-run is **flaky, not green**: re-run it **twice more** and treat green only
+     if it passes all; otherwise fix the race (a real threading bug), don't paper over
+     it.
+6. **A genuinely ambiguous fork is not a stop.** Pick the option most consistent with
+   [threading.md](threading.md)'s *Locked decisions*, note the choice in the commit
+   message, and continue. Do not pause for confirmation on in-scope, reversible work —
+   this plan is that authorization.
+
+**The only stop conditions:**
+
+- **Done** — every milestone box is checked (M1–M6), `zig build` clean, whole
+  `thread-*` suite + guardrail green. Update threading.md's status line to "built" (that
+  is M6's own task) and stop.
+- **Blocked** — a gate is still red after 3 fix attempts, or a step needs something
+  outside the repo (a toolchain change, new hardware, a decision no locked decision
+  covers). Append `> **BLOCKED (M<n>):** <what failed, what was tried, the serial
+  marker missing>` under that milestone, commit the WIP on the branch, and stop. Do not
+  thrash further and do not silently skip the milestone.
+
+Nothing else warrants stopping — not "should I proceed?", not "is this right?". The
+checkboxes + git history are the resumable record; the next iteration picks up from the
+first unchecked box.
+
+---
+
+## M1 — Address-space refcount (kernel foundation, no API, no behaviour change) ✅
+
+The one invariant change threads require, landed and proven **before** anything shares
+an address space. Today aspace is 1:1 with a task and teardown destroys it on any user
+task's exit; make destruction happen on the **last** exit.
+
+- [x] A refcount keyed by the address-space root, held in `scheduler.zig`
+      (`aspace_refs`): `retainAspace` takes a reference in `spawnUserLocked` (on the
+      success path, after the slot + stack are secured), all under the big kernel lock.
+- [x] Both task-teardown paths ([scheduler.zig](../system/kernel/scheduler.zig):
+      `exitUserLocked` and `destroyTaskLocked`) call `releaseAspace`, which decrements
+      and only `destroyAddressSpace`s at **zero**; an unretained space (hand-built test
+      spaces) is destroyed directly, preserving prior behaviour.
+- [x] `-Dtest-case=aspace-refcount`: spawn and reap several ring-3 processes in sequence
+      and assert (via test-observable `liveAspaceCount`/`aspaceDestroyCount`) that the
+      live-space count returns to **baseline** and destructions advance by exactly that
+      many — each space destroyed exactly once, no leak, no double-free. (Refcount
+      observables, not raw frame counts, since kernel stacks are still leaked on exit.)
+
+**Gate (met):** `python3 test/qemu_test.py aspace-refcount` passes
+(`aspace-refcount: spaces released to baseline ok` → `DANOS-TEST-RESULT: PASS`), and the
+full guardrail set passes unchanged — 13/13 (`smoke`, `sched`, `priority`, `smp`,
+`affinity`, `process`, `process-kill`, `supervision`, `fault-recovery`,
+`vfs-client-death`, `ipc`, `ipc-cap`, `display-service`); default `zig build` clean,
+`zig build test` green. The reframing is invisible until an aspace is actually shared.
+
+## M2 — `thread_spawn` + `thread_exit`: a thread runs in the shared address space ✅
+
+Spawn only — no join yet. Prove a second task executes in the **caller's** address
+space and exits cleanly.
+
+- [x] [abi.zig](../system/abi.zig): `thread_spawn = 37`, `thread_exit = 38`. Handlers in
+      process.zig; `thread_spawn` calls `scheduler.spawnThread` (shares the caller's
+      aspace, `retainAspace`); `thread_exit` ends the task like a process `exit(0)`
+      (`terminateCurrent` → `releaseAspace`). The closure pointer is delivered in the new
+      thread's **rdi** via a new `jump_to_user_arg` asm path (`t.user_arg`, 0 for a
+      process) — no naked runtime asm.
+- [x] `library/runtime/thread.zig` (barrel-exported as `runtime.Thread`): `spawn` maps a
+      stack (`mmap`), heap-allocates the `{args}` closure, and calls
+      `thread_spawn(&Closure.entry, stack_top, closure)`; `Closure.entry` (a plain C-ABI
+      Zig fn, closure in rdi) runs the function and calls `thread_exit`. Stack top is
+      16-aligned-minus-8 for the C entry.
+- [x] A `threaded` flag on the user-binary recipe (`addThreadedUserBinary` →
+      `single_threaded = false`); `thread-test` is the first opt-in binary.
+- [x] `-Dtest-case=thread-spawn`: `thread-test` spawns a worker that writes a sentinel to
+      a **shared** global and release-stores `done`; the main thread acquire-polls `done`
+      and asserts the shared global holds the sentinel — proof the worker ran in the same
+      address space.
+
+**Gate (met):** `python3 test/qemu_test.py thread-spawn` passes
+(`thread-test: child ran in shared aspace ok` → `DANOS-TEST-RESULT: PASS`); guardrail set
+16/16 green (incl. `args`/`init`/`process`, which exercise the new `jump_to_user_arg`
+process path with arg 0) plus `aspace-refcount`; `zig build` clean, `zig build test`
+green.
+
+> **Note (deferred to M3+):** the mmap arena is per-*task* (`heap_next`), so two threads
+> in one aspace that both `mmap` would collide. Fine for M2 (only the parent maps, for the
+> child's stack); make the arena per-aspace and the runtime heap thread-safe alongside the
+> `Mutex` work (M5).
+
+## M3 — `join` + `detach` + real parallelism ✅
+
+- [x] `join` over the existing exit-notification path
+      ([process-lifecycle.md](process-lifecycle.md)): `thread_spawn` gained a 4th arg, an
+      `exit_endpoint` handle (resolved + refcounted like `spawnProcessSupervised`, via
+      `spawnThreadSupervised`); `join` blocks in `ipc_reply_wait` on that endpoint until
+      the child-exit notice for its `tid`, then `munmap`s the stack. `detach` relinquishes
+      the join right (its stack is reclaimed at process exit — kernel-reaper reclaim for
+      detached threads is deferred; see note).
+- [x] `runtime.Thread.join` / `detach`, plus `Thread.currentCore()` (a new `current_core`
+      = 39 syscall) for the parallelism proof. `getCurrentId` deferred to M6 (TLS), where
+      a lighter self-id fits. The closure now rides the **thread's own stack** (not the
+      heap) — private per thread, so spawn/join touch no shared heap.
+- [x] `-Dtest-case=thread-join` (`smp: 4`): `thread-test` join mode spawns N=4 workers
+      that each do K=100k `@atomicRmw`-increments on a shared counter and stamp the core
+      they ran on; the main thread joins all N and asserts `counter == N*K` **and**
+      `@popCount(cores_seen) > 1` (genuine cross-core parallelism), then a detached worker
+      proves `detach` runs without a join.
+
+**Gate (met):** `python3 test/qemu_test.py thread-join` passes (`thread-test: join ok` →
+`DANOS-TEST-RESULT: PASS`), robust across 4 runs; guardrail 17/17 green (incl. `smp`,
+`affinity`, `process-kill`, and `args`/`init`/`process` on the exit-endpoint spawn path)
+plus `aspace-refcount`/`thread-spawn`; `zig build` clean, `zig build test` green.
+
+> **Note (deferred):** a detached thread's stack is freed only at process exit (not by the
+> reaper on thread exit) — kernel user-stack tracking + reclaim is a later refinement. And
+> the runtime heap is still not thread-safe: threads that both allocate concurrently would
+> race (the thread *machinery* avoids the heap, but worker code sharing an allocator does
+> not). Both fold into the M5 `Mutex`/allocator work.
+
+## M4 — Futex: the one blocking primitive ✅
+
+- [x] [abi.zig](../system/abi.zig): `futex_wait = 40`, `futex_wake = 41`. A waiter is a
+      `.blocked` task tagged with `Task.futex_addr` (no queue linkage);
+      `futex_wait(addr, expected, timeout_ns)` reads the user word under the big lock,
+      parks iff `*addr == expected`, and returns on wake or timeout; `futex_wake(addr,
+      count)` scans the task table and readies up to `count` matching waiters (same
+      address space). No spinning — a parked waiter leaves its core free to `hlt`. A
+      timed wait also sets `wake_at`, so the timer's `wakeExpired` wakes it; `futex_addr`
+      staying non-zero (only `futex_wake` clears it) is how the waiter tells timeout from
+      a real wake.
+- [x] `runtime.Thread.Futex` (`wait` / `timedWait` / `wake`) over the syscall wrappers.
+- [x] `-Dtest-case=thread-futex` (`smp: 4`): a waiter thread prints `waiting` and
+      `futex_wait`s on a word; the main thread publishes it, prints `waking`, and
+      `futex_wake`s; the waiter prints `woke`. Then a `timedWait` on an unwoken word
+      reports `error.Timeout`.
+
+**Gate (met):** `python3 test/qemu_test.py thread-futex` passes, robust across 3 runs —
+the case's **ordered** regex asserts `waiting → waking → woke → PASS` on the serial
+stream (the handoff proof), and `thread-futex: timeout ok` confirms the timeout.
+Guardrail 18/18 green (incl. `sleep`/`event`/`ipc` blocking paths) + `aspace-refcount`,
+`thread-spawn`, `thread-join`; `zig build` clean, `zig build test` green.
+
+> **Note:** the kernel test checks only the freshest verdict marker via `bufferHas` (the
+> in-memory log ring buffer evicts older lines); ordering is asserted against the full
+> serial stream by the qemu regex instead.
+
+## M5 — `Mutex` + `Condition` + `Semaphore` ✅
+
+- [x] `runtime.Thread.Mutex` (three-state futex mutex: CAS fast path, `futex_wait`/`wake`
+      slow path), `Condition` (`wait`/`timedWait`/`signal`/`broadcast`, a futex sequence
+      counter), `Semaphore` (permits over `Mutex`+`Condition`) — the same state machines
+      `std.Thread` uses, ported onto our `Futex`.
+- [x] `-Dtest-case=thread-mutex` (`smp: 4`): a bounded producer/consumer — 2 producers +
+      2 consumers over one `Mutex` and two `Condition`s move N=2000 unique items through
+      an 8-slot ring; the consumed checksum and tally match exactly (no lost/duplicated
+      item, no overrun) under real cross-core contention. The small ring forces producers
+      to block on full and consumers on empty, exercising `Condition.wait`.
+
+**Gate (met):** `python3 test/qemu_test.py thread-mutex` passes (`thread-mutex: ok` →
+`DANOS-TEST-RESULT: PASS`), robust across 3 runs; guardrail 17/17 green (incl.
+`sleep`/`event`/`ipc`) + all M1–M4 thread cases; `zig build` clean, `zig build test`
+green.
+
+> **Deferred (with rationale):**
+> - **`join` → futex completion word** — the exit-endpoint join (M3) is correct and
+>   tested. A futex-completion join needs the *kernel* to clear+wake a word after the
+>   thread is fully off its stack (a CLONE_CHILD_CLEARTID-style mechanism); doing it in
+>   the thread's own trampoline would let `join` `munmap` the stack while the thread still
+>   runs on it (use-after-free). Left on the exit-endpoint path; the kernel clear-on-exit
+>   is a later, separate refinement.
+> - **Host unit tests for the state machines** — `Mutex`/`Condition` bottom out in the
+>   `futex_*` syscalls, unavailable on the host without a mockable `Futex` seam. The QEMU
+>   `thread-mutex` gate exercises them under real concurrency instead; a host-side mock is
+>   future work.
+
+## M6 — `getCurrentId`, docs, and CI wiring ✅
+
+- [x] `getCurrentId` via a small `thread_self = 42` syscall (`runtime.Thread.getCurrentId`
+      returns the kernel task id). **Per-thread `threadlocal` TLS is deferred** — no
+      consumer needs it, and it would require context-switching `fs.base` per task (real
+      kernel + per-switch cost) for an unused feature; threaded binaries have run fine
+      without it through M2–M5. threading.md's TLS reasoning already scoped it as
+      deferred-unless-needed. When a consumer appears, the shape is: `thread_spawn`
+      allocates a per-thread TLS block, sets `fs.base`, and the context switch saves/
+      restores it.
+- [x] `RwLock` / `WaitGroup` deferred (no consumer yet); they slot onto the same
+      `Futex`/`Mutex`/`Condition` when wanted.
+- [x] All `thread-*` cases wired into [test/qemu_test.py](../test/qemu_test.py)
+      (`thread-spawn`/`-join`/`-futex`/`-mutex`/`-id`); threading.md + docs/README.md
+      status updated to **built**; the worked example is threading.md's win-condition.
+- [x] `-Dtest-case=thread-id` (`smp: 4`): two workers read `getCurrentId`; the main
+      thread confirms all three ids are non-zero and distinct — each thread has its own
+      kernel identity. (Renamed from `thread-tls`, which implied `threadlocal`.)
+
+**Gate (met):** `python3 test/qemu_test.py thread-id` passes; the whole `thread-*` suite
+(`thread-spawn`/`-join`/`-futex`/`-mutex`/`-id`) plus the full guardrail set pass; default
+`zig build` clean, `zig build test` green.
+
+---
+
+## Status: built
+
+M1–M6 complete. danos has `runtime.Thread` — `spawn`/`join`/`detach`, cross-core
+parallelism, futex, and `Mutex`/`Condition`/`Semaphore`, all over a private thread ABI
+behind the runtime. Deferred (with rationale, no consumer yet): `threadlocal` TLS,
+`RwLock`/`WaitGroup`, kernel clear-on-exit for a futex-completion `join`, a per-aspace
+mmap arena / thread-safe runtime heap, and host-side unit tests via a mockable `Futex`.
+
+---
+
+## Deferred (explicitly not in this plan)
+
+- **Cross-process shared-memory futex** — the `(aspace, vaddr)` key can become a
+  physical-address key so two processes share a futex through an [shm](display-v2.md)
+  region. Not needed for intra-process threads.
+- **Per-thread priorities / affinity distinct from the process** — threads inherit the
+  process priority ([scheduling.md](scheduling.md)); revisit only if it earns its keep.
+- **Per-thread signal delivery** — signals stay process-scoped
+  ([process-lifecycle.md](process-lifecycle.md)).
+- **A `pthread`/POSIX surface** — the API is `std.Thread`-shaped Zig, nothing more.
+- **A real `std.Thread` backend** — arrives with self-hosting
+  ([zig-self-hosting.md](zig-self-hosting.md)); it sits on these same primitives, so it
+  swaps the impl under `runtime.Thread`, not the call sites.

docs/threading.md

@@ -0,0 +1,316 @@
+# Threading: `runtime.Thread`, a std-shaped API over a private thread ABI
+
+A note on danos **threads** — several tasks sharing one address space — provided by a
+`runtime.Thread` type that mirrors the shape of Zig's `std.Thread` while keeping every
+kernel entry behind the [runtime](../library/runtime). **Built** (M1–M6, see
+[threading-plan.md](threading-plan.md)): `spawn`/`join`/`detach`, cross-core
+parallelism, a futex (`futex_wait`/`futex_wake`), and a futex-backed
+`Mutex`/`Condition`/`Semaphore`, plus `getCurrentId`/`currentCore`. Deferred by design
+(no consumer yet): per-thread `threadlocal` TLS, `RwLock`/`WaitGroup`, and migrating
+`join` to a futex completion word — see the plan's M5/M6 notes. The analysis is against
+**Zig 0.16** (the pinned toolchain); `std.Thread`'s internals move between releases, so
+treat upstream shapes as "0.16.x."
+
+## The win condition
+
+A danos service can write
+
+```zig
+const t = try runtime.Thread.spawn(.{}, worker, .{ctx});
+// ... do other work concurrently ...
+t.join();
+```
+
+and get real parallelism across cores — with `runtime.Thread.Mutex`,
+`runtime.Thread.Condition`, and `runtime.Thread.Semaphore` available for
+coordination — **without any code path reaching the kernel except through the
+runtime**. The call sites read exactly like `std.Thread`, so the day danos becomes a
+real Zig target (see [self-hosting](#the-self-hosting-endgame)) we swap the
+implementation underneath, not the API above.
+
+## Locked decisions (do not relitigate)
+
+- **We build `runtime.Thread`, not literal `std.Thread`.** It mirrors std's *API and
+  features*; the implementation underneath is danos-native. See
+  [Why not literal std.Thread](#why-not-literal-stdthread).
+- **Threads are a narrow, opt-in capability — not the default concurrency tool.** The
+  default for resilience stays **process + IPC** ([resilience.md](resilience.md),
+  [ipc.md](ipc.md)). See [Where threads fit](#where-threads-fit-the-resilience-tension).
+- **Blocking synchronization is futex-backed, never spin-backed.** Waiters sleep in
+  the kernel so an idle core still halts ([halting.md](halting.md)).
+- **Per-binary opt-in to multi-threaded codegen.** Only a service that asks for
+  threads is built `single_threaded = false`; the rest stay lean and single-threaded.
+- **The thread ABI is private.** New syscalls extend [abi.zig](../system/abi.zig)
+  `SystemCall` and are reached only through `library/runtime` wrappers, exactly like
+  every other danos syscall ([syscall.md](syscall.md)) — numbers stay renumberable.
+
+## Why not literal `std.Thread`
+
+danos's ABI invariant is that the **runtime is the sole holder of the syscall ABI**,
+and that ABI is private and renumberable ([syscall.md](syscall.md) — "unstable
+private ABI"). That is a security and evolvability asset: no compiled binary can
+hardcode a syscall number, and the kernel can renumber freely because only the
+runtime — rebuilt in lockstep — knows the mapping.
+
+`std.Thread` is incompatible with that invariant on two counts:
+
+1. **It selects its backend from `builtin.os.tag`, and issues syscalls directly.**
+   danos targets `.os_tag = .freestanding` ([build.zig](../build.zig)), for which
+   `std.Thread` resolves to an unsupported stub that `@compileError`s. Adding a real
+   backend would either bake danos syscall numbers into std (breaking ABI privacy and
+   renumbering) or fork std to route back through the runtime — a permanent rebase
+   cost that buys nothing the native type doesn't.
+2. **Our user binaries are built `single_threaded = true`** ([build.zig](../build.zig)
+   `addUserBinary`), which compiles threading out entirely and makes atomics and TLS
+   single-threaded. Threads need this flipped per binary regardless.
+
+So we take the *shape* of `std.Thread`, not the *type*. The cost of replicating the
+surface (spawn/join/Mutex/Condition) is small; the cost of the std type is the ABI
+invariant.
+
+## Where threads fit: the resilience tension
+
+Threads are in genuine tension with a resilience-first microkernel, and it is worth
+being explicit so we do not reach for them by reflex.
+
+The reason danos pays for a microkernel is **fault isolation**
+([resilience.md](resilience.md)): a component corrupts its own address space, faults,
+and is **restarted** without touching anyone else — because the boundary *is* the
+address space. Threads deliberately remove that boundary *within* a process:
+
+- Threads share one address space, so one thread's stray write corrupts them all —
+  there is no isolation **between** threads.
+- Threads share fate: a fault in any thread, or a "kill the process" decision, takes
+  down **all** of them. Restartability lives at the process level, not the thread
+  level.
+- Shared mutable state reintroduces data races — the failure class the
+  isolate-and-message model was chosen to avoid.
+
+**Therefore:** the default answer to "make X concurrent" stays *another process over
+IPC* (isolated, independently restartable) or a single event loop with several
+message sources. Reach for a thread only inside **one** service that needs genuine
+**shared-memory, low-latency parallelism** and can accept intra-service fate-sharing —
+e.g. a compositor splitting tile compositing across cores, where per-tile IPC would be
+too chatty. "Input on one thread, display on another" is *not* that case; it wants two
+processes. The isolation boundary stays at process granularity.
+
+## The API surface (mirrors `std.Thread`)
+
+Lives in `library/runtime/thread.zig`, re-exported as `runtime.Thread`.
+
+```zig
+pub const Thread = struct {
+    pub const Id = u32;                       // the kernel task id
+    pub const SpawnConfig = struct {
+        stack_size: usize = default_stack_size,
+        allocator: ?std.mem.Allocator = null, // for the closure + stack bookkeeping
+    };
+    pub const SpawnError = error{ OutOfMemory, ThreadQuotaExceeded, SystemResources };
+
+    pub fn spawn(config: SpawnConfig, comptime function: anytype, args: anytype) SpawnError!Thread;
+    pub fn join(self: Thread) void;           // block until the thread ends, reclaim its stack
+    pub fn detach(self: Thread) void;         // give up the right to join; kernel reclaims on exit
+    pub fn getCurrentId() Id;
+    pub fn yield() void;                       // -> existing `yield` syscall
+
+    pub const Mutex     = struct { pub fn lock(*Mutex) void; pub fn tryLock(*Mutex) bool; pub fn unlock(*Mutex) void; };
+    pub const Condition = struct { pub fn wait(*Condition, *Mutex) void; pub fn timedWait(*Condition, *Mutex, u64) error{Timeout}!void; pub fn signal(*Condition) void; pub fn broadcast(*Condition) void; };
+    pub const Semaphore = struct { pub fn wait(*Semaphore) void; pub fn post(*Semaphore) void; };
+    pub const Futex     = struct { pub fn wait(*const atomic.Value(u32), u32) void; pub fn timedWait(...) error{Timeout}!void; pub fn wake(*const atomic.Value(u32), u32) void; };
+    // RwLock / ResetEvent / WaitGroup follow the same pattern, added as needed.
+};
+```
+
+Deviations from `std.Thread`, called out honestly:
+
+- **The thread function's return value is discarded** (as `std.Thread.join` returns
+  `void`). Return data through shared state or a `Semaphore`/`Condition`, not the
+  return.
+- `getCpuCount()` maps to the existing SMP core count ([smp.md](smp.md)); a service
+  rarely needs it.
+
+## Kernel primitives (new private syscalls)
+
+Four new entries extend [abi.zig](../system/abi.zig) `SystemCall` after
+`shm_physical = 36`, each with a `library/runtime` wrapper:
+
+| Syscall | Signature | Purpose |
+|---|---|---|
+| `thread_spawn` | `(entry, stack_top, arg) -> tid` | create a task sharing the **caller's** address space |
+| `thread_exit`  | `(stack_base, stack_len)` | end the calling thread; hand back its stack range for reclaim |
+| `futex_wait`   | `(addr, expected, timeout_ns) -> status` | block if `*addr == expected`, until woken or timeout |
+| `futex_wake`   | `(addr, count) -> woken` | wake up to `count` waiters on `addr` |
+
+Plus one invariant change with no new syscall: **address-space reference counting**.
+
+## Mechanics
+
+### Address-space reference counting
+
+Today an address space is 1:1 with a task: `spawnUserLocked` records `aspace` on the
+Task, and teardown does `destroyAddressSpace(t.aspace)` when **any** user task exits
+([scheduler.zig](../system/kernel/scheduler.zig)). With threads, several tasks share
+one `aspace`, so the first to exit would rip the address space out from under its
+siblings.
+
+Fix: a small refcount keyed by the address-space root (`createAddressSpace` in
+[process.zig](../system/kernel/process.zig) sets it to 1). `thread_spawn` increments
+it; task teardown decrements and only calls `destroyAddressSpace` at **zero**. All of
+this is already under the big kernel lock, so no new locking. This is the one piece
+that must land and be proven before anything shares an address space.
+
+### `thread_spawn` and the trampoline
+
+The scheduler already accepts an arbitrary `aspace` and does **not** smuggle values
+through registers — `startUserTask` reads the entry/stack from the Task and
+`jumpToUser`s ([scheduler.zig](../system/kernel/scheduler.zig)). That makes the thread
+path clean:
+
+1. The runtime's `spawn` `mmap`s a stack (syscall `4`), heap-allocates a closure —
+   `{ fn_ptr, args_tuple, completion }`, the std "Instance" pattern — and writes the
+   closure pointer to the **top word of the new stack**.
+2. It calls `thread_spawn(entry = &threadTrampoline, stack_top, arg = closure_ptr)`.
+   The kernel calls the same `spawnUserLocked` path with the **caller's aspace**
+   (refcount++), `entry`, and `user_sp = stack_top`.
+3. `threadTrampoline` (a small runtime shim) reads the closure off its stack, calls
+   the user function, then calls `thread_exit`. No new register ABI — the closure
+   pointer rides the stack the runtime set up, mirroring how `startUserTask` avoids
+   register smuggling.
+
+Unlike a process start, there is **no** System V argc/argv/auxv block
+([sysv.md](sysv.md)) — a thread stack carries only the closure pointer.
+
+### Lifetime: exit, join, detach, stack reclaim
+
+- **`thread_exit`** marks the task dead and hands the kernel the thread's user-stack
+  range. The kernel reaps the task on the scheduler (already running on a *kernel*
+  stack, so it can safely unmap the user stack), decrements the aspace refcount, and
+  frees the task slot.
+- **`join` — Stage 1** reuses the existing exit-notification machinery
+  ([process-lifecycle.md](process-lifecycle.md)): `spawn` passes a per-thread
+  `exit_endpoint`, and `join` blocks in `ipc_reply_wait` until the child-exit
+  notification for that `tid` arrives, then `munmap`s the stack. No futex needed to
+  land spawn/join.
+- **`join` — Stage 2 refinement** migrates to the std shape: a `completion` word in
+  the closure that `thread_exit`'s trampoline `futex_wake`s and `join` `futex_wait`s
+  on — dropping the per-thread endpoint. Kept as a refinement so Stage 1 ships first.
+- **`detach`** relinquishes the join right; the kernel reclaims the stack and slot on
+  `thread_exit` (a detached thread's stack range is unmapped by the reaper, since no
+  joiner will).
+
+### Futex, and the sync primitives on top
+
+`futex_wait`/`futex_wake` are the one blocking primitive; `Mutex`, `Condition`, and
+`Semaphore` are ordinary user-space state machines over an `atomic.Value(u32)` that
+call the futex wrappers on the slow path — the same construction `std.Thread` uses,
+so the algorithms port directly.
+
+Keying: threads share an address space, so a **virtual address within that aspace**
+identifies a futex uniquely; the kernel keys its wait queue by `(aspace_root, vaddr)`.
+Keying by the **physical** address instead (translate `vaddr -> paddr` on entry) is a
+deliberate forward door: it lets two *processes* share a futex through an
+[shm](display-v2.md) region later, without changing the API. We start with the
+private-per-aspace key and note the physical-key upgrade.
+
+No spinning: a contended lock parks the task in the kernel and the core is free to run
+other work or `hlt` ([halting.md](halting.md)). This is why futex is a locked
+decision, not a "maybe later."
+
+### TLS and `getCurrentId`
+
+danos sets up no `fs.base` TLS today (fine under `single_threaded`). Two scoped needs:
+
+- **`getCurrentId`** returns the kernel task id — either a trivial syscall or, better,
+  a value the runtime stashes in a per-thread control block.
+- **`threadlocal` variables** need a real per-thread TLS block and `fs.base` set per
+  thread. `thread_spawn` sets `fs.base` to a runtime-allocated per-thread block; full
+  `threadlocal` support is Stage 3, only if a consumer needs it. Nothing in the core
+  spawn/join/mutex path requires `threadlocal`.
+
+### Build: multi-threaded codegen, opt-in
+
+`addUserBinary` gains a `threaded: bool = false` parameter; when set it builds that
+binary `single_threaded = false` so atomics and (later) TLS are real. Threads and
+atomics are unsound in a `single_threaded` image, so a binary must opt in **before**
+it may call `runtime.Thread.spawn`. Everyone else stays single-threaded and lean.
+
+## Interaction with the rest of the kernel
+
+- **Scheduler / SMP** ([scheduling.md](scheduling.md), [smp.md](smp.md)): a thread is
+  just another `Task` with an `aspace` shared with its siblings; the existing
+  per-core ready queues, priorities, and affinity apply unchanged. Threads of one
+  process can run on different cores simultaneously — that is the point.
+- **Halting** ([halting.md](halting.md)): futex-parked waiters keep the "idle core
+  halts" property intact under lock contention — no busy-wait.
+- **Lifecycle** ([process-lifecycle.md](process-lifecycle.md)): killing a process
+  must kill *all* its threads and only then drop the last aspace ref. The kill path
+  already targets a process; it fans out to every task on that aspace.
+- **Resilience** ([resilience.md](resilience.md)): a faulting thread kills its whole
+  process (shared fate). The supervisor restarts the **process**, which respawns its
+  threads from a known-good state — restart granularity stays the process.
+
+## Build-out plan (staged, each gate serial-checkable)
+
+The ordered, `/loop`-runnable milestones live in
+**[threading-plan.md](threading-plan.md)** (shaped like
+[display-v2-plan.md](display-v2-plan.md)): every milestone lands on its own and ends in
+a verifiable gate (`python3 test/qemu_test.py <case>`, asserting serial markers;
+`zig build test` for host unit tests). The stages below are the shape it expands.
+
+- **Stage 0 — address-space refcount.** Refcount on the aspace root; teardown destroys
+  at zero. No API yet; nothing shares an aspace, so refcount is 1 everywhere.
+  *Gate:* the full QEMU suite stays green (no regression) — proves the reframing is
+  invisible until used.
+- **Stage 1 — spawn / join / detach.** `thread_spawn` + `thread_exit`, the trampoline,
+  stacks via `mmap`, join over the exit-endpoint, the `threaded` build flag.
+  *Gate:* `-Dtest-case=thread-spawn` — a threaded test service spawns N threads that
+  each `@atomicRmw`-increment a shared counter, the parent joins all N, and asserts
+  the total is exactly N × iterations. Runs `smp` (multi-core) to prove real
+  parallelism.
+- **Stage 2 — blocking synchronization.** `futex_wait`/`futex_wake` + `Futex`,
+  `Mutex`, `Condition`, `Semaphore`; optionally migrate join to a futex completion
+  word. *Gate:* `-Dtest-case=thread-mutex` — a bounded producer/consumer over a
+  `Mutex` + `Condition` moves K items with no lost wakeups and no busy-wait (assert
+  the consumer blocked, e.g. via a low idle tick count).
+- **Stage 3 — polish.** Per-thread TLS / `fs.base` and `threadlocal` (only if a
+  consumer needs it), `RwLock`/`WaitGroup` as demanded, and this doc's cases wired
+  into [test/qemu_test.py](../test/qemu_test.py).
+
+## Conventions
+
+Follow [coding-standards.md](coding-standards.md): spell out non-acronym
+abbreviations, kebab-case file names, no `Co-Authored-By` trailers. New syscalls
+extend [abi.zig](../system/abi.zig) `SystemCall` + a `library/runtime` wrapper
+([syscall.md](syscall.md)). `runtime.Thread` is a first-class runtime module, the same
+way `runtime.process` ([process-lifecycle.md](process-lifecycle.md)) and `runtime.ipc`
+are — user code never names a syscall.
+
+## Non-goals
+
+- **No preemptive user-space signals delivered to a specific thread.** Signals stay
+  process-scoped ([process-lifecycle.md](process-lifecycle.md)).
+- **No thread priorities distinct from the process.** Threads inherit the process
+  priority; per-thread priority is a later question if it ever earns its keep.
+- **No cross-process shared-memory futex yet** — the physical-address key leaves the
+  door open, but the first cut is private-per-aspace.
+- **No `pthread`/POSIX surface.** The API is `std.Thread`-shaped Zig, nothing more.
+
+## The self-hosting endgame
+
+When danos becomes a real Zig target and we (eventually) add a danos backend to std
+([zig-self-hosting.md](zig-self-hosting.md)), `std.Thread` can sit *on top of* these
+same kernel primitives — the danos `std.Thread.Impl` would call the very
+`thread_spawn`/`futex_*` wrappers `runtime.Thread` already uses. Because
+`runtime.Thread` was built API-compatible from day one, that transition swaps the
+implementation, not a single call site. Designing to the std shape now is what makes
+the later self-hosting lift cheap.
+
+## Further reading
+
+- [scheduling.md](scheduling.md), [smp.md](smp.md) — the task model these threads join.
+- [resilience.md](resilience.md), [vision.md](vision.md) — why isolation is the default
+  and threads are the exception.
+- [syscall.md](syscall.md), [ipc.md](ipc.md) — the private ABI and the messaging model
+  threads sit beside.
+- [halting.md](halting.md) — the idle/halt property futex-backed blocking preserves.
+- [zig-self-hosting.md](zig-self-hosting.md) — the target this bends toward.

docs/vdso.md

@@ -0,0 +1,206 @@
+# The vDSO — the public system-call boundary
+
+> **Status:** design note, not built. The runtime today issues raw `syscall`
+> instructions from `library/runtime/system-call.zig` using the numbers in
+> `system/abi.zig`. This note designs the layer that replaces that arrangement:
+> a **kernel-supplied, C-ABI entry library** mapped into every process — the
+> only supported way into the kernel — so the raw numbers can stay private,
+> be renumbered at will, and eventually be randomised per boot.
+
+## Why: the ABI danos promises, and the one it doesn't
+
+`system/abi.zig` is the **private** kernel ↔ runtime contract. Its header says
+so: the numbers are an implementation detail the runtime hides and may
+renumber, the same split as libSystem over the XNU syscalls on macOS or win32
+over the NT syscalls on Windows. Linux — with its world-visible, frozen
+syscall table — is the outlier, not the norm.
+
+That stance has consequences the moment binaries exist that we don't rebuild
+ourselves:
+
+1. **Third-party binaries** (docs/zig-self-hosting.md) must keep working across
+   kernel updates. If they contain raw `syscall` instructions with today's
+   numbers baked in, every renumbering breaks the world — the ABI would be
+   *de facto* public no matter what the header says. Go on macOS made exactly
+   this mistake: it issued XNU syscalls directly instead of going through
+   libSystem, and macOS updates repeatedly broke every Go binary until Go
+   switched to the library like everyone else.
+2. **Not everything is Zig.** A Rust or C program can't import the `runtime`
+   module. The public boundary has to be expressible in the one calling
+   convention every language speaks: the C ABI.
+3. **Randomised syscall numbers** — a hardening option we want open — only
+   work if no user binary anywhere knows a number at build time. The binding
+   must happen at *load time*, from something the kernel controls.
+
+All three point at the same well-known shape: a **vDSO** (virtual dynamic
+shared object). The kernel carries a small blob of user-mode code, maps it
+into every process at spawn, and that blob — not the application — contains
+the `syscall` instructions. Fuchsia works exactly this way: its vDSO is the
+*only* kernel entry, version-matched by construction because the kernel itself
+injects it. Because the kernel and the blob ship as one artifact, there is
+**no version skew, no loader, no search path, and no shared file on disk** —
+which is what makes this the resilient way to have a private ABI
+(docs/resilience.md), where a conventional `ld.so` + `/lib/libdanos.so`
+arrangement would add a loader to every spawn and a single shared point of
+failure.
+
+The public danos ABI then has exactly two layers, neither of which is
+`abi.zig`:
+
+| Layer | Contract | Spoken by |
+|-------|----------|-----------|
+| **vDSO** | C-ABI functions, this note | every language's thin shim (`runtime.system` for Zig, a `-sys` crate for Rust, a header for C) |
+| **IPC wire protocols** | byte layouts over `ipc_call` ([vfs-protocol.md](vfs-protocol.md) is the first one documented) | any client that can lay out bytes |
+
+Everything above those — the heap, `runtime.fs`, the service harness — is
+per-language convenience, compiled into each binary from source, exactly as
+today. Nothing about the Zig runtime's shape changes; it just stops being the
+*only* door.
+
+## The blob
+
+A single copy of the vDSO code lives in the kernel image (built by
+`build.zig` as a tiny freestanding object, embedded like the AP trampoline).
+At boot the kernel finalises it once — this is where randomised numbers would
+be patched in — and thereafter maps the **same physical pages** read-execute
+into every process's address space. The blob is:
+
+- **Position-independent.** It is mapped at a per-process randomised base, so
+  it must be PIC (rip-relative addressing only — no relocations to process).
+- **Stateless and re-entrant.** No writable data. Anything stateful belongs to
+  the process, not the vDSO.
+- **Architecture-specific.** The x86-64 blob wraps `syscall`; an aarch64 blob
+  wraps `svc #0`. It lives beside the other per-architecture kernel sources
+  (`system/kernel/architecture/<arch>/`), selected the same way the
+  `architecture` module is (docs/arch.md).
+
+### Shape: a function table, not an ELF
+
+A real `.so` with a dynamic symbol table is the conventional vDSO shape, but
+linking against one at load time needs a dynamic linker in every binary —
+machinery danos deliberately doesn't have. Instead the v1 shape is the
+simplest thing that is still a stable contract — a **function-pointer table**
+at the vDSO base:
+
+```
+offset 0   u64  magic        'danosVDS' — a mapped-the-wrong-thing guard
+offset 8   u64  api_level    incremented when the table grows
+offset 16  u64  count        number of table entries that follow
+offset 24  u64  table[count] function pointers into the vDSO's own code
+```
+
+Table *indices* are the public constants (published in a C header,
+`danos.h`), assigned once and append-only — the same discipline the IPC
+protocols use for operation values. The pointers point at stubs inside the
+blob; what those stubs put in `rax` is nobody's business but the kernel's.
+A language shim binds in one step: read the base from the init block, check
+the magic, keep the table pointer. Feature detection for a binary built
+against older headers is `count`/`api_level` — a kernel never removes or
+reorders entries.
+
+(If danos ever grows a real dynamic linker, the same blob can additionally
+present an ELF `dynsym` without breaking the table — Fuchsia's vDSO is
+likewise both a mappable blob and a linkable `.so`. That is a later
+convenience, not a requirement.)
+
+### Delivery: the auxiliary vector
+
+The kernel already builds a System V entry block — argc, argv, envp
+terminator, **auxiliary vector** — on every new process's stack
+(`buildEntryStack`, read by `runtime.start`). The vDSO base rides in a new
+auxv entry, exactly Linux's `AT_SYSINFO_EHDR` move. No new syscall, no magic
+address, and a language shim finds it the same portable way on every
+architecture.
+
+## The function surface
+
+One table entry per kernel call, C ABI (System V AMD64), names prefixed
+`danos_`. The current `SystemCall` set maps directly; integer arguments and
+returns are `u64`, errors return as negative values exactly as today.
+
+The calls that return two values in `rax:rdx` today — `dma_alloc`
+(vaddr + paddr), `msi_bind` (address + data), `shm_create` (vaddr + handle) —
+become functions returning a two-`u64` struct. The System V ABI returns a
+16-byte struct in `rax:rdx`, so the stub is a plain `syscall; ret` — the
+C-ABI spelling of the existing convention, at zero cost.
+
+Grouped as `abi.zig` groups them:
+
+| Group | Functions |
+|-------|-----------|
+| process | `danos_exit`, `danos_yield`, `danos_sleep`, `danos_spawn`, `danos_process_enumerate`, `danos_process_kill`, `danos_process_exit_reason`, `danos_process_subscribe`, `danos_process_signal`, `danos_signal_bind` |
+| memory | `danos_mmap`, `danos_munmap`, `danos_dma_alloc`, `danos_dma_free`, `danos_shm_create`, `danos_shm_map`, `danos_shm_physical` |
+| ipc | `danos_endpoint_create`, `danos_ipc_register`, `danos_ipc_lookup`, `danos_ipc_call`, `danos_ipc_reply_wait`, `danos_ipc_send` |
+| devices | `danos_device_enumerate`, `danos_device_claim`, `danos_device_register`, `danos_mmio_map`, `danos_irq_bind`, `danos_irq_ack`, `danos_msi_bind`, `danos_io_read`, `danos_io_write` |
+| time | `danos_clock`, `danos_wall_clock`, `danos_timer_bind` |
+| diagnostics | `danos_debug_write`, `danos_klog_read` |
+
+The constants that ride alongside the calls — mmap protection bits, DMA
+flags, notification badge bits, `ExitReason`, `Signal`, well-known service
+ids, `page_size`, the IPC message maximum — move to the public header too:
+they are wire values a Rust program needs verbatim. What stays private in
+`abi.zig` is exactly the thing the vDSO exists to hide: the `SystemCall`
+numbers and the trap convention.
+
+## Enforcement, and an honest threat model
+
+Renumbering only has teeth if the kernel **refuses syscalls that don't come
+from the vDSO**. The check is cheap: on kernel entry, the saved user `rip`
+must lie inside the calling process's vDSO mapping; otherwise the process is
+killed with a fault-class exit reason (its supervisor restarts or gives up,
+docs/process-lifecycle.md — a foreign-syscall attempt is a bug or an attack,
+never something to limp past). Fuchsia enforces exactly this.
+
+What this buys, precisely:
+
+- **ABI freedom** — the real prize. The numbers can change per release or per
+  boot and nothing outside the kernel image cares. The private ABI stays
+  actually private, permanently.
+- **A single audited chokepoint** for kernel entry, per process, at a
+  randomised address.
+- **Raised bar for exploits**: shellcode can't issue a hard-coded `syscall`;
+  it must first discover the per-process vDSO base (ASLR) and call through
+  it.
+
+What it does *not* buy: an attacker with arbitrary code execution in a
+process can still *call* the vDSO functions — they are mapped executable in
+that process, and return-oriented chains reach them. Syscall randomisation is
+hardening, not a security boundary; the security boundary remains the
+capability model (what the process's endpoints and device claims let it do).
+It is worth building anyway — for the ABI freedom first and the hardening
+second — but the design should never be sold as more than that.
+
+## Migration
+
+Phased so every step ships alone (the M-milestone discipline):
+
+1. **The blob + the table.** Build the vDSO, map it at spawn, deliver the
+   base via auxv. `runtime.system-call.zig` binds through the table when the
+   auxv entry is present, falls back to raw `syscall` when absent — the whole
+   tree keeps booting during the transition.
+2. **Cut the runtime over.** Delete the raw stubs; `runtime` no longer
+   imports the `SystemCall` numbers at all (`abi.zig`'s enum becomes
+   kernel-internal). The QEMU suite passing proves the table carries the
+   whole system.
+3. **Enforce + randomise.** Add the `rip`-range check, then per-boot number
+   randomisation patched into the blob at kernel init. A test boots with
+   randomisation on and runs the full suite.
+4. **The other languages.** Publish `danos.h`; a Rust `danos-sys` crate wraps
+   the table. This is also the seam `std.os.danos` calls through when the Zig
+   self-hosting fork lands (docs/zig-self-hosting.md) — the vDSO is what
+   makes that seam stable across kernel versions.
+
+## What deliberately stays out
+
+- **No dynamic linker, no `/lib/*.so`.** The vDSO is kernel-injected precisely
+  so danos binaries can stay fully static above it. Sharing *library code*
+  across processes stays what it is today: a service behind IPC, or source
+  compiled into each binary.
+- **No file/device I/O in the vDSO.** The microkernel line doesn't move: the
+  vDSO wraps the same deliberately tiny table (docs/syscall.md); files are
+  still the VFS server's business over IPC.
+- **No fast-path user-mode implementations yet.** Linux's vDSO exists mostly
+  to answer `gettimeofday` without a kernel entry. `danos_clock` could one
+  day read the calibrated TSC in user mode the same way — the blob is where
+  such an optimisation would live — but that is an optimisation, not part of
+  this design's contract.

docs/vfs-protocol.md

@@ -0,0 +1,170 @@
+# The VFS wire protocol
+
+> **Status:** built and spoken today between `runtime.fs` (the client) and the
+> VFS server (`system/services/vfs`), with mounted backends (the FAT server)
+> speaking the same protocol behind the router. The Zig source of truth is
+> `system/services/vfs/protocol.zig` (the `vfs-protocol` module), whose unit
+> tests pin the sizes and values below. This page is the **language-neutral
+> wire specification** of that contract — what a Rust or C client implements
+> ([vdso.md](vdso.md) explains why the IPC protocols, not the syscall
+> numbers, are danos's public ABI).
+
+## Transport
+
+A VFS exchange is one synchronous IPC rendezvous (`ipc_call`,
+docs/ipc.md): the client sends one message and blocks; the server replies
+with one message. The endpoint is found by well-known service id
+(`ipc_lookup`, service id **1** = vfs).
+
+- A message is at most **256 bytes** (`message_maximum`).
+- A request is a fixed 32-byte **Request** header followed by an inline
+  payload of at most **224 bytes** (`maximum_payload`) — a path, or write
+  bytes. There is no multi-message request: paths and single reads/writes
+  must fit, and larger transfers loop (see *read* / *write*).
+- A reply is a fixed 24-byte **Reply** header followed by an inline payload —
+  read bytes, a `FileStatus`, or a `DirectoryEntry`.
+- All integers are **little-endian**; layouts are C layout for x86-64
+  (`extern struct`), offsets given below so nothing need be inferred.
+
+The kernel never parses any of this — it only moves the bytes
+(docs/syscall.md); files are entirely a user-space affair.
+
+## Request header — 32 bytes
+
+| offset | size | field | meaning |
+|-------:|-----:|-------|---------|
+| 0 | 4 | `operation` | an **Operation** value (below) |
+| 4 | 4 | — | padding |
+| 8 | 8 | `node` | the server-side open-node id from a prior `open`; 0 for path-based operations |
+| 16 | 8 | `offset` | byte position for read/write; entry index (cursor) for readdir; else 0 |
+| 24 | 4 | `len` | payload length for path/write operations; requested byte count for read |
+| 28 | 4 | `flags` | open flags (below); else 0 |
+
+## Reply header — 24 bytes
+
+| offset | size | field | meaning |
+|-------:|-----:|-------|---------|
+| 0 | 4 | `status` | **0 = success**, negative = failure (signed) |
+| 4 | 4 | — | padding |
+| 8 | 8 | `node` | the new open-node id (for `open`); else 0 |
+| 16 | 4 | `len` | reply payload length in bytes |
+| 20 | 4 | — | padding |
+
+On failure the router replies `status = -1`; a mounted backend's negative
+status is forwarded to the client verbatim. A richer errno vocabulary is
+future work — clients must treat *any* negative status as failure, not match
+on -1.
+
+## Operations
+
+Values are append-only and never renumbered (the same evolution rule every
+danos protocol follows); an unrecognised operation gets a `status = -1`
+reply.
+
+| value | operation | request payload | reply |
+|------:|-----------|-----------------|-------|
+| 0 | `open` | the path (`len` = its length), `flags` as below | `node` = open-node id |
+| 1 | `close` | — (`node` set) | status only |
+| 2 | `read` | — (`node`, `offset`, `len` = wanted count) | `len` bytes read, payload = the bytes; `len` 0 at end of file |
+| 3 | `write` | the bytes (`node`, `offset`, `len` = count) | `len` = bytes accepted (may be short — loop) |
+| 4 | `status` | — (`node` set) | payload = **FileStatus** (24 bytes) |
+| 5 | `readdir` | — (`node` = a directory, `offset` = cursor) | payload = one **DirectoryEntry** + name; `len` 0 at end |
+| 6 | `mount` | the mount-point path; the backend endpoint rides as the call's **capability** | status only |
+| 7 | `unmount` | the mount-point path | status only |
+| 8 | `mkdir` | the path | status only |
+| 9 | `unlink` | the path | status only |
+| 10 | `rename` | old path, one `0x00`, new path (`len` = total) | status only |
+
+Notes per operation:
+
+- **open** — paths are absolute (`/mnt/usb/notes.txt`) or bare names
+  (`greeting`); bare names resolve in the VFS's flat ramfs, absolute paths
+  route through the mount table (below). The returned `node` is an id in the
+  *router's* open table; clients never see a backend's own ids.
+- **read / write** — a single exchange moves at most 224 bytes
+  (`maximum_payload`); the client loops, advancing `offset` by the returned
+  `len`, until done (read) or the slice is written (write). A `write` reply
+  shorter than requested is progress, not an error; a `len` of 0 means no
+  forward progress — stop rather than spin.
+- **readdir** — `offset` is a **cursor: the entry index**, not a byte
+  position. Each call returns exactly one entry; the client increments the
+  cursor by 1. A reply with `len` 0 is end-of-directory. The directory must
+  have been opened with the `directory` flag.
+- **mount** — the one operation that passes a **capability**: the caller
+  (a filesystem server, e.g. FAT) sends its own request endpoint as the
+  `ipc_call` capability argument, and the router forwards everything under
+  the mount point to it — speaking this same protocol, with paths rewritten
+  relative to the mount. Prefixes match at path boundaries only
+  (`/mnt/usb` never captures `/mnt/usbextra`); the longest matching prefix
+  wins.
+- **rename** — same-directory rename only (the router requires old and new to
+  resolve under one mount).
+
+## Open flags
+
+Bitwise OR in `Request.flags`, meaningful for `open` only:
+
+| bit | name | meaning |
+|----:|------|---------|
+| 1 | `create` | create the file if it does not exist |
+| 2 | `directory` | open a directory node for `readdir` rather than a file |
+| 4 | `truncate` | truncate an existing file to zero length on open (replace, don't overwrite in place) |
+
+## FileStatus — 24 bytes (the `status` reply payload)
+
+| offset | size | field | meaning |
+|-------:|-----:|-------|---------|
+| 0 | 8 | `size` | file size in bytes |
+| 8 | 4 | `kind` | a **NodeKind** value |
+| 12 | 4 | — | padding |
+| 16 | 8 | `mtime` | modification time, Unix epoch seconds UTC; 0 if the backend keeps none |
+
+## DirectoryEntry — 16 bytes + name (the `readdir` reply payload)
+
+| offset | size | field | meaning |
+|-------:|-----:|-------|---------|
+| 0 | 4 | `kind` | a **NodeKind** value |
+| 4 | 4 | `name_len` | length of the name that follows |
+| 8 | 8 | `size` | the entry's size in bytes |
+| 16 | `name_len` | name | the entry's name, not NUL-terminated |
+
+## NodeKind
+
+Aligned to the FSH file-type table
+(docs/danos-file-system-hierarchy-FSH.md):
+
+| value | kind |
+|------:|------|
+| 0 | regular file |
+| 1 | directory |
+| 2 | character device |
+| 3 | block device |
+| 4 | symbolic link |
+| 5 | fifo |
+| 6 | socket |
+
+Clients should map unknown values to *regular* rather than reject — the
+table can grow.
+
+## Lifetimes and trust
+
+Open-node ids live in the server. A client that dies without closing leaks
+nothing permanently: the VFS subscribes to the kernel's published process-exit
+events (docs/process-lifecycle.md) and releases a dead client's handles,
+closing forwarded backend nodes best-effort. Ids are plain integers, not
+capabilities — the VFS trusts its callers with each other's ids today, which
+is acceptable while every client is part of the system image and worth
+revisiting (per-client id namespaces) before third-party binaries arrive.
+
+## Evolution rules
+
+What a non-Zig implementation may rely on, and what it must not:
+
+- Operation values, flag bits, `NodeKind` values, and struct layouts are
+  **append-only and frozen once shipped** — the unit tests in `protocol.zig`
+  pin them exactly so a refactor can't silently move them.
+- The 256-byte message ceiling is a property of the current IPC transport,
+  not a promise; clients should read `maximum_payload`-shaped limits from the
+  reply lengths they actually get (loop-until-done), not hard-code 224.
+- Negative statuses beyond -1 will appear (an errno vocabulary); success is
+  exactly 0.

library/runtime/root.zig

@@ -0,0 +1,19 @@
+//! The root module every user binary is compiled through (build.zig,
+//! `addUserBinary`). The program's own file is imported as `program`, and this
+//! shim contributes the declarations Zig resolves from the compilation root —
+//! `main` (dispatched by runtime.start) and the panic handler — and pulls in the
+//! `_start` entry shim. A program therefore only defines `pub fn main`; nothing
+//! else is required in its source file.
+
+const runtime = @import("runtime");
+const program = @import("program");
+
+/// Resolved as `@import("root").main` by runtime.start's comptime dispatch.
+pub const main = program.main;
+
+/// The panic handler for every safety check in the image (runtime.start.panic).
+pub const panic = runtime.panic;
+
+comptime {
+    _ = &runtime.start._start; // pull the runtime entry shim into the image
+}

library/runtime/runtime.zig

@@ -4,12 +4,11 @@
 //! shim. It is compiled into each binary (inheriting its `.large` code model and
 //! freestanding target), so all user programs share one implementation.
 //!
-//! A user binary needs three lines:
+//! A user binary only defines a `pub fn main() void` or
-//!   const runtime = @import("runtime");
+//! `pub fn main(init: runtime.process.Init) void` (arguments arrive via `init`).
-//!   pub const panic = runtime.panic;
+//! The panic handler and the `_start` entry pull live in the shared compilation
-//!   comptime { _ = &runtime.start._start; }   // pull the entry shim in
+//! root, library/runtime/root.zig, which build.zig wires around every program —
-//! and a `pub fn main() void` or `pub fn main(init: runtime.process.Init) void`
+//! nothing to declare per source file.
-//! (arguments arrive via `init`).
 
 pub const system = @import("system.zig");
 /// Monotonic time, delays, and deadlines over the kernel clock/sleep/timer syscalls
@@ -65,12 +64,16 @@ pub const scanout_protocol = @import("scanout-protocol");
 /// `std.os.danos` are staged. See docs/zig-self-hosting.md.
 pub const fs = @import("fs.zig");
 
-/// Re-exported so a user binary can `pub const panic = runtime.panic;`.
+/// Re-exported so the root shim (root.zig) can install it as the panic handler.
 pub const panic = start.panic;
 
 /// Process entry types: the `Init` handed to `main`, and its `Arguments`.
 pub const process = @import("process.zig");
 
+/// Threads: `runtime.Thread`, std.Thread-shaped, over the private thread ABI
+/// (docs/threading.md). A binary must be built multi-threaded to spawn.
+pub const Thread = @import("thread.zig").Thread;
+
 /// The service harness: one replyWait loop folding requests, signals, and
 /// notifications into callbacks (docs/process-lifecycle.md).
 pub const service = @import("service.zig");

library/runtime/start.zig

@@ -1,6 +1,7 @@
 //! The user-space process entry shim. Every user binary roots `_start` here (via
-//! `entry = _start` in build.zig) and forces this file to be analysed with
+//! `entry = _start` in build.zig); the shared compilation root, root.zig, forces
-//! `comptime { _ = &runtime.start._start; }`, so the whole runtime is linked in.
+//! this file to be analysed with `comptime { _ = &runtime.start._start; }`, so
+//! the whole runtime is linked in.
 
 const std = @import("std");
 const system = @import("system.zig");
@@ -36,7 +37,7 @@ export fn rt_start(stack: [*]const u64) callconv(.c) noreturn {
 /// Comptime-dispatch on root.main's signature, in the spirit of std's start.zig:
 /// zero parameters or one `process.Init`; returns void, noreturn, u8, !void, or !u8.
 fn callMain(init: process.Init) u8 {
-    const root = @import("root"); // the user binary's root source file
+    const root = @import("root"); // root.zig, re-exporting the program's main
     const main_information = @typeInfo(@TypeOf(root.main)).@"fn";
 
     const call_arguments = switch (main_information.params.len) {

library/runtime/thread.zig

@@ -0,0 +1,260 @@
+//! `runtime.Thread` — threads for danos, shaped like Zig's `std.Thread` but built on
+//! danos's private thread ABI (docs/threading.md). Several tasks share one address
+//! space; `spawn` starts one, the kernel delivers the closure pointer in the new
+//! thread's rdi, a plain Zig trampoline runs the user function and calls `thread_exit`,
+//! and `join` blocks on the thread's exit notification. See docs/threading.md for why
+//! this mirrors `std.Thread`'s API rather than being the literal type.
+//!
+//! The closure (the function's captured args) lives at the **top of the thread's own
+//! stack**, not the heap — each thread's stack is private, so there is no shared-heap
+//! concurrency in the spawn/join machinery (the runtime heap is not yet thread-safe).
+//! A binary must be built multi-threaded (`addThreadedUserBinary`) before it may spawn.
+
+const std = @import("std");
+const abi = @import("abi");
+const sc = @import("system-call.zig");
+const system = @import("system.zig");
+const ipc = @import("ipc.zig");
+
+/// A thread stack, if the caller does not override it. 64 KiB of mmap'd, zeroed pages.
+pub const default_stack_size: usize = 64 * 1024;
+
+pub const Thread = struct {
+    /// The kernel task id of the spawned thread.
+    tid: u32,
+    /// The endpoint the kernel notifies when this thread ends — what `join` blocks on.
+    exit_endpoint: ipc.Handle,
+    /// The mmap'd stack, reclaimed by `join` (or at process exit after `detach`).
+    stack_base: usize,
+    stack_size: usize,
+
+    pub const Id = u32;
+
+    pub const SpawnConfig = struct {
+        /// Bytes of stack, rounded up to whole pages by the kernel's mmap.
+        stack_size: usize = default_stack_size,
+    };
+
+    pub const SpawnError = error{
+        /// The kernel refused the thread, the stack mmap failed, or no endpoint was free.
+        SystemResources,
+    };
+
+    /// Start `function(args...)` on a new thread sharing this address space. Mirrors
+    /// `std.Thread.spawn`. The thread's return value is discarded (as in `std.Thread`);
+    /// return data through shared state.
+    pub fn spawn(config: SpawnConfig, comptime function: anytype, args: anytype) SpawnError!Thread {
+        const Args = @TypeOf(args);
+        const Closure = struct {
+            args: Args,
+            /// Entered directly by the kernel with `self` in rdi (C ABI). Runs the user
+            /// function, then ends the thread — never returns.
+            fn entry(self_addr: usize) callconv(.c) noreturn {
+                const self: *@This() = @ptrFromInt(self_addr);
+                @call(.auto, function, self.args);
+                exitThread();
+            }
+        };
+
+        // The endpoint the kernel posts this thread's exit notification to.
+        const endpoint = ipc.createIpcEndpoint() orelse return error.SystemResources;
+
+        const base = system.mmap(config.stack_size, system.PROT_READ | system.PROT_WRITE);
+        if (system.mmapFailed(base)) return error.SystemResources;
+
+        // Lay the closure at the very top of the thread's own stack, then start the
+        // thread's rsp just below it (16-aligned minus 8, the alignment a `call` leaves
+        // for a C-ABI entry) so the growing stack never overwrites the args.
+        var closure_addr = (base + config.stack_size) - @sizeOf(Closure);
+        closure_addr &= ~@as(usize, @alignOf(Closure) - 1); // align the closure down
+        const closure: *Closure = @ptrFromInt(closure_addr);
+        closure.* = .{ .args = args };
+
+        var stack_top = closure_addr & ~@as(usize, 15); // 16-align below the closure
+        stack_top -= 8; // ...then rsp % 16 == 8 at the C entry
+
+        const tid = threadSpawn(@intFromPtr(&Closure.entry), stack_top, closure_addr, endpoint);
+        if (threadSpawnFailed(tid)) {
+            _ = system.munmap(base, config.stack_size);
+            return error.SystemResources;
+        }
+        return .{ .tid = @intCast(tid), .exit_endpoint = endpoint, .stack_base = base, .stack_size = config.stack_size };
+    }
+
+    /// Block until this thread finishes, then reclaim its stack. Mirrors
+    /// `std.Thread.join`. The exit endpoint is private to this thread, so the first
+    /// child-exit notification on it is this thread's.
+    pub fn join(self: Thread) void {
+        var receive: [0]u8 = undefined;
+        while (true) {
+            const got = ipc.replyWait(self.exit_endpoint, &.{}, &receive, null);
+            if (got.isChildExit() and got.childProcessId() == self.tid) break;
+        }
+        _ = system.munmap(self.stack_base, self.stack_size);
+    }
+
+    /// Relinquish the right to join: never wait for or reclaim this thread. Its stack is
+    /// reclaimed at process exit (docs/threading-plan.md M3 — kernel-reaper stack reclaim
+    /// for detached threads is a later refinement). Mirrors `std.Thread.detach`.
+    pub fn detach(self: Thread) void {
+        _ = self;
+    }
+
+    /// The calling thread's id (its kernel task id). Mirrors `std.Thread.getCurrentId`.
+    pub fn getCurrentId() Id {
+        return @intCast(sc.systemCall0(.thread_self));
+    }
+
+    /// The dense 0-based index of the core the calling thread is running on. A danos
+    /// extension beyond `std.Thread`, used to observe genuine cross-core parallelism.
+    pub fn currentCore() Id {
+        return @intCast(sc.systemCall0(.current_core));
+    }
+
+    /// `std.Thread.Futex`-shaped block/wake on a `u32` atomic — the primitive the
+    /// blocking `Mutex`/`Condition`/`Semaphore` are built on. Waiters park in the
+    /// kernel (no busy-wait), so an idle core still halts (docs/halting.md).
+    pub const Futex = struct {
+        /// Block while `ptr.* == expect`. Returns when woken by `wake`, or promptly if
+        /// the value already differs (safe against spurious returns, as in std): the
+        /// caller re-checks its condition in a loop.
+        pub fn wait(ptr: *const std.atomic.Value(u32), expect: u32) void {
+            _ = futexWait(@intFromPtr(ptr), expect, 0);
+        }
+
+        /// As `wait`, but returns `error.Timeout` if `timeout_ns` elapses first.
+        pub fn timedWait(ptr: *const std.atomic.Value(u32), expect: u32, timeout_ns: u64) error{Timeout}!void {
+            if (futexWait(@intFromPtr(ptr), expect, timeout_ns) == abi.futex_timed_out) return error.Timeout;
+        }
+
+        /// Wake up to `max_waiters` threads blocked on `ptr`.
+        pub fn wake(ptr: *const std.atomic.Value(u32), max_waiters: u32) void {
+            _ = futexWake(@intFromPtr(ptr), max_waiters);
+        }
+    };
+
+    /// A mutual-exclusion lock, `std.Thread.Mutex`-shaped. The classic three-state
+    /// futex mutex (unlocked / locked / contended): the fast path is a single CAS, and
+    /// only a contended lock ever enters the kernel.
+    pub const Mutex = struct {
+        state: std.atomic.Value(u32) = std.atomic.Value(u32).init(unlocked),
+
+        const unlocked: u32 = 0;
+        const locked: u32 = 1;
+        const contended: u32 = 2;
+
+        /// Try to take the lock without blocking; returns whether it was acquired.
+        pub fn tryLock(m: *Mutex) bool {
+            return m.state.cmpxchgStrong(unlocked, locked, .acquire, .monotonic) == null;
+        }
+
+        /// Acquire the lock, blocking in the kernel while it is contended.
+        pub fn lock(m: *Mutex) void {
+            if (m.state.cmpxchgStrong(unlocked, locked, .acquire, .monotonic) != null) m.lockSlow();
+        }
+
+        fn lockSlow(m: *Mutex) void {
+            @branchHint(.cold);
+            // Mark the lock contended and take it as soon as it falls unlocked; park on
+            // the futex while it stays contended. Marking contended may cause a spurious
+            // wake on unlock (harmless), never a missed one.
+            while (m.state.swap(contended, .acquire) != unlocked) {
+                Futex.wait(&m.state, contended);
+            }
+        }
+
+        /// Release the lock; wake one waiter if the lock was contended.
+        pub fn unlock(m: *Mutex) void {
+            if (m.state.swap(unlocked, .release) == contended) Futex.wake(&m.state, 1);
+        }
+    };
+
+    /// A condition variable, `std.Thread.Condition`-shaped. Spurious wakeups are
+    /// allowed — always wait in a predicate loop with the mutex held. Built on a futex
+    /// sequence counter: a waiter samples the seq, drops the mutex, and parks until the
+    /// seq changes (a signal that races the unlock bumps the seq, so it is not missed).
+    pub const Condition = struct {
+        seq: std.atomic.Value(u32) = std.atomic.Value(u32).init(0),
+
+        /// Atomically release `mutex` and block until signalled, then re-acquire it.
+        pub fn wait(c: *Condition, mutex: *Mutex) void {
+            const seq = c.seq.load(.acquire);
+            mutex.unlock();
+            Futex.wait(&c.seq, seq);
+            mutex.lock();
+        }
+
+        /// As `wait`, but returns `error.Timeout` if `timeout_ns` elapses first. The
+        /// mutex is re-acquired either way.
+        pub fn timedWait(c: *Condition, mutex: *Mutex, timeout_ns: u64) error{Timeout}!void {
+            const seq = c.seq.load(.acquire);
+            mutex.unlock();
+            const timed_out = if (Futex.timedWait(&c.seq, seq, timeout_ns)) |_| false else |_| true;
+            mutex.lock();
+            if (timed_out) return error.Timeout;
+        }
+
+        /// Wake one waiter.
+        pub fn signal(c: *Condition) void {
+            _ = c.seq.fetchAdd(1, .release);
+            Futex.wake(&c.seq, 1);
+        }
+
+        /// Wake all waiters.
+        pub fn broadcast(c: *Condition) void {
+            _ = c.seq.fetchAdd(1, .release);
+            Futex.wake(&c.seq, std.math.maxInt(u32));
+        }
+    };
+
+    /// A counting semaphore, `std.Thread.Semaphore`-shaped: a permit count guarded by a
+    /// `Mutex` + `Condition`.
+    pub const Semaphore = struct {
+        mutex: Mutex = .{},
+        cond: Condition = .{},
+        permits: usize = 0,
+
+        /// Take a permit, blocking until one is available.
+        pub fn wait(s: *Semaphore) void {
+            s.mutex.lock();
+            defer s.mutex.unlock();
+            while (s.permits == 0) s.cond.wait(&s.mutex);
+            s.permits -= 1;
+        }
+
+        /// Return a permit and wake a waiter.
+        pub fn post(s: *Semaphore) void {
+            s.mutex.lock();
+            defer s.mutex.unlock();
+            s.permits += 1;
+            s.cond.signal();
+        }
+    };
+};
+
+/// thread_spawn(entry, stack_top, arg, exit_endpoint) -> tid, or a wrapped error.
+fn threadSpawn(entry: usize, stack_top: usize, arg: usize, exit_endpoint: ipc.Handle) usize {
+    return sc.systemCall4(.thread_spawn, entry, stack_top, arg, exit_endpoint);
+}
+
+/// The kernel returns a real (small) task id on success and a wrapped `-1` on failure;
+/// no valid task id ever exceeds a u32.
+inline fn threadSpawnFailed(ret: usize) bool {
+    return ret > std.math.maxInt(u32);
+}
+
+/// End the calling thread. Never returns.
+fn exitThread() noreturn {
+    _ = sc.systemCall0(.thread_exit);
+    unreachable;
+}
+
+/// futex_wait(addr, expect, timeout_ns) -> status (abi.futex_*).
+fn futexWait(addr: usize, expect: u32, timeout_ns: u64) usize {
+    return sc.systemCall3(.futex_wait, addr, expect, timeout_ns);
+}
+
+/// futex_wake(addr, count) -> number woken.
+fn futexWake(addr: usize, count: u32) usize {
+    return sc.systemCall2(.futex_wake, addr, count);
+}

system/abi.zig

@@ -63,9 +63,20 @@ pub const SystemCall = enum(u64) {
     shm_create = 34, // shm_create(len) -> vaddr (rax), handle (rdx): a shareable, zeroed, cacheable RAM region mapped into this AS; the handle is a capability passed to another process as an ipc_call send_cap (docs/display-v2.md)
     shm_map = 35, // shm_map(cap) -> vaddr: map the shared region named by a received capability into this AS (the same physical pages the creator sees)
     shm_physical = 36, // shm_physical(cap) -> paddr: the guest-physical base of a shared region held by capability, so a driver can program it into a device (e.g. virtio-gpu attach_backing); the pages are contiguous (docs/display-v2.md)
+    thread_spawn = 37, // thread_spawn(entry, stack_top, arg, exit_endpoint) -> tid: start a task sharing the caller's address space at `entry` on `stack_top`, `arg` in rdi; exit_endpoint (a handle, or no_cap) is notified when it ends — how join waits (docs/threading.md)
+    thread_exit = 38, // thread_exit(): end the calling thread, dropping one reference to its address space (destroyed on the last)
+    current_core = 39, // current_core() -> index: the dense 0-based index of the core the caller is running on (for parallelism/affinity introspection)
+    futex_wait = 40, // futex_wait(addr, expected, timeout_ns) -> status: if *addr == expected, block until woken or the timeout; returns futex_woken/mismatch/timed_out (docs/threading.md)
+    futex_wake = 41, // futex_wake(addr, count) -> woken: wake up to `count` tasks blocked in futex_wait on `addr` in this address space
+    thread_self = 42, // thread_self() -> tid: the calling thread's kernel task id (runtime.Thread.getCurrentId)
     _,
 };
 
+/// `futex_wait` return codes (in rax).
+pub const futex_woken: u64 = 0; // woken by a futex_wake
+pub const futex_mismatch: u64 = 1; // *addr != expected on entry; the caller did not block
+pub const futex_timed_out: u64 = 2; // the timeout elapsed before a wake
+
 /// How a process ended — recorded by the kernel at death, queried by the
 /// supervisor with `process_exit_reason`, and the input to its restart decision
 /// (docs/process-lifecycle.md): a clean exit meant to stop, a fault wants a

system/drivers/pci-bus/pci-bus.zig

@@ -263,8 +263,3 @@ pub fn main(init: runtime.process.Init) void {
         .on_message = onMessage,
     });
 }
-
-pub const panic = runtime.panic;
-comptime {
-    _ = &runtime.start._start; // pull the runtime entry shim into the image
-}

system/drivers/ps2-bus/keyboard.zig

@@ -177,8 +177,3 @@ pub fn main(init: runtime.process.Init) void {
         }
     }
 }
-
-pub const panic = runtime.panic;
-comptime {
-    _ = &runtime.start._start;
-}

system/drivers/ps2-bus/mouse.zig

@@ -60,7 +60,7 @@ pub fn main(init: runtime.process.Init) void {
         _ = runtime.system.write("/system/drivers/ps2-bus/mouse: out of memory\n");
         return;
     };
-    if (device.findDeviceDescriptorByHid(buffer, hid) == null) {
+    if (ps2.findMouseDescriptor(buffer) == null) {
         writeLine("/system/drivers/ps2-bus/mouse: no device for hid {s}\n", .{hid});
         return;
     }
@@ -137,8 +137,3 @@ pub fn main(init: runtime.process.Init) void {
         }
     }
 }
-
-pub const panic = runtime.panic;
-comptime {
-    _ = &runtime.start._start;
-}

system/drivers/ps2-bus/ps2-bus.zig

@@ -244,7 +244,7 @@ pub fn main() void {
     // to the *port*, whatever device identify found on it.
     var maybe_auxiliary_interrupt: ?struct { device_id: u64, interrupt_index: u64, gsi: u64 } = null;
     if (port_device_types[@intFromEnum(ps2.Port.two)] != null) {
-        if (device.findDeviceDescriptorByHid(buffer, acpi_ids.HardwareId.ps2_mouse.hid())) |descriptor| {
+        if (ps2.findMouseDescriptor(buffer)) |descriptor| {
             if (findInterruptResourceIndex(descriptor)) |auxiliary_index| {
                 if (device.claim(descriptor.id) and device.irqBind(descriptor.id, auxiliary_index, endpoint)) {
                     maybe_auxiliary_interrupt = .{
@@ -308,8 +308,3 @@ pub fn main() void {
         reply_len = handleAttach(receive[0..got.len], got, &reply_buffer);
     }
 }
-
-pub const panic = runtime.panic;
-comptime {
-    _ = &runtime.start._start;
-}

system/drivers/ps2-bus/ps2-library.zig

@@ -270,6 +270,22 @@ pub const DeviceType = enum(u32) {
     }
 };
 
+/// The `_HID`s a PS/2 pointing device (the controller's aux channel) can enumerate
+/// under. It is the same 8042 mouse channel whichever id the firmware chose:
+/// QEMU/OVMF report the generic `.ps2_mouse` (PNP0F13), VirtualBox reports
+/// `.microsoft_ps2_mouse` (PNP0F03). Both mean "the mouse on port two".
+pub const mouse_hardware_ids = [_]acpi_ids.HardwareId{ .ps2_mouse, .microsoft_ps2_mouse };
+
+/// Find the aux (mouse) device's ACPI node, whichever of the PS/2-mouse `_HID`s the
+/// firmware used — the bus needs it to bind IRQ12, and the mouse driver to confirm
+/// its device is present. Returns the first match, or null if none is reported.
+pub fn findMouseDescriptor(buffer: []device.DeviceDescriptor) ?device.DeviceDescriptor {
+    for (mouse_hardware_ids) |id| {
+        if (device.findDeviceDescriptorByHid(buffer, id.hid())) |descriptor| return descriptor;
+    }
+    return null;
+}
+
 // --- the bus <-> child-driver forwarding protocol -----------------------------
 //
 // The 8042's ports and IRQ1 live on the PNP0303 node that only the ps2-bus driver

system/drivers/usb-hid/keyboard.zig

@@ -167,8 +167,3 @@ pub fn main(init: runtime.process.Init) void {
         }
     }
 }
-
-pub const panic = runtime.panic;
-comptime {
-    _ = &runtime.start._start;
-}

system/drivers/usb-hid/mouse.zig

@@ -133,8 +133,3 @@ pub fn main(init: runtime.process.Init) void {
         }
     }
 }
-
-pub const panic = runtime.panic;
-comptime {
-    _ = &runtime.start._start;
-}

system/drivers/usb-storage/usb-storage.zig

@@ -180,8 +180,3 @@ pub fn main(init: runtime.process.Init) void {
         .on_message = onMessage,
     });
 }
-
-pub const panic = runtime.panic;
-comptime {
-    _ = &runtime.start._start;
-}

system/drivers/usb-xhci-bus/usb-xhci-bus.zig

@@ -417,8 +417,3 @@ pub fn main(init: runtime.process.Init) void {
         .on_notification = onNotification,
     });
 }
-
-pub const panic = runtime.panic;
-comptime {
-    _ = &runtime.start._start; // pull the runtime entry shim into the image
-}

system/drivers/virtio-gpu/virtio-gpu.zig

@@ -615,8 +615,3 @@ pub fn main(init: runtime.process.Init) void {
         .on_message = onMessage,
     });
 }
-
-pub const panic = runtime.panic;
-comptime {
-    _ = &runtime.start._start; // pull the runtime entry shim into the image
-}

system/kernel/architecture/x86_64/cpu.zig

@@ -687,6 +687,15 @@ pub fn jumpToUser(entry: u64, stack_top: u64) noreturn {
     jump_to_user(entry, stack_top);
 }
 
+/// As `jumpToUser`, but delivers `arg0` in the user's `rdi` — how a fresh thread
+/// receives its closure pointer (docs/threading.md). A normal process is dropped
+/// with `arg0 = 0`, which its `_start` ignores (it reads argv off the stack).
+extern fn jump_to_user_arg(rip: u64, rsp: u64, arg0: u64) callconv(.c) noreturn;
+
+pub fn jumpToUserArg(entry: u64, stack_top: u64, arg0: u64) noreturn {
+    jump_to_user_arg(entry, stack_top, arg0);
+}
+
 /// Route CPU exceptions to `handler`, which receives the trap frame and does not
 /// return. Until set, faults just halt the core.
 pub fn setFaultHandler(handler: *const fn (*const CpuState) noreturn) void {

system/kernel/architecture/x86_64/isr.s

@@ -123,6 +123,22 @@ jump_to_user:
     swapgs                 # user GS base (isr_common/syscall swap back on entry)
     iretq
 
+# jump_to_user_arg(rdi = user rip, rsi = user rsp, rdx = user rdi/arg0): as
+# jump_to_user, but delivers arg0 in the user's rdi — how a fresh **thread**
+# receives its closure pointer (docs/threading.md). rdi carries the rip only until
+# it is pushed into the iretq frame, after which we overwrite it with the arg.
+.global jump_to_user_arg
+jump_to_user_arg:
+    cli
+    push $0x1B              # user SS  (0x18 | RPL 3)
+    push %rsi              # user RSP
+    push $0x202            # RFLAGS: IF | reserved-1
+    push $0x23             # user CS  (0x20 | RPL 3)
+    push %rdi             # user RIP  (consumes rdi)
+    mov %rdx, %rdi        # user rdi = arg0 (the thread's closure pointer)
+    swapgs                 # user GS base (isr_common/syscall swap back on entry)
+    iretq
+
 # --- ring 3 entry/exit ------------------------------------------------------
 
 # enter_user(rdi = user rip, rsi = user rsp, rdx = &TSS.rsp0)

system/kernel/process.zig

@@ -227,6 +227,20 @@ fn system_call(state: *architecture.CpuState) void {
         .shm_create => systemShmCreate(state),
         .shm_map => systemShmMap(state),
         .shm_physical => systemShmPhysical(state),
+        .thread_spawn => systemThreadSpawn(state),
+        .current_core => systemCurrentCore(state),
+        .thread_self => systemThreadSelf(state),
+        .futex_wait => systemFutexWait(state),
+        .futex_wake => systemFutexWake(state),
+        .thread_exit => {
+            // A thread ends like a process exit(0), but only this task: its
+            // resources are released and its address-space reference dropped (the
+            // space survives while sibling threads hold it). docs/threading.md.
+            if (scheduler.currentIsUserProcess()) {
+                scheduler.current().exit_reason = .exited;
+                terminateCurrent();
+            } else architecture.userExit();
+        },
         _ => fail(state),
     }
 }
@@ -642,6 +656,97 @@ fn systemSpawn(state: *architecture.CpuState) void {
     fail(state); // no bundled binary by that name
 }
 
+/// thread_spawn(entry, stack_top, arg) -> tid: start a task that shares the **caller's**
+/// address space (docs/threading.md). The runtime supplies `entry` (its thread
+/// trampoline), a stack it mmap'd, and the closure pointer, which the kernel delivers in
+/// the new thread's rdi. The entry and stack must lie in the user half; the new thread is
+/// supervised by the caller and inherits its priority. Only a user process may spawn.
+fn systemThreadSpawn(state: *architecture.CpuState) void {
+    const entry = architecture.systemCallArg(state, 0);
+    const stack_top = architecture.systemCallArg(state, 1);
+    const arg = architecture.systemCallArg(state, 2);
+    const exit_handle = architecture.systemCallArg(state, 3);
+    const t = scheduler.current();
+    if (t.aspace == 0) return fail(state); // kernel tasks own no address space to share
+    if (entry == 0 or entry >= user_half_end) return fail(state);
+    if (stack_top == 0 or stack_top > user_half_end) return fail(state);
+    // The endpoint the thread notifies on exit (how join waits), or none.
+    const exit_endpoint: ?*ipc.Endpoint = if (exit_handle == abi.no_cap)
+        null
+    else
+        ipc.resolveHandle(t, exit_handle) orelse return failErr(state, ipc.EBADF);
+    const tid = spawnThreadSupervised(t.aspace, entry, stack_top, arg, t.priority, t.id, exit_endpoint) orelse return fail(state);
+    architecture.setSystemCallResult(state, tid);
+}
+
+/// Spawn a thread sharing `aspace`, taking the exit-endpoint reference under the **same**
+/// lock as the spawn (as `spawnProcessSupervised` does), so the thread cannot die before
+/// its reference exists. Returns the new thread id, or null on resource exhaustion.
+fn spawnThreadSupervised(aspace: u64, entry: u64, stack_top: u64, arg: u64, priority: scheduler.Priority, supervisor: u32, exit_endpoint: ?*ipc.Endpoint) ?u32 {
+    const flags = sync.enter();
+    defer sync.leave(flags);
+    const tid = scheduler.spawnUserLocked(aspace, entry, stack_top, arg, priority, "thread", supervisor, if (exit_endpoint) |e| @ptrCast(e) else null) orelse return null;
+    if (exit_endpoint) |endpoint| endpoint.refcount += 1; // the thread holds it birth-to-death
+    return tid;
+}
+
+/// current_core() -> index: the dense 0-based index of the core the caller runs on.
+fn systemCurrentCore(state: *architecture.CpuState) void {
+    architecture.setSystemCallResult(state, scheduler.currentCpuIndex());
+}
+
+/// thread_self() -> tid: the calling thread's kernel task id.
+fn systemThreadSelf(state: *architecture.CpuState) void {
+    architecture.setSystemCallResult(state, scheduler.currentId());
+}
+
+/// futex_wait(addr, expected, timeout_ns) -> status (docs/threading.md): if the 4-byte
+/// user word at `addr` still equals `expected`, block until a futex_wake on `addr` or
+/// (if timeout_ns > 0) the deadline. The compare and the block are one critical section,
+/// so a concurrent futex_wake cannot slip between them. Returns futex_woken / mismatch /
+/// timed_out.
+fn systemFutexWait(state: *architecture.CpuState) void {
+    const addr = architecture.systemCallArg(state, 0);
+    const expected: u32 = @truncate(architecture.systemCallArg(state, 1));
+    const timeout_ns = architecture.systemCallArg(state, 2);
+    const t = scheduler.current();
+    if (t.aspace == 0) return fail(state);
+    if (addr == 0 or (addr & 3) != 0 or addr + 4 > user_half_end) return fail(state);
+
+    const flags = sync.enter();
+    var word_bytes: [4]u8 = undefined;
+    if (!ipc.copyFromUser(t.aspace, addr, &word_bytes)) {
+        sync.leave(flags);
+        return fail(state);
+    }
+    if (std.mem.readInt(u32, &word_bytes, .little) != expected) {
+        sync.leave(flags);
+        architecture.setSystemCallResult(state, abi.futex_mismatch);
+        return;
+    }
+    const timeout_ms = if (timeout_ns == 0) 0 else (timeout_ns + 999_999) / 1_000_000;
+    const result = scheduler.futexWaitLocked(addr, timeout_ms);
+    sync.leave(flags);
+    architecture.setSystemCallResult(state, switch (result) {
+        .woken => abi.futex_woken,
+        .timed_out => abi.futex_timed_out,
+    });
+}
+
+/// futex_wake(addr, count) -> woken: wake up to `count` tasks blocked in futex_wait on
+/// `addr` in the caller's address space.
+fn systemFutexWake(state: *architecture.CpuState) void {
+    const addr = architecture.systemCallArg(state, 0);
+    const count: u32 = @truncate(architecture.systemCallArg(state, 1));
+    const t = scheduler.current();
+    if (t.aspace == 0) return fail(state);
+    if (addr == 0 or (addr & 3) != 0 or addr + 4 > user_half_end) return fail(state);
+    const flags = sync.enter();
+    const woken = scheduler.futexWakeLocked(t.aspace, addr, count);
+    sync.leave(flags);
+    architecture.setSystemCallResult(state, woken);
+}
+
 /// process_enumerate(buffer, maximum) -> total: snapshot the task table into the
 /// caller's buffer (up to `maximum` `abi.ProcessDescriptor` entries), returning
 /// the total live-task count — the exact shape of `device_enumerate`, so a `ps`
@@ -1433,7 +1538,7 @@ pub fn spawnProcessSupervised(image: []const u8, priority: u3, argv: []const []c
         architecture.mapUserPageInto(aspace, page_virtual, stack_frame, true, false); // RW + NX
     }
 
-    const child = scheduler.spawnUserLocked(aspace, parsed.entry, user_sp, priority, argv[0], supervisor, if (exit_endpoint) |endpoint| @ptrCast(endpoint) else null) orelse
+    const child = scheduler.spawnUserLocked(aspace, parsed.entry, user_sp, 0, priority, argv[0], supervisor, if (exit_endpoint) |endpoint| @ptrCast(endpoint) else null) orelse
         return error.OutOfMemory;
     // The child holds a reference to its exit endpoint from birth to death. Taken
     // only now, after nothing can fail; the lock is still held, so the child

system/kernel/scheduler.zig

@@ -78,6 +78,11 @@ pub const Task = struct {
     aspace: u64 = 0,
     user_ip: u64 = 0, // user-mode entry point (user task only)
     user_sp: u64 = 0, // user-mode stack pointer (user task only)
+    user_arg: u64 = 0, // value delivered in the user's rdi at first entry: 0 for a
+    // process (its _start ignores it), the closure pointer for a thread (docs/threading.md)
+    // The user address this task is blocked on in futex_wait (0 = not futex-waiting).
+    // Cleared to 0 by futexWakeLocked as the "woken, not timed out" signal (docs/threading.md).
+    futex_addr: u64 = 0,
     // Next free virtual address in this task's mmap grant arena (0 = uninitialised;
     // process.zig lazily seeds it to the arena base on the first mmap). Bumped up
     // as the user heap grows; user task only.
@@ -136,6 +141,67 @@ pub const ipc_maximum_handles = 16;
 pub const HandleObject = struct { kind: u8, ptr: *anyopaque };
 
 var tasks = [_]Task{.{}} ** maximum_tasks;
+
+/// Address-space reference counts: one live entry per address space, counting the
+/// tasks that share it. An address space is 1:1 with a process today; threads
+/// (docs/threading.md) will push a count above 1, and `destroyAddressSpace` must run
+/// only when the **last** task on an address space exits. All access is under the big
+/// kernel lock. There can be no more live address spaces than tasks, so the table is
+/// sized to the task pool and never overflows in practice.
+const AspaceRef = struct { root: u64 = 0, count: u32 = 0 };
+var aspace_refs = [_]AspaceRef{.{}} ** maximum_tasks;
+var aspace_destroy_count: u64 = 0;
+
+/// Take a reference to address space `root` (0 = a kernel task, which owns none).
+/// Returns false only if the ref table is full — bounded by `maximum_tasks`, so in
+/// practice it never is. Caller holds the kernel lock.
+fn retainAspace(root: u64) bool {
+    if (root == 0) return true;
+    var free: ?*AspaceRef = null;
+    for (&aspace_refs) |*entry| {
+        if (entry.count != 0 and entry.root == root) {
+            entry.count += 1;
+            return true;
+        }
+        if (entry.count == 0 and free == null) free = entry;
+    }
+    const slot = free orelse return false;
+    slot.* = .{ .root = root, .count = 1 };
+    return true;
+}
+
+/// Drop a reference to `root`; destroy the address space when the **last** one drops.
+/// A `root` with no entry — never retained, e.g. a hand-built test space — is
+/// destroyed directly, preserving the pre-refcount behaviour. Caller holds the lock.
+fn releaseAspace(root: u64) void {
+    if (root == 0) return;
+    for (&aspace_refs) |*entry| {
+        if (entry.count == 0 or entry.root != root) continue;
+        entry.count -= 1;
+        if (entry.count == 0) {
+            entry.root = 0;
+            architecture.destroyAddressSpace(root);
+            aspace_destroy_count += 1;
+        }
+        return;
+    }
+    architecture.destroyAddressSpace(root);
+    aspace_destroy_count += 1;
+}
+
+/// Test-observable: how many address spaces are live (entries with a nonzero count).
+pub fn liveAspaceCount() u32 {
+    var live: u32 = 0;
+    for (&aspace_refs) |*entry| {
+        if (entry.count != 0) live += 1;
+    }
+    return live;
+}
+
+/// Test-observable: total address-space destructions since boot.
+pub fn aspaceDestroyCount() u64 {
+    return aspace_destroy_count;
+}
 var next_id: u32 = 1;
 
 /// Per-CPU scheduler state: the task each core is running, its own idle task, and a
@@ -327,9 +393,15 @@ pub fn spawnOn(entry: *const fn () void, priority: Priority, cpu: u32) bool {
 /// out of memory.
 /// **Caller must hold the kernel lock** (the loader that builds `aspace` holds it
 /// across the whole spawn, so the address space and the task appear atomically).
-pub fn spawnUserLocked(aspace: u64, entry: u64, user_sp: u64, priority: Priority, task_name: []const u8, supervisor: u32, exit_endpoint: ?*anyopaque) ?u32 {
+pub fn spawnUserLocked(aspace: u64, entry: u64, user_sp: u64, user_arg: u64, priority: Priority, task_name: []const u8, supervisor: u32, exit_endpoint: ?*anyopaque) ?u32 {
     const t = freeSlot() orelse return null;
     const stack = heap.allocator().alloc(u8, stack_size) catch return null;
+    // Take this task's reference to the address space before we commit the slot, so a
+    // failure here leaves nothing to unwind (the caller still owns the raw `aspace`).
+    if (!retainAspace(aspace)) {
+        heap.allocator().free(stack);
+        return null;
+    }
     t.* = .{
         .id = next_id,
         .state = .ready,
@@ -338,6 +410,7 @@ pub fn spawnUserLocked(aspace: u64, entry: u64, user_sp: u64, priority: Priority
         .aspace = aspace,
         .user_ip = entry,
         .user_sp = user_sp,
+        .user_arg = user_arg,
         .supervisor = supervisor,
         .exit_endpoint = exit_endpoint,
     };
@@ -363,7 +436,7 @@ fn startUserTask() void {
     // No serial chatter here: this runs on every spawn, unserialized against
     // user-space writes, and its output used to shear concurrent log lines in
     // half — the largest source of corrupted markers in the QEMU scenarios.
-    architecture.jumpToUser(t.user_ip, t.user_sp); // noreturn
+    architecture.jumpToUserArg(t.user_ip, t.user_sp, t.user_arg); // noreturn (arg0 = 0 for a process)
 }
 
 /// The unlocked task-creation primitive. Caller must hold the kernel lock (or be the
@@ -446,6 +519,49 @@ pub fn sleep(ms: u64) void {
     sync.leave(flags);
 }
 
+// --- futex: block/wake on a user address (docs/threading.md) ----------------
+//
+// A futex waiter is not linked into any queue — it is simply a `.blocked` task
+// tagged with the address it waits on (`futex_addr`). Waking scans the task table
+// (bounded) for matching waiters. A timed wait also sets `wake_at`, so the timer's
+// `wakeExpired` can wake it; `futex_addr` stays non-zero in that case, which is how
+// the waiter tells a timeout from a real wake.
+
+pub const FutexResult = enum { woken, timed_out };
+
+/// Block the current task on futex `addr` until woken, or (if `timeout_ms > 0`) the
+/// deadline. **Precondition:** the big kernel lock is held and the caller has already
+/// checked, under this same lock, that the futex word equals the expected value — so
+/// no wake can be missed. Returns with the lock still held.
+pub fn futexWaitLocked(addr: u64, timeout_ms: u64) FutexResult {
+    const t = current();
+    t.futex_addr = addr;
+    t.wake_at = if (timeout_ms > 0) architecture.millis() + timeout_ms else 0;
+    t.state = .blocked;
+    schedule(); // woken by futexWakeLocked (clears futex_addr) or wakeExpired (timeout)
+    const woken = t.futex_addr == 0;
+    t.futex_addr = 0;
+    t.wake_at = 0;
+    return if (woken) .woken else .timed_out;
+}
+
+/// Wake up to `count` tasks blocked in `futex_wait` on `addr` in address space
+/// `aspace`. Precondition: the big kernel lock is held. Returns how many woke.
+pub fn futexWakeLocked(aspace: u64, addr: u64, count: u32) u32 {
+    var woken: u32 = 0;
+    for (&tasks) |*t| {
+        if (woken >= count) break;
+        if (t.state == .blocked and t.aspace == aspace and t.futex_addr == addr) {
+            t.futex_addr = 0; // the "woken, not timed out" signal to futexWaitLocked
+            t.wake_at = 0;
+            t.state = .ready;
+            enqueue(t);
+            woken += 1;
+        }
+    }
+    return woken;
+}
+
 // --- event-based blocking -------------------------------------------------
 //
 // A WaitQueue is a set of tasks blocked waiting for something (a resource, a
@@ -720,7 +836,7 @@ pub fn exitUserLocked() noreturn {
         const kroot = architecture.kernelPageTable();
         architecture.loadPageTable(kroot); // off the process tables before freeing them
         pc.loaded_aspace = kroot;
-        architecture.destroyAddressSpace(as);
+        releaseAspace(as); // destroys only when this was the last task on the space
     }
     dying.state = .free;
     dying.aspace = 0;
@@ -741,7 +857,7 @@ pub fn exitUserLocked() noreturn {
 /// task isn't running). The kernel stack is leaked, as in `exitUser` (no reaper
 /// yet). Precondition: the big kernel lock is held.
 pub fn destroyTaskLocked(t: *Task) void {
-    if (t.aspace != 0) architecture.destroyAddressSpace(t.aspace);
+    if (t.aspace != 0) releaseAspace(t.aspace); // destroys only on the last reference
     t.aspace = 0;
     t.kill_pending = false;
     t.in_system_call = false;

system/kernel/tests.zig

@@ -139,6 +139,18 @@ pub fn run(case: []const u8, boot_information: *const BootInformation) void {
         userPfTest();
     } else if (eql(case, "fault-recovery")) {
         faultRecoveryTest(boot_information);
+    } else if (eql(case, "aspace-refcount")) {
+        aspaceRefcountTest(boot_information);
+    } else if (eql(case, "thread-spawn")) {
+        threadSpawnTest(boot_information);
+    } else if (eql(case, "thread-join")) {
+        threadJoinTest(boot_information);
+    } else if (eql(case, "thread-futex")) {
+        threadFutexTest(boot_information);
+    } else if (eql(case, "thread-mutex")) {
+        threadMutexTest(boot_information);
+    } else if (eql(case, "thread-id")) {
+        threadIdTest(boot_information);
     } else if (eql(case, "args")) {
         argsTest(boot_information);
     } else if (eql(case, "init")) {
@@ -1374,7 +1386,7 @@ fn spawnFaultingProcess() ?u32 {
     architecture.mapUserPageInto(aspace, process.stack_base_virtual, stack_frame, true, false); // RW + NX
 
     // Supervised by the calling test task, so exitReasonOf can read the verdict.
-    const id = scheduler.spawnUserLocked(aspace, process.code_virtual, process.stack_base_virtual + abi.page_size, 4, "fault-probe", scheduler.currentId(), null) orelse {
+    const id = scheduler.spawnUserLocked(aspace, process.code_virtual, process.stack_base_virtual + abi.page_size, 0, 4, "fault-probe", scheduler.currentId(), null) orelse {
         architecture.destroyAddressSpace(aspace);
         return null;
     };
@@ -1429,6 +1441,254 @@ fn faultRecoveryTest(boot_information: *const BootInformation) void {
     result();
 }
 
+/// Address-space refcount (docs/threading-plan.md M1): every process holds exactly one
+/// reference to its address space, released when it dies, so `destroyAddressSpace` runs
+/// exactly once per space — no leak, no double-free. Spawn and kill several ring-3
+/// processes (the faulting probe, reaped by the kernel) and confirm the count of live
+/// address spaces returns to baseline while destructions advance by exactly that many.
+/// This is the foundation threads (shared address spaces) build on: the refactor must be
+/// invisible while every space still has exactly one task.
+fn aspaceRefcountTest(boot_information: *const BootInformation) void {
+    _ = boot_information;
+    log("DANOS-TEST-BEGIN: aspace-refcount\n", .{});
+    const base_live = scheduler.liveAspaceCount();
+    const base_destroyed = scheduler.aspaceDestroyCount();
+    const rounds: u32 = 5;
+    var killed: u32 = 0;
+    var round: u32 = 0;
+    while (round < rounds) : (round += 1) {
+        process.fault_kill_count = 0;
+        const probe = spawnFaultingProcess() orelse break;
+        _ = probe;
+        // Let the probe fault on its first instruction and be reaped.
+        scheduler.setPriority(1);
+        const deadline = architecture.millis() + 5000;
+        while (process.fault_kill_count < 1 and architecture.millis() < deadline) scheduler.yield();
+        scheduler.setPriority(4);
+        if (process.fault_kill_count >= 1) killed += 1;
+    }
+    check("all probes spawned and were killed", killed == rounds);
+    check("live address-space count returned to baseline", scheduler.liveAspaceCount() == base_live);
+    check("each address space destroyed exactly once", scheduler.aspaceDestroyCount() == base_destroyed + rounds);
+    if (killed == rounds and scheduler.liveAspaceCount() == base_live and
+        scheduler.aspaceDestroyCount() == base_destroyed + rounds)
+        log("aspace-refcount: spaces released to baseline ok\n", .{});
+    result();
+}
+
+/// Thread spawn (docs/threading-plan.md M2): the `thread-test` service spawns a worker
+/// thread that writes a shared global; the main thread, polling that memory, observes the
+/// write — proving `runtime.Thread.spawn` started a task in the **same** address space
+/// (a separate process could not touch it). The service's own marker is the verdict.
+fn threadSpawnTest(boot_information: *const BootInformation) void {
+    log("DANOS-TEST-BEGIN: thread-spawn\n", .{});
+    if (boot_information.initial_ramdisk_len == 0) {
+        check("bootloader handed over an initial_ramdisk", false);
+        result();
+        return;
+    }
+    const image = @as([*]const u8, @ptrFromInt(boot_handoff.physicalToVirtual(boot_information.initial_ramdisk_base)))[0..boot_information.initial_ramdisk_len];
+    const rd = initial_ramdisk.Reader.init(image) orelse {
+        check("initial_ramdisk image is valid", false);
+        result();
+        return;
+    };
+
+    check("thread-test spawned", spawnNamed(rd, "thread-test"));
+
+    // Wait for the service's verdict marker (it polls shared memory the worker wrote).
+    const ok_marker = "thread-test: child ran in shared aspace ok";
+    const fail_marker = "thread-test: FAIL";
+    scheduler.setPriority(1);
+    const deadline = architecture.millis() + 12000;
+    while (architecture.millis() < deadline) {
+        if (bufferHas(ok_marker) or bufferHas(fail_marker)) break;
+        scheduler.yield();
+    }
+    scheduler.setPriority(4);
+
+    check("a worker thread ran in the shared address space (shared write observed)", bufferHas(ok_marker));
+    check("the thread path reported no failure", !bufferHas(fail_marker));
+    result();
+}
+
+/// Thread join + parallelism (docs/threading-plan.md M3): `thread-test` in join mode
+/// spawns N workers that each do K atomic increments on a shared counter and stamp the
+/// core they ran on; it `join`s all N and asserts the total is exactly N*K (every worker
+/// ran, join waited for each) and that >1 core was used (genuine parallelism), then a
+/// detached worker proves `detach`. Its single verdict marker is the case result.
+fn threadJoinTest(boot_information: *const BootInformation) void {
+    log("DANOS-TEST-BEGIN: thread-join\n", .{});
+    if (boot_information.initial_ramdisk_len == 0) {
+        check("bootloader handed over an initial_ramdisk", false);
+        result();
+        return;
+    }
+    const image = @as([*]const u8, @ptrFromInt(boot_handoff.physicalToVirtual(boot_information.initial_ramdisk_base)))[0..boot_information.initial_ramdisk_len];
+    const rd = initial_ramdisk.Reader.init(image) orelse {
+        check("initial_ramdisk image is valid", false);
+        result();
+        return;
+    };
+
+    // Spawn thread-test in join mode (argv selects the mode).
+    var started = false;
+    var i: u32 = 0;
+    while (i < rd.count) : (i += 1) {
+        const item = rd.entry(i) orelse continue;
+        if (!eql(item.name, "thread-test")) continue;
+        started = if (process.spawnProcess(item.blob, 4, &.{ "thread-test", "join" })) true else |_| false;
+        break;
+    }
+    check("thread-test (join mode) spawned", started);
+
+    const ok_marker = "thread-test: join ok";
+    const fail_marker = "thread-test: FAIL";
+    scheduler.setPriority(1);
+    const deadline = architecture.millis() + 15000;
+    while (architecture.millis() < deadline) {
+        if (bufferHas(ok_marker) or bufferHas(fail_marker)) break;
+        scheduler.yield();
+    }
+    scheduler.setPriority(4);
+
+    check("N worker threads joined; counter exact (N*K) and >1 core used", bufferHas(ok_marker));
+    check("no thread failure reported", !bufferHas(fail_marker));
+    result();
+}
+
+/// Futex (docs/threading-plan.md M4): `thread-test` in futex mode has a waiter thread
+/// block in `futex_wait` on a word; the main thread publishes the word and `futex_wake`s
+/// it. The serial order `waiting → waking → woke` shows the kernel handoff (the waiter
+/// parked and was woken, not spun), and a `timedWait` on an unwoken word reports a
+/// timeout. The verdict marker is emitted only after both hold.
+fn threadFutexTest(boot_information: *const BootInformation) void {
+    log("DANOS-TEST-BEGIN: thread-futex\n", .{});
+    if (boot_information.initial_ramdisk_len == 0) {
+        check("bootloader handed over an initial_ramdisk", false);
+        result();
+        return;
+    }
+    const image = @as([*]const u8, @ptrFromInt(boot_handoff.physicalToVirtual(boot_information.initial_ramdisk_base)))[0..boot_information.initial_ramdisk_len];
+    const rd = initial_ramdisk.Reader.init(image) orelse {
+        check("initial_ramdisk image is valid", false);
+        result();
+        return;
+    };
+
+    var started = false;
+    var i: u32 = 0;
+    while (i < rd.count) : (i += 1) {
+        const item = rd.entry(i) orelse continue;
+        if (!eql(item.name, "thread-test")) continue;
+        started = if (process.spawnProcess(item.blob, 4, &.{ "thread-test", "futex" })) true else |_| false;
+        break;
+    }
+    check("thread-test (futex mode) spawned", started);
+
+    const ok_marker = "thread-futex: ok";
+    const fail_marker = "thread-futex: FAIL";
+    scheduler.setPriority(1);
+    const deadline = architecture.millis() + 15000;
+    while (architecture.millis() < deadline) {
+        if (bufferHas(ok_marker) or bufferHas(fail_marker)) break;
+        scheduler.yield();
+    }
+    scheduler.setPriority(4);
+
+    // Only the freshest marker is checked here — the kernel's log ring buffer may have
+    // evicted the earlier ones by now. The waiting/waking/woke ordering (the handoff
+    // proof) is asserted against the full serial stream by the qemu case's regex; the
+    // verdict marker is emitted by thread-test only after the wake AND the timeout hold.
+    check("futex handoff + timeout completed (verdict reached, no failure)", bufferHas(ok_marker) and !bufferHas(fail_marker));
+    result();
+}
+
+/// Mutex + Condition (docs/threading-plan.md M5): `thread-test` in mutex mode runs a
+/// bounded producer/consumer — P producers and C consumers over one `Mutex` and two
+/// `Condition`s move N unique items through a small ring. Every item is produced once;
+/// if the lock and condition variables are correct under real cross-core contention,
+/// the consumed checksum and tally match exactly (no lost or duplicated item, no
+/// overrun). The verdict marker is emitted only when both match.
+fn threadMutexTest(boot_information: *const BootInformation) void {
+    log("DANOS-TEST-BEGIN: thread-mutex\n", .{});
+    if (boot_information.initial_ramdisk_len == 0) {
+        check("bootloader handed over an initial_ramdisk", false);
+        result();
+        return;
+    }
+    const image = @as([*]const u8, @ptrFromInt(boot_handoff.physicalToVirtual(boot_information.initial_ramdisk_base)))[0..boot_information.initial_ramdisk_len];
+    const rd = initial_ramdisk.Reader.init(image) orelse {
+        check("initial_ramdisk image is valid", false);
+        result();
+        return;
+    };
+
+    var started = false;
+    var i: u32 = 0;
+    while (i < rd.count) : (i += 1) {
+        const item = rd.entry(i) orelse continue;
+        if (!eql(item.name, "thread-test")) continue;
+        started = if (process.spawnProcess(item.blob, 4, &.{ "thread-test", "mutex" })) true else |_| false;
+        break;
+    }
+    check("thread-test (mutex mode) spawned", started);
+
+    const ok_marker = "thread-mutex: ok";
+    const fail_marker = "thread-mutex: FAIL";
+    scheduler.setPriority(1);
+    const deadline = architecture.millis() + 20000;
+    while (architecture.millis() < deadline) {
+        if (bufferHas(ok_marker) or bufferHas(fail_marker)) break;
+        scheduler.yield();
+    }
+    scheduler.setPriority(4);
+
+    check("producer/consumer over Mutex+Condition moved every item exactly once", bufferHas(ok_marker) and !bufferHas(fail_marker));
+    result();
+}
+
+/// Thread identity (docs/threading-plan.md M6): `thread-test` in id mode spawns two
+/// workers that each read `runtime.Thread.getCurrentId`; the main thread confirms all
+/// three ids are non-zero and distinct — proof each thread has its own kernel identity.
+fn threadIdTest(boot_information: *const BootInformation) void {
+    log("DANOS-TEST-BEGIN: thread-id\n", .{});
+    if (boot_information.initial_ramdisk_len == 0) {
+        check("bootloader handed over an initial_ramdisk", false);
+        result();
+        return;
+    }
+    const image = @as([*]const u8, @ptrFromInt(boot_handoff.physicalToVirtual(boot_information.initial_ramdisk_base)))[0..boot_information.initial_ramdisk_len];
+    const rd = initial_ramdisk.Reader.init(image) orelse {
+        check("initial_ramdisk image is valid", false);
+        result();
+        return;
+    };
+
+    var started = false;
+    var i: u32 = 0;
+    while (i < rd.count) : (i += 1) {
+        const item = rd.entry(i) orelse continue;
+        if (!eql(item.name, "thread-test")) continue;
+        started = if (process.spawnProcess(item.blob, 4, &.{ "thread-test", "id" })) true else |_| false;
+        break;
+    }
+    check("thread-test (id mode) spawned", started);
+
+    const ok_marker = "thread-id: ok";
+    const fail_marker = "thread-id: FAIL";
+    scheduler.setPriority(1);
+    const deadline = architecture.millis() + 12000;
+    while (architecture.millis() < deadline) {
+        if (bufferHas(ok_marker) or bufferHas(fail_marker)) break;
+        scheduler.yield();
+    }
+    scheduler.setPriority(4);
+
+    check("each thread has a distinct, non-zero getCurrentId", bufferHas(ok_marker) and !bufferHas(fail_marker));
+    result();
+}
+
 /// The full PID-1 path: the bootloader read /system/services/init off the boot volume and
 /// handed it over; load it as a user ELF and spawn it as a real ring-3 process
 /// — the same call the normal boot path makes — then confirm it beats. init

system/services/acpi/acpi.zig

@@ -694,8 +694,3 @@ fn rd16(bytes: []const u8, off: usize) u64 {
 fn rd32(bytes: []const u8, off: usize) u64 {
     return rd16(bytes, off) | (rd16(bytes, off + 2) << 16);
 }
-
-pub const panic = runtime.panic;
-comptime {
-    _ = &runtime.start._start; // pull the runtime entry shim into the image
-}

system/services/args-echo/args-echo.zig

@@ -48,8 +48,3 @@ pub fn main(init: runtime.process.Init) void {
     len += 1;
     _ = runtime.system.write(buffer[0..len]);
 }
-
-pub const panic = runtime.panic;
-comptime {
-    _ = &runtime.start._start; // pull the runtime entry shim into the image
-}

system/services/crash-test/crash-test.zig

@@ -37,8 +37,3 @@ pub fn main(init: runtime.process.Init) void {
     const poison: *volatile u32 = @ptrFromInt(0xdead0000);
     poison.* = 1; // the restart machinery's fuel: a real segmentation fault
 }
-
-pub const panic = runtime.panic;
-comptime {
-    _ = &runtime.start._start; // pull the runtime entry shim into the image
-}

system/services/device-list/device-list.zig

@@ -81,8 +81,3 @@ pub fn main() void {
         }
     }
 }
-
-pub const panic = runtime.panic;
-comptime {
-    _ = &runtime.start._start; // pull the runtime entry shim into the image
-}

system/services/device-manager/device-manager.zig

@@ -586,8 +586,3 @@ pub fn main(init: runtime.process.Init) void {
         .on_notification = onNotification,
     });
 }
-
-pub const panic = runtime.panic;
-comptime {
-    _ = &runtime.start._start; // pull the runtime entry shim into the image
-}

system/services/display-demo/display-demo.zig

@@ -45,7 +45,6 @@ pub fn main() void {
     var dx: i32 = 8;
     var frame: u32 = 0;
 
-
     var mouse = input.subscribeMouse(); // type: ?input.MouseSubscriber
     if (mouse == null) _ = system.write("display-demo: no mouse; animating without it\n");
 
@@ -85,8 +84,3 @@ fn clamp(v: i32, lo: i32, hi: i32) i32 {
 fn createFailed() void {
     _ = system.write("display-demo: create failed\n");
 }
-
-pub const panic = runtime.panic;
-comptime {
-    _ = &runtime.start._start;
-}

system/services/display/display.zig

@@ -455,8 +455,3 @@ pub fn main() void {
         .on_notification = onNotification,
     });
 }
-
-pub const panic = runtime.panic;
-comptime {
-    _ = &runtime.start._start;
-}

system/services/fat/fat-test.zig

@@ -101,8 +101,3 @@ pub fn main(init: runtime.process.Init) void {
     }
     _ = runtime.system.write("fat-test: root listing was empty\n");
 }
-
-pub const panic = runtime.panic;
-comptime {
-    _ = &runtime.start._start;
-}

system/services/fat/fat.zig

@@ -242,8 +242,3 @@ pub fn main() void {
         .on_message = onMessage,
     });
 }
-
-pub const panic = runtime.panic;
-comptime {
-    _ = &runtime.start._start;
-}

system/services/fdt/fdt.zig

@@ -28,8 +28,3 @@ pub fn main(init: runtime.process.Init) void {
     // supervisor reads a clean exit as "meant to stop" — correct for a
     // placeholder.
 }
-
-pub const panic = runtime.panic;
-comptime {
-    _ = &runtime.start._start; // pull the runtime entry shim into the image
-}

system/services/init/init.zig

@@ -20,6 +20,7 @@
 const std = @import("std");
 const runtime = @import("runtime");
 const power = runtime.power_protocol;
+const build_options = @import("build_options");
 
 /// Where the kernel boot log is persisted on the USB FAT volume — an 8.3 name at
 /// the mount root (see system/services/log-flush). init writes it at shutdown;
@@ -90,10 +91,13 @@ pub fn main() void {
     // triggers the same shutdown path.
     subscribePower();
 
-    // A re-arming timer drives the liveness heartbeat: proof PID 1 is alive
+    // A re-arming timer drives the liveness heartbeat — proof PID 1 is alive (the
-    // (the init test's marker) while the loop stays free to receive signals,
+    // init test's marker) and a -Dserial diagnostic. It is a serial/test-build-only
-    // power events, and children's exit notifications.
+    // concern: a flashable (serial-off) image runs a purely event-driven PID 1 that
-    _ = runtime.system.timerOnce(supervision_endpoint, 1000);
+    // wakes only for real work (signals, power events, children's exits), never for a
+    // periodic beat. `build_options.serial` is comptime, so the heartbeat — its timer
+    // and the handler below — folds away entirely when serial is off.
+    if (build_options.serial) _ = runtime.system.timerOnce(supervision_endpoint, 1000);
 
     var receive: [power.message_maximum]u8 = undefined;
     while (true) {
@@ -102,7 +106,7 @@ pub fn main() void {
             if (signals.has(.terminate)) shutDown();
             continue;
         }
-        if (got.isTimer()) {
+        if (build_options.serial and got.isTimer()) {
             _ = runtime.system.write("/system/services/init: heartbeat\n");
             _ = runtime.system.timerOnce(supervision_endpoint, 1000);
             continue;
@@ -215,8 +219,3 @@ fn shutDown() void {
     // If S5 did not take, init has nothing left to do but idle.
     while (true) runtime.system.sleep(1000);
 }
-
-pub const panic = runtime.panic;
-comptime {
-    _ = &runtime.start._start; // pull the runtime entry shim into the image
-}

system/services/input-source/input-source.zig

@@ -33,8 +33,3 @@ pub fn main() void {
         system.sleep(200);
     }
 }
-
-pub const panic = runtime.panic;
-comptime {
-    _ = &runtime.start._start;
-}

system/services/input-test/input-test.zig

@@ -49,8 +49,3 @@ pub fn main() void {
         }
     }
 }
-
-pub const panic = runtime.panic;
-comptime {
-    _ = &runtime.start._start;
-}

system/services/input/input.zig

@@ -138,8 +138,3 @@ pub fn main() void {
         reply_len = handle(receive[0..got.len], got, &reply_buffer);
     }
 }
-
-pub const panic = runtime.panic;
-comptime {
-    _ = &runtime.start._start;
-}

system/services/log-flush/log-flush.zig

@@ -60,8 +60,3 @@ pub fn main() void {
     var line: [96]u8 = undefined;
     _ = runtime.system.write(std.fmt.bufPrint(&line, "log-flush: wrote {d} bytes to {s}\n", .{ written, log_path }) catch return);
 }
-
-pub const panic = runtime.panic;
-comptime {
-    _ = &runtime.start._start; // pull the runtime entry shim into the image
-}

system/services/process-test/process-test.zig

@@ -178,8 +178,3 @@ pub fn main(init: runtime.process.Init) void {
 
     _ = runtime.system.write("process-test: ok\n");
 }
-
-pub const panic = runtime.panic;
-comptime {
-    _ = &runtime.start._start; // pull the runtime entry shim into the image
-}

system/services/shm-client/shm-client.zig

@@ -44,8 +44,3 @@ pub fn main() void {
         _ = system.write("shm: call failed\n");
     };
 }
-
-pub const panic = runtime.panic;
-comptime {
-    _ = &runtime.start._start;
-}

system/services/shm-server/shm-server.zig

@@ -42,8 +42,3 @@ fn onMessage(message: []const u8, reply: []u8, sender: u32, capability: ?ipc.Han
 pub fn main() void {
     runtime.service.run(64, .{ .service = .shm_test, .on_message = onMessage });
 }
-
-pub const panic = runtime.panic;
-comptime {
-    _ = &runtime.start._start;
-}

system/services/thread-test/thread-test.zig

@@ -0,0 +1,310 @@
+//! thread-test — danos's multi-threaded exerciser (docs/threading-plan.md M2, M3).
+//!
+//! Two modes, chosen by argv[1] (default "spawn"):
+//!   spawn — M2: one worker writes a shared global; the main thread observes it, proving
+//!           `runtime.Thread.spawn` started a task in the **same** address space.
+//!   join  — M3: N workers each do K atomic increments on a shared counter and stamp the
+//!           core they ran on; the main thread `join`s all N and checks the total is
+//!           exactly N*K (every worker ran, join waited) and that >1 core was used
+//!           (genuine parallelism). Then a detached worker proves `detach` runs and
+//!           needs no join.
+//!
+//! Built multi-threaded (`addThreadedUserBinary`) so atomics/shared reads are real.
+
+const std = @import("std");
+const runtime = @import("runtime");
+
+fn write(comptime s: []const u8) void {
+    _ = runtime.system.write(s);
+}
+
+// --- M2: spawn mode ---------------------------------------------------------
+
+var shared_value: u32 = 0;
+var spawn_done = std.atomic.Value(u32).init(0);
+const sentinel: u32 = 0xA5A5;
+
+fn spawnWorker() void {
+    shared_value = sentinel;
+    spawn_done.store(1, .release);
+}
+
+fn runSpawnMode() void {
+    write("thread-test: starting\n");
+    _ = runtime.Thread.spawn(.{}, spawnWorker, .{}) catch {
+        write("thread-test: FAIL spawn refused\n");
+        return;
+    };
+    var spins: usize = 0;
+    while (spawn_done.load(.acquire) == 0 and spins < 50_000_000) : (spins += 1) {
+        runtime.system.yield();
+    }
+    if (spawn_done.load(.acquire) == 1 and shared_value == sentinel) {
+        write("thread-test: child ran in shared aspace ok\n");
+    } else {
+        write("thread-test: FAIL worker did not update shared memory\n");
+    }
+}
+
+// --- M3: join mode ----------------------------------------------------------
+
+const worker_count: u32 = 4;
+const iterations: u64 = 100_000;
+
+var counter = std.atomic.Value(u64).init(0);
+var cores_seen = std.atomic.Value(u32).init(0);
+
+fn joinWorker() void {
+    var i: u64 = 0;
+    while (i < iterations) : (i += 1) {
+        _ = counter.fetchAdd(1, .monotonic);
+        if (i % 1000 == 0) stampCore(); // periodic: catches cross-core migration too
+    }
+    stampCore();
+}
+
+fn stampCore() void {
+    const core = runtime.Thread.currentCore();
+    if (core < 32) _ = cores_seen.fetchOr(@as(u32, 1) << @intCast(core), .monotonic);
+}
+
+var detach_done = std.atomic.Value(u32).init(0);
+
+fn detachWorker() void {
+    detach_done.store(1, .release);
+}
+
+fn runJoinMode() void {
+    write("thread-test: join mode starting\n");
+
+    var threads: [worker_count]runtime.Thread = undefined;
+    var spawned: u32 = 0;
+    while (spawned < worker_count) : (spawned += 1) {
+        threads[spawned] = runtime.Thread.spawn(.{}, joinWorker, .{}) catch break;
+    }
+    if (spawned != worker_count) {
+        write("thread-test: FAIL could not spawn all workers\n");
+        return;
+    }
+    for (threads[0..spawned]) |t| t.join();
+
+    const total = counter.load(.acquire);
+    const cores = @popCount(cores_seen.load(.acquire));
+    if (total != worker_count * iterations) {
+        write("thread-test: FAIL counter mismatch (a worker was lost or join did not wait)\n");
+        return;
+    }
+    if (cores <= 1) {
+        write("thread-test: FAIL workers never ran on more than one core\n");
+        return;
+    }
+
+    // detach: the worker runs and we never join it.
+    const dt = runtime.Thread.spawn(.{}, detachWorker, .{}) catch {
+        write("thread-test: FAIL detach spawn refused\n");
+        return;
+    };
+    dt.detach();
+    var spins: usize = 0;
+    while (detach_done.load(.acquire) == 0 and spins < 50_000_000) : (spins += 1) {
+        runtime.system.yield();
+    }
+    if (detach_done.load(.acquire) != 1) {
+        write("thread-test: FAIL detached worker did not run\n");
+        return;
+    }
+
+    write("thread-test: join ok\n"); // the M3 verdict marker
+}
+
+// --- M4: futex mode ---------------------------------------------------------
+
+const Futex = runtime.Thread.Futex;
+
+var futex_word = std.atomic.Value(u32).init(0);
+var waiter_parked = std.atomic.Value(u32).init(0);
+
+fn futexWaiter() void {
+    write("thread-futex: waiting\n");
+    waiter_parked.store(1, .release);
+    // Block while the word is still 0; the waker sets it to 1 and wakes us.
+    while (futex_word.load(.acquire) == 0) {
+        Futex.wait(&futex_word, 0);
+    }
+    write("thread-futex: woke\n");
+}
+
+fn runFutexMode() void {
+    write("thread-futex: starting\n");
+
+    const waiter = runtime.Thread.spawn(.{}, futexWaiter, .{}) catch {
+        write("thread-futex: FAIL spawn refused\n");
+        return;
+    };
+    // Let the waiter reach its wait, then give it a beat to actually park in-kernel.
+    var spins: usize = 0;
+    while (waiter_parked.load(.acquire) == 0 and spins < 50_000_000) : (spins += 1) {
+        runtime.system.yield();
+    }
+    runtime.system.sleep(50);
+
+    // The handshake: publish the value, then wake the parked waiter.
+    futex_word.store(1, .release);
+    write("thread-futex: waking\n");
+    Futex.wake(&futex_word, 1);
+
+    waiter.join(); // returns once the waiter woke and printed "woke"
+
+    // Timeout: nobody ever wakes this word, so timedWait must report a timeout.
+    var lonely = std.atomic.Value(u32).init(0);
+    if (Futex.timedWait(&lonely, 0, 100_000_000)) |_| {
+        write("thread-futex: FAIL timedWait did not time out\n");
+        return;
+    } else |_| {}
+    write("thread-futex: timeout ok\n");
+
+    write("thread-futex: ok\n"); // the M4 verdict marker
+}
+
+// --- M5: mutex mode (bounded producer/consumer over Mutex + Condition) ------
+
+const Mutex = runtime.Thread.Mutex;
+const Condition = runtime.Thread.Condition;
+
+const producers: u32 = 2;
+const consumers: u32 = 2;
+const per_producer: u32 = 1000;
+const per_consumer: u32 = 1000; // producers*per_producer == consumers*per_consumer (balanced)
+const total_items: u32 = producers * per_producer;
+const ring_cap: usize = 8; // small, so producers block on full and consumers on empty
+
+var ring: [ring_cap]u32 = undefined;
+var ring_count: usize = 0;
+var ring_head: usize = 0;
+var ring_tail: usize = 0;
+
+var pc_mutex = Mutex{};
+var not_full = Condition{};
+var not_empty = Condition{};
+
+// Verified outside the lock: the checksum and tally of everything consumed.
+var consumed_sum = std.atomic.Value(u64).init(0);
+var consumed_count = std.atomic.Value(u32).init(0);
+
+fn producer(base: u32) void {
+    var i: u32 = 0;
+    while (i < per_producer) : (i += 1) {
+        const item = base + i;
+        pc_mutex.lock();
+        while (ring_count == ring_cap) not_full.wait(&pc_mutex);
+        ring[ring_tail] = item;
+        ring_tail = (ring_tail + 1) % ring_cap;
+        ring_count += 1;
+        pc_mutex.unlock();
+        not_empty.signal();
+    }
+}
+
+fn consumer() void {
+    var i: u32 = 0;
+    while (i < per_consumer) : (i += 1) {
+        pc_mutex.lock();
+        while (ring_count == 0) not_empty.wait(&pc_mutex);
+        const item = ring[ring_head];
+        ring_head = (ring_head + 1) % ring_cap;
+        ring_count -= 1;
+        pc_mutex.unlock();
+        not_full.signal();
+        _ = consumed_sum.fetchAdd(item, .monotonic);
+        _ = consumed_count.fetchAdd(1, .monotonic);
+    }
+}
+
+fn runMutexMode() void {
+    write("thread-mutex: starting\n");
+
+    var threads: [producers + consumers]runtime.Thread = undefined;
+    var n: usize = 0;
+    var p: u32 = 0;
+    while (p < producers) : (p += 1) {
+        threads[n] = runtime.Thread.spawn(.{}, producer, .{p * per_producer}) catch {
+            write("thread-mutex: FAIL producer spawn\n");
+            return;
+        };
+        n += 1;
+    }
+    var c: u32 = 0;
+    while (c < consumers) : (c += 1) {
+        threads[n] = runtime.Thread.spawn(.{}, consumer, .{}) catch {
+            write("thread-mutex: FAIL consumer spawn\n");
+            return;
+        };
+        n += 1;
+    }
+    for (threads[0..n]) |t| t.join();
+
+    // Every item 0..total_items-1 was produced exactly once; if the mutex/condition are
+    // correct, each was consumed exactly once, so the checksum matches.
+    const expected_sum: u64 = @as(u64, total_items) * (total_items - 1) / 2;
+    if (consumed_count.load(.acquire) != total_items) {
+        write("thread-mutex: FAIL wrong number of items consumed\n");
+        return;
+    }
+    if (consumed_sum.load(.acquire) != expected_sum) {
+        write("thread-mutex: FAIL checksum mismatch (item lost or duplicated)\n");
+        return;
+    }
+    write("thread-mutex: ok\n"); // the M5 verdict marker
+}
+
+// --- M6: id mode (getCurrentId identity) ------------------------------------
+
+var worker_ids: [2]std.atomic.Value(u32) = .{ std.atomic.Value(u32).init(0), std.atomic.Value(u32).init(0) };
+
+fn idWorker(slot: usize) void {
+    worker_ids[slot].store(runtime.Thread.getCurrentId(), .release);
+}
+
+fn runIdMode() void {
+    write("thread-id: starting\n");
+    const main_id = runtime.Thread.getCurrentId();
+
+    const t0 = runtime.Thread.spawn(.{}, idWorker, .{@as(usize, 0)}) catch {
+        write("thread-id: FAIL spawn\n");
+        return;
+    };
+    const t1 = runtime.Thread.spawn(.{}, idWorker, .{@as(usize, 1)}) catch {
+        write("thread-id: FAIL spawn\n");
+        return;
+    };
+    t0.join();
+    t1.join();
+
+    const id0 = worker_ids[0].load(.acquire);
+    const id1 = worker_ids[1].load(.acquire);
+    // Each thread has its own kernel task id: all three distinct and non-zero.
+    if (main_id == 0 or id0 == 0 or id1 == 0) {
+        write("thread-id: FAIL a thread reported id 0\n");
+        return;
+    }
+    if (id0 == id1 or id0 == main_id or id1 == main_id) {
+        write("thread-id: FAIL thread ids collided\n");
+        return;
+    }
+    write("thread-id: ok\n"); // the M6 verdict marker
+}
+
+pub fn main(init: runtime.process.Init) void {
+    const mode = init.arguments.get(1) orelse "spawn";
+    if (std.mem.eql(u8, mode, "join")) {
+        runJoinMode();
+    } else if (std.mem.eql(u8, mode, "futex")) {
+        runFutexMode();
+    } else if (std.mem.eql(u8, mode, "mutex")) {
+        runMutexMode();
+    } else if (std.mem.eql(u8, mode, "id")) {
+        runIdMode();
+    } else {
+        runSpawnMode();
+    }
+}

system/services/vfs/vfs-test.zig

@@ -60,8 +60,3 @@ pub fn main(init: runtime.process.Init) void {
     }
     _ = runtime.system.write("vfstest: mismatch\n");
 }
-
-pub const panic = runtime.panic;
-comptime {
-    _ = &runtime.start._start;
-}

system/services/vfs/vfs.zig

@@ -394,8 +394,3 @@ pub fn main() void {
         .on_notification = onNotification,
     });
 }
-
-pub const panic = runtime.panic;
-comptime {
-    _ = &runtime.start._start;
-}

test/qemu_test.py

@@ -293,6 +293,52 @@ CASES = [
      "timeout": 60,
      "expect": r"DANOS-TEST-RESULT: PASS",
      "fail":   r"DANOS-TEST-RESULT: FAIL"},
+
+    # docs/threading-plan.md M1: address-space refcount — spaces destroyed exactly
+    # once per process, no leak/double-free (the foundation shared-aspace threads need).
+    {"name": "aspace-refcount",
+     "timeout": 60,
+     "expect": r"DANOS-TEST-RESULT: PASS",
+     "fail":   r"DANOS-TEST-RESULT: FAIL"},
+
+    # docs/threading-plan.md M2: runtime.Thread.spawn — a worker thread runs in the
+    # caller's address space (a shared-memory write, observed by the main thread).
+    {"name": "thread-spawn",
+     "timeout": 60,
+     "expect": r"DANOS-TEST-RESULT: PASS",
+     "fail":   r"DANOS-TEST-RESULT: FAIL"},
+
+    # docs/threading-plan.md M3: join + parallelism — N workers each do K atomic
+    # increments (total exactly N*K after join) and run on >1 core; plus detach.
+    {"name": "thread-join",
+     "smp":    4,
+     "timeout": 60,
+     "expect": r"DANOS-TEST-RESULT: PASS",
+     "fail":   r"DANOS-TEST-RESULT: FAIL"},
+
+    # docs/threading-plan.md M4: futex — a thread parks in futex_wait and is woken by
+    # futex_wake (serial order waiting/waking/woke), and timedWait reports a timeout.
+    {"name": "thread-futex",
+     "smp":    4,
+     "timeout": 60,
+     "expect": r"thread-futex: waiting[\s\S]*thread-futex: waking[\s\S]*thread-futex: woke[\s\S]*DANOS-TEST-RESULT: PASS",
+     "fail":   r"DANOS-TEST-RESULT: FAIL"},
+
+    # docs/threading-plan.md M5: Mutex + Condition — a bounded producer/consumer moves
+    # N unique items across cores; the consumed checksum matches exactly (no loss).
+    {"name": "thread-mutex",
+     "smp":    4,
+     "timeout": 60,
+     "expect": r"DANOS-TEST-RESULT: PASS",
+     "fail":   r"DANOS-TEST-RESULT: FAIL"},
+
+    # docs/threading-plan.md M6: thread identity — getCurrentId is distinct and non-zero
+    # for the main thread and two workers.
+    {"name": "thread-id",
+     "smp":    4,
+     "timeout": 60,
+     "expect": r"DANOS-TEST-RESULT: PASS",
+     "fail":   r"DANOS-TEST-RESULT: FAIL"},
     # Process arguments: argv arrives on the SysV entry stack (argv[0] = the spawned
     # name, argv[1..] = the system_spawn argument blob) and echoes back intact.
     {"name": "args",

tools/make-iso-image.py

@@ -0,0 +1,261 @@
+#!/usr/bin/env python3
+"""Wrap the FAT32 boot volume in a hybrid ISO — the flashable danos release image.
+
+Mirrors tools/make-fat-image.py in spirit: pure Python 3 standard library, no
+external tools (no xorriso / mkisofs / isohybrid). The output is one file that
+boots both ways release media is consumed:
+
+  * Flashed raw to a USB stick (Etcher, dd): the ISO's system area carries an
+    MBR whose single partition (type 0xEF, "EFI System") points at the FAT32
+    image embedded in the ISO, so UEFI firmware finds the ESP and runs
+    \\EFI\\BOOT\\BOOTX64.efi off it.
+  * Burned to optical media: an El Torito boot catalog with an EFI platform
+    entry points at the same embedded FAT image.
+
+The ISO9660 filesystem itself is minimal but valid — a primary volume
+descriptor, the El Torito boot record, path tables, and a root directory that
+lists the boot catalog and the FAT image — so inspection tools can open it.
+
+    make-iso-image.py <out.iso> <esp.img>
+    make-iso-image.py --verify <out.iso>
+
+All timestamp fields are zero ("not specified") so the build is reproducible.
+"""
+
+import struct
+import sys
+
+ISO_SECTOR = 2048
+
+# Fixed layout, in ISO sectors (LBA). Sectors 0-15 are the system area (the
+# hybrid MBR lives in its first 512 bytes); volume descriptors start at 16.
+PVD_LBA = 16          # primary volume descriptor
+BOOT_RECORD_LBA = 17  # El Torito boot record volume descriptor
+TERMINATOR_LBA = 18   # volume descriptor set terminator
+PATH_TABLE_L_LBA = 19
+PATH_TABLE_M_LBA = 20
+ROOT_DIR_LBA = 21     # root directory (one sector holds our four records)
+CATALOG_LBA = 22      # El Torito boot catalog
+ESP_LBA = 23          # the embedded FAT32 image starts here
+
+MBR_PARTITION_TYPE_ESP = 0xEF
+
+
+def both16(value):
+    """ISO9660 both-byte-order encoding: little-endian then big-endian."""
+    return struct.pack("<H", value) + struct.pack(">H", value)
+
+
+def both32(value):
+    return struct.pack("<I", value) + struct.pack(">I", value)
+
+
+def directory_record(identifier, lba, size, flags):
+    length = 33 + len(identifier)
+    if length % 2:
+        length += 1  # records are padded to even length
+    record = bytearray(length)
+    record[0] = length
+    record[2:10] = both32(lba)
+    record[10:18] = both32(size)
+    # record[18:25] is the recording date; zero = unspecified (reproducible).
+    record[25] = flags  # 0x02 = directory
+    record[28:32] = both16(1)  # volume sequence number
+    record[32] = len(identifier)
+    record[33:33 + len(identifier)] = identifier
+    return bytes(record)
+
+
+def primary_volume_descriptor(total_sectors, path_table_size):
+    sector = bytearray(ISO_SECTOR)
+    sector[0] = 1  # type: primary
+    sector[1:6] = b"CD001"
+    sector[6] = 1  # version
+    sector[8:40] = b"DANOS".ljust(32)  # system identifier
+    sector[40:72] = b"DANOS".ljust(32)  # volume identifier
+    sector[80:88] = both32(total_sectors)
+    sector[120:124] = both16(1)  # volume set size
+    sector[124:128] = both16(1)  # volume sequence number
+    sector[128:132] = both16(ISO_SECTOR)  # logical block size
+    sector[132:140] = both32(path_table_size)
+    sector[140:144] = struct.pack("<I", PATH_TABLE_L_LBA)
+    sector[148:152] = struct.pack(">I", PATH_TABLE_M_LBA)
+    sector[156:190] = directory_record(b"\x00", ROOT_DIR_LBA, ISO_SECTOR, 0x02)
+    sector[190:318] = b" " * 128  # volume set identifier
+    sector[318:446] = b" " * 128  # publisher
+    sector[446:574] = b" " * 128  # data preparer
+    sector[574:702] = b"DANOS MAKE-ISO-IMAGE".ljust(128)  # application
+    sector[702:739] = b" " * 37  # copyright file
+    sector[739:776] = b" " * 37  # abstract file
+    sector[776:813] = b" " * 37  # bibliographic file
+    unspecified_date = b"0" * 16 + b"\x00"
+    for offset in (813, 830, 847, 864):  # creation/modification/expiry/effective
+        sector[offset:offset + 17] = unspecified_date
+    sector[881] = 1  # file structure version
+    return bytes(sector)
+
+
+def boot_record_descriptor():
+    sector = bytearray(ISO_SECTOR)
+    sector[0] = 0  # type: boot record
+    sector[1:6] = b"CD001"
+    sector[6] = 1
+    sector[7:39] = b"EL TORITO SPECIFICATION".ljust(32, b"\x00")
+    sector[71:75] = struct.pack("<I", CATALOG_LBA)
+    return bytes(sector)
+
+
+def terminator_descriptor():
+    sector = bytearray(ISO_SECTOR)
+    sector[0] = 255
+    sector[1:6] = b"CD001"
+    sector[6] = 1
+    return bytes(sector)
+
+
+def path_table(byte_order):
+    # A single entry: the root directory.
+    return (struct.pack("BB", 1, 0)
+            + struct.pack(byte_order + "I", ROOT_DIR_LBA)
+            + struct.pack(byte_order + "H", 1)
+            + b"\x00\x00")  # identifier 0x00 + pad to even
+
+
+def root_directory(esp_size):
+    # Records must be sorted by identifier; BOOT.CAT < EFI.IMG holds.
+    entries = (directory_record(b"\x00", ROOT_DIR_LBA, ISO_SECTOR, 0x02)
+               + directory_record(b"\x01", ROOT_DIR_LBA, ISO_SECTOR, 0x02)
+               + directory_record(b"BOOT.CAT;1", CATALOG_LBA, ISO_SECTOR, 0)
+               + directory_record(b"EFI.IMG;1", ESP_LBA, esp_size, 0))
+    return entries + b"\x00" * (ISO_SECTOR - len(entries))
+
+
+def boot_catalog(esp_size):
+    # Validation entry: EFI platform (0xEF), checksummed so its 16-bit words sum
+    # to zero, closed by the 0x55AA key bytes.
+    validation = bytearray(32)
+    validation[0] = 0x01
+    validation[1] = 0xEF
+    validation[4:28] = b"danos".ljust(24, b"\x00")
+    validation[30] = 0x55
+    validation[31] = 0xAA
+    checksum = (-sum(struct.unpack("<16H", validation))) & 0xFFFF
+    validation[28:30] = struct.pack("<H", checksum)
+    # Initial/default entry: bootable, no emulation, image at ESP_LBA. The
+    # sector-count field is 16-bit (units of 512 bytes) so it can't span a large
+    # ESP; UEFI firmware sizes the FAT filesystem from its own BPB, and the
+    # image's boot files sit well inside the capped span regardless.
+    default = bytearray(32)
+    default[0] = 0x88  # bootable
+    default[1] = 0x00  # no emulation
+    sector_count = min(0xFFFF, esp_size // 512)
+    default[6:8] = struct.pack("<H", sector_count)
+    default[8:12] = struct.pack("<I", ESP_LBA)
+    catalog = bytes(validation) + bytes(default)
+    return catalog + b"\x00" * (ISO_SECTOR - len(catalog))
+
+
+def hybrid_mbr(esp_size):
+    """The system-area MBR that makes the ISO flashable: one ESP partition."""
+    mbr = bytearray(512)
+    mbr[440:444] = b"dano"  # disk signature (fixed: reproducible builds)
+    start_lba = ESP_LBA * (ISO_SECTOR // 512)
+    partition = struct.pack(
+        "<B3sB3sII",
+        0x80,                      # status: active (harmless; helps picky firmware)
+        b"\xFE\xFF\xFF",           # CHS start: maxed out, LBA is authoritative
+        MBR_PARTITION_TYPE_ESP,    # type: EFI System
+        b"\xFE\xFF\xFF",           # CHS end
+        start_lba,
+        esp_size // 512,
+    )
+    mbr[446:462] = partition
+    mbr[510] = 0x55
+    mbr[511] = 0xAA
+    return bytes(mbr)
+
+
+def build(out_path, esp_path):
+    with open(esp_path, "rb") as handle:
+        esp = handle.read()
+    if len(esp) % ISO_SECTOR:
+        esp += b"\x00" * (ISO_SECTOR - len(esp) % ISO_SECTOR)
+    esp_sectors = len(esp) // ISO_SECTOR
+    total_sectors = ESP_LBA + esp_sectors
+
+    table_l = path_table("<")
+    image = bytearray(total_sectors * ISO_SECTOR)
+    image[0:512] = hybrid_mbr(len(esp))
+    image[PVD_LBA * ISO_SECTOR:(PVD_LBA + 1) * ISO_SECTOR] = \
+        primary_volume_descriptor(total_sectors, len(table_l))
+    image[BOOT_RECORD_LBA * ISO_SECTOR:(BOOT_RECORD_LBA + 1) * ISO_SECTOR] = \
+        boot_record_descriptor()
+    image[TERMINATOR_LBA * ISO_SECTOR:(TERMINATOR_LBA + 1) * ISO_SECTOR] = \
+        terminator_descriptor()
+    image[PATH_TABLE_L_LBA * ISO_SECTOR:PATH_TABLE_L_LBA * ISO_SECTOR + len(table_l)] = table_l
+    table_m = path_table(">")
+    image[PATH_TABLE_M_LBA * ISO_SECTOR:PATH_TABLE_M_LBA * ISO_SECTOR + len(table_m)] = table_m
+    image[ROOT_DIR_LBA * ISO_SECTOR:(ROOT_DIR_LBA + 1) * ISO_SECTOR] = root_directory(len(esp))
+    image[CATALOG_LBA * ISO_SECTOR:(CATALOG_LBA + 1) * ISO_SECTOR] = boot_catalog(len(esp))
+    image[ESP_LBA * ISO_SECTOR:] = esp
+
+    with open(out_path, "wb") as handle:
+        handle.write(image)
+    print(f"make-iso-image: wrote {out_path} "
+          f"({total_sectors * ISO_SECTOR // (1024 * 1024)} MiB hybrid ISO, "
+          f"ESP at LBA {ESP_LBA}, {esp_sectors} sectors)")
+
+
+def verify(path):
+    with open(path, "rb") as handle:
+        data = handle.read()
+    # The hybrid MBR (the Etcher/dd boot path).
+    if data[510] != 0x55 or data[511] != 0xAA:
+        sys.exit("verify: missing MBR 0x55AA signature")
+    status, _, part_type, _, part_start, part_sectors = \
+        struct.unpack_from("<B3sB3sII", data, 446)
+    if part_type != MBR_PARTITION_TYPE_ESP:
+        sys.exit(f"verify: MBR partition type 0x{part_type:02X}, expected 0xEF (ESP)")
+    # The ISO9660 descriptors (the optical boot path).
+    if data[PVD_LBA * ISO_SECTOR + 1:PVD_LBA * ISO_SECTOR + 6] != b"CD001":
+        sys.exit("verify: no primary volume descriptor")
+    boot_record = data[BOOT_RECORD_LBA * ISO_SECTOR:(BOOT_RECORD_LBA + 1) * ISO_SECTOR]
+    if not boot_record.startswith(b"\x00CD001") or \
+            not boot_record[7:30].startswith(b"EL TORITO SPECIFICATION"):
+        sys.exit("verify: no El Torito boot record")
+    catalog_lba = struct.unpack_from("<I", boot_record, 71)[0]
+    catalog = data[catalog_lba * ISO_SECTOR:(catalog_lba + 1) * ISO_SECTOR]
+    if catalog[0] != 0x01 or catalog[1] != 0xEF or catalog[30:32] != b"\x55\xAA":
+        sys.exit("verify: boot catalog validation entry is not an EFI entry")
+    if sum(struct.unpack("<16H", catalog[0:32])) & 0xFFFF != 0:
+        sys.exit("verify: boot catalog validation checksum is wrong")
+    if catalog[32] != 0x88:
+        sys.exit("verify: default catalog entry is not bootable")
+    boot_lba = struct.unpack_from("<I", catalog, 40)[0]
+    # Both paths must agree on where the ESP lives, and it must be a FAT32 image.
+    if boot_lba * (ISO_SECTOR // 512) != part_start:
+        sys.exit(f"verify: catalog boot image (LBA {boot_lba}) and MBR partition "
+                 f"(sector {part_start}) disagree")
+    esp = data[boot_lba * ISO_SECTOR:]
+    if len(esp) < part_sectors * 512:
+        sys.exit("verify: MBR partition extends past the end of the file")
+    if esp[510] != 0x55 or esp[511] != 0xAA or esp[82:90] != b"FAT32   ":
+        sys.exit("verify: embedded image is not a FAT32 boot volume")
+    print(f"verify: {path} is a hybrid ISO — MBR ESP partition (sector {part_start}, "
+          f"{part_sectors} sectors, active={status == 0x80}) and El Torito EFI entry "
+          f"both point at the embedded FAT32 image")
+
+
+def main(argv):
+    if len(argv) == 3 and argv[1] == "--verify":
+        verify(argv[2])
+        return 0
+    if len(argv) != 3:
+        sys.exit("usage: make-iso-image.py <out.iso> <esp.img>\n"
+                 "       make-iso-image.py --verify <out.iso>")
+    build(argv[1], argv[2])
+    return 0
+
+
+if __name__ == "__main__":
+    sys.exit(main(sys.argv))