display: track the mouse with a listener thread (Shape A) + cursor

The display's first use of threads (docs/threading.md, docs/display.md). The
compositor stays the single owner of the framebuffer — only the main
service.run loop touches the backend and layer stack — and a dedicated
mouse-listener thread runs beside it:

  - Listener: blocks on input.subscribeMouse(), accumulates relative dx/dy
    into an absolute cursor position clamped to the screen, and hands it to
    the compositor. It never touches the compositor, so no lock guards the
    framebuffer; a parked next() lets the core halt.
  - CursorChannel: a single-slot latest-value cell under a Thread.Mutex (the
    renderer wants where the cursor is now, not a replay of deltas), with a
    coalesced self-ipc.send poke that wakes the main loop — parked in
    replyWait — as a message-notification. At most one poke is queued while
    the last is undrained, so a fast mouse can't flood the endpoint.
  - Render: the cursor is a top-z compositor layer; on the poke the main loop
    moves it via configure + present (which damages old + new footprints).

The display binary opts into threads (addThreadedUserBinary), and
input-source gains a "mouse" mode that publishes pure motion to drive it.

Two kernel-level findings this surfaced, both fixed:

  1. IPC handles do not cross threads. The handle table lives on the Task, so
     the listener can't reuse the main loop's endpoint handle — it
     ipc.lookup(.display)s its own handle to the same endpoint to poke through.

  2. Concurrent IPC from two threads raced unlocked kernel state. The display
     is the first process issuing IPC syscalls from two threads at once, which
     exposed a data race (flaky #GP in installEntry): create_ipc_endpoint /
     ipc_register / ipc_lookup allocate from the kernel heap and mutate the
     global registry, endpoint refcounts, and handle tables without the big
     kernel lock. They were safe only while a process couldn't race itself.
     They now sync.enter() like call/reply_wait/send already did (the kernel
     heap has no lock of its own yet — heap.zig: "a lock comes with
     threads/SMP" — so the big lock keeps its callers serialized).

Test: -Dtest-case=display-cursor (smp:4) spawns the input service, the
threaded display, and input-source in mouse mode; asserts the display's
"cursor tracking mouse ok" marker once the cursor has tracked a run of motion
end to end. Verified green 6/6 under stress (the race hit ~1-in-4 before the
lock fix) and in the full 29-case QEMU guardrail suite; zig build test clean.

system/kernel/process.zig

@@ -256,6 +256,13 @@ fn failErr(state: *architecture.CpuState, errno: i64) void {
 /// create_ipc_endpoint() -> handle: allocate an endpoint and install it in the
 /// caller's handle table.
 fn systemCreateIpcEndpoint(state: *architecture.CpuState) void {
+    // Under the big kernel lock: this allocates from the kernel heap and mutates the
+    // caller's handle table. A multi-threaded process (e.g. the display's compositor +
+    // mouse-listener threads) can drive this concurrently from two cores, so the endpoint
+    // allocation and every other lock holder must serialize (heap.zig: "a lock comes with
+    // threads/SMP").
+    const flags = sync.enter();
+    defer sync.leave(flags);
     const endpoint = ipc.createIpcEndpoint() orelse return failErr(state, ipc.ENOMEM);
     const h = ipc.installHandle(scheduler.current(), endpoint);
     if (h < 0) {
@@ -268,6 +275,10 @@ fn systemCreateIpcEndpoint(state: *architecture.CpuState) void {
 /// ipc_register(service_id, handle): publish the caller's endpoint under a
 /// well-known id so other processes can find it.
 fn systemIpcRegister(state: *architecture.CpuState) void {
+    // Under the big kernel lock: mutates the global service registry and endpoint
+    // refcounts, which threads of the same (or another) process can race.
+    const flags = sync.enter();
+    defer sync.leave(flags);
     const id: u32 = @truncate(architecture.systemCallArg(state, 0));
     const endpoint = ipc.resolveHandle(scheduler.current(), architecture.systemCallArg(state, 1)) orelse return failErr(state, ipc.EBADF);
     architecture.setSystemCallResult(state, @bitCast(ipc.register(id, endpoint)));
@@ -276,6 +287,11 @@ fn systemIpcRegister(state: *architecture.CpuState) void {
 /// ipc_lookup(service_id) -> handle: find a published endpoint and install a
 /// handle to it in the caller.
 fn systemIpcLookup(state: *architecture.CpuState) void {
+    // Under the big kernel lock: reads the global registry, takes an endpoint reference,
+    // and installs a handle — all racy against concurrent threads (this is the path the
+    // display's mouse-listener thread takes to reach the compositor endpoint).
+    const flags = sync.enter();
+    defer sync.leave(flags);
     const id: u32 = @truncate(architecture.systemCallArg(state, 0));
     const endpoint = ipc.lookup(id) orelse return failErr(state, ipc.ENOENT);
     const h = ipc.installHandle(scheduler.current(), endpoint);