The application surface: enumerate, subscribe, and device-list (M18.3)

Applications ask the device manager for the tree (enumerate: a header
plus ChildEntry records) and subscribe to published add/remove events by
handing their endpoint over as the call's capability — the input-service
pattern; events are the same ChildAdded/ChildRemoved structs the bus
drivers send, one encoding in both directions. device-list is the first
client: it prints the tree, subscribes, and narrates the events through
a driver restart. The protocol's message maximum is capped at the
kernel's IPC MESSAGE_MAXIMUM (256 bytes, ten entries per reply; paging
joins the protocol when a tree outgrows one message). The startUserTask
debug print is gone: it wrote to serial unserialized against user-space
lines and sheared concurrent log markers in half — the root cause of the
scenario flakes.

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

@@ -0,0 +1,88 @@
+//! device-list — the `ps` analog for the device tree (docs/device-manager.md
+//! M18.3): asks the device manager for the tree over IPC, prints it, then
+//! subscribes and prints every published add/remove event. The manager is the
+//! one answer to "what devices exist" for user space; nothing here touches a
+//! device_* system call.
+
+const std = @import("std");
+const runtime = @import("runtime");
+const protocol = runtime.device_manager_protocol;
+
+fn writeLine(comptime fmt: []const u8, arguments: anytype) void {
+    var line: [96]u8 = undefined;
+    _ = runtime.system.write(std.fmt.bufPrint(&line, fmt, arguments) catch return);
+}
+
+pub fn main() void {
+    var manager: ?runtime.ipc.Handle = null;
+    var tries: u32 = 0;
+    while (manager == null and tries < 200) : (tries += 1) {
+        manager = runtime.ipc.lookup(.device_manager);
+        if (manager == null) runtime.system.sleep(20);
+    }
+    const h = manager orelse {
+        _ = runtime.system.write("device-list: no device manager\n");
+        return;
+    };
+
+    // The snapshot — polled briefly, because at boot the bus drivers may still
+    // be scanning: an empty first answer usually just means "too early".
+    var reply: [protocol.message_maximum]u8 = undefined;
+    var count: u32 = 0;
+    var length: usize = 0;
+    tries = 0;
+    while (tries < 20) : (tries += 1) {
+        const request = protocol.Enumerate{};
+        length = runtime.ipc.call(h, std.mem.asBytes(&request), &reply) catch 0;
+        if (length >= @sizeOf(protocol.EnumerateReply)) {
+            count = std.mem.bytesToValue(protocol.EnumerateReply, reply[0..@sizeOf(protocol.EnumerateReply)]).count;
+            if (count != 0) break;
+        }
+        runtime.system.sleep(100);
+    }
+    writeLine("device-list: {d} devices\n", .{count});
+    var offset: usize = @sizeOf(protocol.EnumerateReply);
+    var index: u32 = 0;
+    while (index < count and offset + @sizeOf(protocol.ChildEntry) <= length) : (index += 1) {
+        const entry = std.mem.bytesToValue(protocol.ChildEntry, reply[offset..][0..@sizeOf(protocol.ChildEntry)]);
+        writeLine("device-list: device {d} port {d} identity {d}\n", .{ entry.parent, entry.bus_address, entry.identity });
+        offset += @sizeOf(protocol.ChildEntry);
+    }
+
+    // The subscription: our endpoint rides as the call's capability; events
+    // arrive as buffered messages carrying the same structs the bus sends.
+    const endpoint = runtime.ipc.createIpcEndpoint() orelse {
+        _ = runtime.system.write("device-list: no endpoint\n");
+        return;
+    };
+    const subscribe = protocol.Subscribe{};
+    _ = runtime.ipc.callCap(h, std.mem.asBytes(&subscribe), &reply, endpoint) catch {
+        _ = runtime.system.write("device-list: subscribe failed\n");
+        return;
+    };
+    _ = runtime.system.write("device-list: subscribed\n");
+
+    var receive: [protocol.message_maximum]u8 = undefined;
+    while (true) {
+        const got = runtime.ipc.replyWait(endpoint, &.{}, &receive, null);
+        if (!got.isMessage() or got.len < 1) continue;
+        switch (receive[0]) {
+            @intFromEnum(protocol.Operation.child_added) => {
+                if (got.len < protocol.child_added_size) continue;
+                const event = std.mem.bytesToValue(protocol.ChildAdded, receive[0..protocol.child_added_size]);
+                writeLine("device-list: added (device {d} port {d})\n", .{ event.parent, event.bus_address });
+            },
+            @intFromEnum(protocol.Operation.child_removed) => {
+                if (got.len < protocol.child_removed_size) continue;
+                const event = std.mem.bytesToValue(protocol.ChildRemoved, receive[0..protocol.child_removed_size]);
+                writeLine("device-list: removed (device {d} port {d})\n", .{ event.parent, event.bus_address });
+            },
+            else => {},
+        }
+    }
+}
+
+pub const panic = runtime.panic;
+comptime {
+    _ = &runtime.start._start; // pull the runtime entry shim into the image
+}

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

@@ -19,11 +19,13 @@ pub const Role = enum(u8) {
     device = 2,
 };
 
-/// The message kinds. `enumerate`/`subscribe` land in M18.3.
+/// The message kinds.
 pub const Operation = enum(u8) {
     hello = 1,
     child_added = 2,
     child_removed = 3,
+    enumerate = 4,
+    subscribe = 5,
 };
 
 /// `Hello.device_id` for a driver that serves no enumerated device (a test
@@ -97,5 +99,41 @@ pub const ReportReply = extern struct {
     reserved: u32 = 0,
 };
 
+/// An application asking for the tree (M18.3): the reply is an EnumerateReply
+/// header followed by `count` ChildEntry records.
+pub const Enumerate = extern struct {
+    operation: u8 = @intFromEnum(Operation.enumerate),
+    reserved0: u8 = 0,
+    reserved1: u16 = 0,
+    reserved2: u32 = 0,
+};
+
+pub const EnumerateReply = extern struct {
+    status: i32,
+    /// ChildEntry records following this header.
+    count: u32,
+};
+
+pub const ChildEntry = extern struct {
+    parent: u64,
+    bus_address: u64,
+    identity: u64,
+};
+
+/// An application subscribing to published add/remove events (the input-service
+/// pattern): the subscriber's endpoint rides as the call's **capability**, and
+/// events arrive on it as buffered messages whose payload is the same
+/// ChildAdded / ChildRemoved struct the bus drivers send — one encoding, both
+/// directions.
+pub const Subscribe = extern struct {
+    operation: u8 = @intFromEnum(Operation.subscribe),
+    reserved0: u8 = 0,
+    reserved1: u16 = 0,
+    reserved2: u32 = 0,
+};
+
 /// Upper bound on any message in this protocol — sizes the endpoint buffers.
-pub const message_maximum = 64;
+/// Capped by the kernel's IPC MESSAGE_MAXIMUM (256): an EnumerateReply carries
+/// up to ten ChildEntry records per call, plenty for the mirror's current
+/// bounds; paging joins the protocol if a tree ever outgrows one message.
+pub const message_maximum = 256;

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

@@ -108,6 +108,25 @@ var manager_endpoint: runtime.ipc.Handle = 0;
 var test_restart_mode = false;
 var test_usb_restart_mode = false;
 var test_usb_killed = false;
+var test_kill_pid: u32 = 0;
+var test_kill_due_ns: u64 = 0;
+
+/// The application subscribers (M18.3, the input-service pattern): endpoints
+/// handed over as capabilities, each receiving every child add/remove as a
+/// buffered message. A subscriber whose endpoint stops accepting (it died) is
+/// dropped on the failed send.
+const maximum_subscribers = 8;
+var subscribers: [maximum_subscribers]?runtime.ipc.Handle = .{null} ** maximum_subscribers;
+
+/// Publish one event (a ChildAdded or ChildRemoved struct, the same encoding
+/// the bus drivers send) to every subscriber.
+fn publishEvent(event: []const u8) void {
+    for (&subscribers) |*slot| {
+        if (slot.*) |handle| {
+            if (!runtime.ipc.send(handle, event)) slot.* = null; // dead subscriber
+        }
+    }
+}
 
 /// The manager's mirror of what bus drivers report (docs/device-manager.md "the
 /// tree"): the children, keyed by (parent, bus address), each remembering which
@@ -144,11 +163,14 @@ fn addChild(parent: u64, bus_address: u64, identity: u64, reporter: u32) bool {
 /// Prune every child a dead driver instance reported: the children describe
 /// protocol state (slots, rings) that died with the process — keeping the nodes
 /// would be keeping a lie. The restarted instance rediscovers and re-reports.
+/// Watchers hear the honest story: removed now, added again on rediscovery.
 fn pruneChildrenOf(reporter: u32) void {
     for (&children) |*child| {
         if (child.used and child.reporter == reporter) {
             writeLine("device-manager: child removed (device {d} port {d})\n", .{ child.parent, child.bus_address });
             child.used = false;
+            const event = protocol.ChildRemoved{ .parent = child.parent, .bus_address = child.bus_address };
+            publishEvent(std.mem.asBytes(&event));
         }
     }
 }
@@ -259,6 +281,11 @@ fn onDriverExit(driver: *Driver) void {
 /// sweep serves every armed deadline.
 fn sweepDeadlines() void {
     const now = system.clock();
+    if (test_kill_pid != 0 and now >= test_kill_due_ns) {
+        writeLine("device-manager: test mode: killing the reporter\n", .{});
+        _ = system.kill(test_kill_pid);
+        test_kill_pid = 0;
+    }
     for (&drivers) |*driver| {
         if (!driver.used) continue;
         switch (driver.state) {
@@ -318,11 +345,13 @@ fn initialise(endpoint: runtime.ipc.Handle) bool {
     return true;
 }
 
-fn onMessage(message: []const u8, reply: []u8, sender: u32) usize {
+fn onMessage(message: []const u8, reply: []u8, sender: u32, capability: ?runtime.ipc.Handle) usize {
     if (message.len < 1) return 0;
     switch (message[0]) {
         @intFromEnum(protocol.Operation.child_added) => return onChildAdded(message, reply, sender),
         @intFromEnum(protocol.Operation.child_removed) => return onChildRemoved(message, reply, sender),
+        @intFromEnum(protocol.Operation.enumerate) => return onEnumerate(reply),
+        @intFromEnum(protocol.Operation.subscribe) => return onSubscribe(reply, capability),
         @intFromEnum(protocol.Operation.hello) => {},
         else => return 0,
     }
@@ -355,15 +384,19 @@ fn onChildAdded(message: []const u8, reply: []u8, sender: u32) usize {
     if (driverByProcess(sender)) |driver| {
         if (!addChild(report.parent, report.bus_address, report.identity, sender)) status = -1;
         writeLine("device-manager: child added (device {d} port {d}, identity {d}) by {s}\n", .{ report.parent, report.bus_address, report.identity, driver.name() });
+        if (status == 0) publishEvent(message[0..protocol.child_added_size]);
     } else {
         status = -1;
     }
     const report_reply = protocol.ReportReply{ .status = status };
     @memcpy(reply[0..@sizeOf(protocol.ReportReply)], std.mem.asBytes(&report_reply));
     if (test_usb_restart_mode and !test_usb_killed and childCountOf(sender) >= 2) {
+        // Delayed, not immediate: the device-list scenario's subscriber needs a
+        // window to enumerate and subscribe before the events start.
         test_usb_killed = true;
-        writeLine("device-manager: test mode: killing the reporter\n", .{});
-        _ = system.kill(sender);
+        test_kill_pid = sender;
+        test_kill_due_ns = system.clock() + 2_000_000_000;
+        _ = system.timerOnce(manager_endpoint, 2100);
     }
     return @sizeOf(protocol.ReportReply);
 }
@@ -386,6 +419,40 @@ fn onChildRemoved(message: []const u8, reply: []u8, sender: u32) usize {
     return @sizeOf(protocol.ReportReply);
 }
 
+/// An application asked for the tree: the mirror, as a header plus entries.
+fn onEnumerate(reply: []u8) usize {
+    var count: u32 = 0;
+    var offset: usize = @sizeOf(protocol.EnumerateReply);
+    for (&children) |*child| {
+        if (!child.used) continue;
+        if (offset + @sizeOf(protocol.ChildEntry) > reply.len) break;
+        const entry = protocol.ChildEntry{ .parent = child.parent, .bus_address = child.bus_address, .identity = child.identity };
+        @memcpy(reply[offset..][0..@sizeOf(protocol.ChildEntry)], std.mem.asBytes(&entry));
+        offset += @sizeOf(protocol.ChildEntry);
+        count += 1;
+    }
+    const header = protocol.EnumerateReply{ .status = 0, .count = count };
+    @memcpy(reply[0..@sizeOf(protocol.EnumerateReply)], std.mem.asBytes(&header));
+    return offset;
+}
+
+/// An application subscribed: its endpoint arrived as the call's capability.
+fn onSubscribe(reply: []u8, capability: ?runtime.ipc.Handle) usize {
+    var status: i32 = -1;
+    if (capability) |handle| {
+        for (&subscribers) |*slot| {
+            if (slot.* == null) {
+                slot.* = handle;
+                status = 0;
+                break;
+            }
+        }
+    }
+    const report_reply = protocol.ReportReply{ .status = status };
+    @memcpy(reply[0..@sizeOf(protocol.ReportReply)], std.mem.asBytes(&report_reply));
+    return @sizeOf(protocol.ReportReply);
+}
+
 fn onNotification(badge: u64) void {
     if (badge & runtime.ipc.notify_exit_bit != 0) {
         const dead: u32 = @intCast(badge & ~(runtime.ipc.notify_badge_bit | runtime.ipc.notify_exit_bit));

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

@@ -51,8 +51,9 @@ fn awaitChildExit(endpoint: runtime.ipc.Handle) u32 {
 /// The harness-run child of the signals test: echoes requests, logs the two
 /// signals it handles. Terminate makes run() return, and returning from main is
 /// the clean exit the parent reads as ExitReason.exited.
-fn echo(message: []const u8, reply: []u8, sender: u32) usize {
+fn echo(message: []const u8, reply: []u8, sender: u32, capability: ?runtime.ipc.Handle) usize {
     _ = sender;
+    _ = capability;
     const n = @min(message.len, reply.len);
     @memcpy(reply[0..n], message[0..n]);
     return n;

system/services/vfs/vfs.zig

@@ -91,7 +91,8 @@ fn releaseClientHandles(client: u32) void {
 }
 
 /// Handle one request from `sender`; write the reply into `out`, return its length.
-fn handle(message: []const u8, out: []u8, sender: u32) usize {
+fn handle(message: []const u8, out: []u8, sender: u32, capability: ?runtime.ipc.Handle) usize {
+    _ = capability;
     if (message.len < protocol.request_size) return fail(out);
     const request = std.mem.bytesToValue(protocol.Request, message[0..protocol.request_size]);
     const payload = message[protocol.request_size..];