add macOS support by alfaoz · Pull Request #57 · nowrep/dualsensectl

alfaoz · 2026-03-20T20:12:18Z

summary

adds macOS support to dualsensectl. all changes are behind #ifdef __APPLE__ — linux code is untouched.

replaces libudev with IOKit/CoreFoundation for device monitoring (IOHIDManager callbacks for hotplug)
swaps hidapi-hidraw for the generic hidapi (uses IOKit backend on macOS)
handles macOS IOKit stripping the HID report ID byte from read buffers
adds nanosleep compat shim for thrd_sleep
platform-conditional meson.build (auto-detects macOS vs Linux dependencies)

tested on

macOS sequoia 15, arm64 (apple silicon)
dualsense (CFL-ZCT1) over USB and bluetooth

feature	bluetooth	usb
device listing	yes	yes
battery	yes	yes
firmware info	yes	yes
lightbar	yes	yes
player leds	yes	yes
adaptive triggers (all modes)	yes	yes
mic led	yes	yes
speaker routing	—	yes
volume	—	yes
microphone on/off	—	yes
monitor (plug/unplug)	yes	yes

build on macOS

brew install hidapi meson
meson setup build
meson compile -C build

replace libudev with IOKit/CoreFoundation for device monitoring, swap hidapi-hidraw for the generic hidapi (IOKit backend), handle macOS-specific HID report ID stripping, and add nanosleep compat for thrd_sleep. all changes behind #ifdef __APPLE__ — linux code is untouched. tested on macOS sequoia (arm64) over both USB and bluetooth.

Copilot

Pull request overview

This PR adds macOS support to dualsensectl by introducing macOS-specific HID handling and device monitoring while keeping Linux behavior intact via platform conditionals.

Changes:

Adds platform-conditional Meson dependencies to switch between Linux (hidapi-hidraw + libudev) and macOS (hidapi + IOKit/CoreFoundation frameworks).
Introduces macOS-specific handling for HID input report layout differences (report ID stripping).
Implements a macOS monitor command backend using IOHIDManager callbacks.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 5 comments.

File	Description
meson.build	Adds OS-conditional dependencies and linker args for macOS vs Linux.
main.c	Adds macOS-specific includes/shims, input report parsing adjustments, and an IOKit-based monitor implementation.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

- fix potential NULL dereference in serial number fprintf - handle macOS report ID stripping in update command (not just battery) - fix CF object leak on IOHIDManagerOpen error path - use bracket syntax for meson array appends

Copilot

Pull request overview

Copilot reviewed 2 out of 2 changed files in this pull request and generated 2 comments.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

alfaoz · 2026-03-25T11:06:30Z

hello!

alfaoz · 2026-06-01T20:46:14Z

bump

egormanga

My two cents; never used a Mac.

egormanga · 2026-08-07T23:44:12Z

 		ds_report = (struct dualsense_input_report *)&data[1];
+#ifdef __APPLE__
+	} else if (res == DS_INPUT_REPORT_USB_SIZE - 1) {
+		/* macOS IOKit strips report ID */


Maybe replace the entire branch on Darwin instead of falling back?

in my testing macos always stripped the report id, but i kept the generic branch so a build where hidapi doesn't strip still works. i'd rather not hard-assume iokit behavior here lol

egormanga · 2026-08-07T23:46:38Z

+        size_t len = strlen(buf);
+        if (len == 17) {
+            /* Replace dashes with colons if needed, uppercase */
+            for (int i = 0; i < 17; i++) {


Trap for future refactoring!

Suggested change

for (int i = 0; i < 17; i++) {

for (int i = 0; i < len; i++) {

fair! will change.

egormanga · 2026-08-07T23:46:45Z

+                if (buf[i] == '-') buf[i] = ':';
+                serial_number[i] = toupper(buf[i]);
+            }
+            serial_number[17] = '\0';


Suggested change

serial_number[17] = '\0';

serial_number[len] = '\0';

egormanga · 2026-08-07T23:47:12Z

+static void iokit_device_added(void *context, IOReturn result, void *sender, IOHIDDeviceRef device)
+{
+    (void)context; (void)result; (void)sender;
+    char serial_number[18] = "00:00:00:00:00:00";


Suggested change

char serial_number[18] = "00:00:00:00:00:00";

char serial_number[] = "00:00:00:00:00:00";

same here. will change.

egormanga · 2026-08-07T23:47:35Z

+static void iokit_device_removed(void *context, IOReturn result, void *sender, IOHIDDeviceRef device)
+{
+    (void)context; (void)result; (void)sender;
+    char serial_number[18] = "00:00:00:00:00:00";


Suggested change

char serial_number[18] = "00:00:00:00:00:00";

char serial_number[] = "00:00:00:00:00:00";

egormanga · 2026-08-07T23:48:51Z

+
+    IOReturn ret = IOHIDManagerOpen(manager, kIOHIDOptionsTypeNone);
+    if (ret != kIOReturnSuccess) {
+        fprintf(stderr, "Failed to open IOHIDManager: 0x%x\n", ret);


Suggested change

fprintf(stderr, "Failed to open IOHIDManager: 0x%x\n", ret);

fprintf(stderr, "Failed to open IOHIDManager: %#04x\n", ret);

will change!

nowrep · 2026-08-08T09:24:40Z

Can you please put all the macos specific code to a new file (eg. macos.c)?

egormanga · 2026-08-09T00:49:57Z

@nowrep would a separate file really be more maintainable than preproc guards? I think I'm more with @alfaoz on that

alfaoz · 2026-08-09T02:53:00Z

Can you please put all the macos specific code to a new file (eg. macos.c)?

well most of the macos stuff is a couple lines inside shared functions, splitting those out would need either duplicating them or a short shim, do we want that?.

the monitor code is self contained though, i can move that plus the serial helper into macos.c and leave the small ifdefs where they are. would that work?

alfaoz · 2026-08-14T15:32:41Z

moved the self-contained monitor implementation to macos.c; retained the small platform-specific report branches in main.c

egormanga

All minor; I have no issue with this getting merged even as-is.
@nowrep FYI

egormanga · 2026-09-12T12:28:16Z

+                return;
+            }


Can we fail-fast instead? Missed that on my initial review, but a proper (and functionally equivalent) practice would be if (len < 17) continue; and flattening the rest.

agreed, will flatten it to one early-out and keep the dash/uppercase pass at top level.

egormanga · 2026-09-12T12:30:26Z

+    (void)result;
+    (void)sender;


Perhaps start these names with an underscore as well?

there's no underscore convention anywhere in main.c. and the (void) casts are what actually silence -Wunused-parameter portably. if you feel strongly about ts, it's a two-line rename, just say so.

There isn't a single (void) in there, either. But it's your call.

egormanga · 2026-09-12T12:31:20Z

+/* Portable thrd_sleep replacement using nanosleep on macOS */
+#ifdef __APPLE__
+#define thrd_sleep(ts, rem) nanosleep((ts), (rem))
+#endif


Perhaps move that to macos.h too?

yep, makes sense will move it there rq

egormanga · 2026-09-12T12:34:03Z

+            }
+        }
+    }
+    strncpy(serial_number, "00:00:00:00:00:00", 18);


nit: this assumes serial_number size is enough which it may or may not be depending on API/ABI.

will pass the buffer size in explicitly and snprintf into it

egormanga · 2026-09-12T12:35:49Z

+    (void)sender;
+
+    struct monitor_context *monitor = context;
+    char serial_number[] = "00:00:00:00:00:00";


Why do we need a default if that gets filled in in the get_serial_from_hid_device()?

Also specify an explicit size to avoid a possible future bounds fault.

will make it a plain char serial_number[18] and hand sizeof() to the getter, which also covers your strncpy note above

alfaoz · 2026-09-12T12:49:45Z

will push a quick fix

egormanga · 2026-09-12T20:32:59Z

 };

-static void get_serial_from_hid_device(IOHIDDeviceRef device, char serial_number[18])
+static void get_serial_from_hid_device(IOHIDDeviceRef device, char *serial_number, size_t size)


Note you can char serial_number[static size] for sanity & documentation purposes as well

egormanga · 2026-09-12T20:33:36Z

-static void get_serial_from_hid_device(IOHIDDeviceRef device, char serial_number[18])
+static void get_serial_from_hid_device(IOHIDDeviceRef device, char *serial_number, size_t size)
 {
+    char buf[64];


why 64 btw? a magic number

missed it lol was as headroom

Copilot AI review requested due to automatic review settings March 20, 2026 20:12

Copilot started reviewing on behalf of alfaoz March 20, 2026 20:12 View session

Copilot AI reviewed Mar 20, 2026

View reviewed changes

Comment thread meson.build Outdated

Comment thread meson.build Outdated

Comment thread main.c Outdated

Comment thread main.c

Comment thread main.c Outdated

address review feedback

f907d2f

- fix potential NULL dereference in serial number fprintf - handle macOS report ID stripping in update command (not just battery) - fix CF object leak on IOHIDManagerOpen error path - use bracket syntax for meson array appends

alfaoz requested a review from Copilot March 21, 2026 09:35

Copilot started reviewing on behalf of alfaoz March 21, 2026 09:36 View session

Copilot AI reviewed Mar 21, 2026

View reviewed changes

Comment thread meson.build

Comment thread main.c Outdated

egormanga suggested changes Aug 7, 2026

View reviewed changes

alfaoz added 2 commits August 9, 2026 06:00

apply review suggestions

68e3881

move macOS monitor to separate file

ff0939c

address review follow-ups

b6123cc

egormanga approved these changes Sep 12, 2026

View reviewed changes

quick minor fixes and refactors

f1bcda8

egormanga approved these changes Sep 12, 2026

View reviewed changes

serial buffer: named length, static array param

1f402b5

	for (int i = 0; i < 17; i++) {
	for (int i = 0; i < len; i++) {

	char serial_number[18] = "00:00:00:00:00:00";
	char serial_number[] = "00:00:00:00:00:00";

	fprintf(stderr, "Failed to open IOHIDManager: 0x%x\n", ret);
	fprintf(stderr, "Failed to open IOHIDManager: %#04x\n", ret);

Conversation

alfaoz commented Mar 20, 2026

summary

tested on

build on macOS

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

alfaoz commented Mar 25, 2026

Uh oh!

alfaoz commented Jun 1, 2026

Uh oh!

egormanga left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

nowrep commented Aug 8, 2026

Uh oh!

egormanga commented Aug 9, 2026

Uh oh!

alfaoz commented Aug 9, 2026

Uh oh!

alfaoz commented Aug 14, 2026

Uh oh!

egormanga left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment