Bug 99881

Same chip in MacBookPro11,4 and MacBookPro11,5.

I'm sorry, not sure I understand... I am talking about the bluetooth keyboard and mouse, what do they have to do with MBP?

Hi Aayla,

Sorry if I misunderstood you. I thought you were talking about the overall Bluetooth support as the ticket is registered to the 'Bluetooth' component.

When Bluetooth is not enabled, the use of a Bluetooth keyboard and mouse (whatever the brand) is accomplished through the bluetooth chip directly with some special Apple circuitry and the keyboard and mouse are seen by Linux as a USB keyboard and a USB mouse. This is done to have early keyboard and mouse support during boot (for instance you use the arrows keys in Refind).

However the keyboard and mouse are seen like standard USB keyboard, and I'm not sure it can be recognized as apple hardware.

Anyway, as I have a apple bluetooth keyboard and trackpad, I experience the same problem. As the Bluetooth chip is not supported, I only have basic support of the keyboard and trackpad, so no two finger scrolling nor special keys (as Fn is always on like on a PC keyboard).

I see... I read somewhere that the problem is indeed in the fact that the bluetooth controller is run in HID mode and not in HCI but I couldn't get it to switch (with udev's hid2hci)...

Created attachment 188821 [details]
Patch for bluetooth

This hacky patch helps, after that bluetooth works like a charm.

I took it from here:
http://www.spinics.net/lists/linux-bluetooth/msg64098.html

Unfortunately that did not work - bluetooth device is still detected and listed as using btusb driver but it simply receives nothing from the keyboard or mouse...

It is strange, it works for me as expected, I could pair both my keyboard and mouse, and I can make the tethering network from my smartphone.

Thethering from the smartphone is handled by a different driver and it works for me as well (has always worked). But this keyboard... To be honest, I'm sick of trying and am tempted to just close the bug and either stick to the internal laptop one (as I am at the moment) or just get a normal wireless keyboard.

P.S. What kernel version are you using? I'm still on 4.0.5 due to issues with the graphics that appeared in later versions and haven't been resolved.

I forgot to mention, I do not use Apple devices, my mouse is Logitech M555b, and my keyboard is Logitech K810.

I am using the latest released kernel 4.2.1.
I do not experience any issues with the graphics, moreover, I do not need to use kernel command line arguments like i915.enable_ips=0 and intel_iommu=igfx_off anymore as I did in 4.0 and 4.1.

Oh well, that may explain why, I'm pretty sure any other BT keyboard would work although I haven't tried.
Thanks, I'll give the latest version a go, hopefully the graphics would work!

So using kernel 4.2.2 - same thing:
Using unmodified souurces the BT controller is picked up by usbhid driver with which the apple BT keyboard works but not the FN button (and mouse scrolling does not work on the magic mouse), which was the reason I started this bug.
Using the patch from comment 5, the BT controller is picked up by btusb but then the apple BT keyboard does not work AT ALL (tried both building btusb into the kernel as well as as a module - no difference).

Created attachment 203441 [details]
Binding of MM2 to hid-magicmouse

This only binds the magic mouse 2 to the hid-magicmouse module, scrolling still doesn't work but it is a first step. I will pair with one of my macs and start trying to capture the scrolling. First guess would indicate that scrolling, force touch, and other gestures are all sent over BTLE.

I have some details on the difference between the Fn key on Apple Magic Keyboard (the new one) and Apple Wireless Keyboard (the old one). Both are bluetooth keyboards. I used Wireshark to sniff HID packets over Bluetooth.

In the old keyboard, Fn key sends a HID packet with protocol code 0x11 (unknown to Wireshark) and a single byte payload of 0x10 for key down and 0x00 for key up.

In the new keyboard, Fn key sends a HID packet with protocol code 0x01 (ie keyboard) and value of the *7th* key set to 0x02 for key down, and 0x00 for key up.

Note the 7th key. Wireshark only shows up to 6 keycodes in a keyboard HID packet which I assume is the standard, but the new keyboard sends a 7th keycode for the Fn key.

Hope this helps.

I would LOVE to get my hands on that wireshark. Is it full bluetooth or BLE by any chance?

Thanks,

Michael Schlies

(In reply to Mansour Behabadi from comment #14)
> I have some details on the difference between the Fn key on Apple Magic
> Keyboard (the new one) and Apple Wireless Keyboard (the old one). Both are
> bluetooth keyboards. I used Wireshark to sniff HID packets over Bluetooth.
> 
> In the old keyboard, Fn key sends a HID packet with protocol code 0x11
> (unknown to Wireshark) and a single byte payload of 0x10 for key down and
> 0x00 for key up.
> 
> In the new keyboard, Fn key sends a HID packet with protocol code 0x01 (ie
> keyboard) and value of the *7th* key set to 0x02 for key down, and 0x00 for
> key up.
> 
> Note the 7th key. Wireshark only shows up to 6 keycodes in a keyboard HID
> packet which I assume is the standard, but the new keyboard sends a 7th
> keycode for the Fn key.
> 
> Hope this helps.

(In reply to Michael Schlies from comment #15)
> I would LOVE to get my hands on that wireshark. Is it full bluetooth or BLE
> by any chance?

I have the keyboard for another day or two and I can capture and upload any scenarios that helps this issue. Just let me know what you need.

I don't recall seeing any BLE stuff but I could be wrong.

Actually, do you happen to have captures from the mouse or trackpad? I think I got the Fn key working this afternoon.

Thanks!

> Actually, do you happen to have captures from the mouse or trackpad?

Unfortunately I don't own those.

> I think I got the Fn key working this afternoon.

Awesome!

I have the 2015 magic mouse and keyboard and I would like to help, but I don't really know how to capture HID packets. I've only used the likes of showkey. I also don't currently have access to them, since I'm away from my office until at some point next week, but if someone refers me to some links on how I can help with that, I'd be glad.

I have all three devices, I just can't say I have done bluetooth captures of them and I think the trackpad and mouse are going to be the most difficult of these to setup.

Hello Guys.

I have set a bug bounty over at bountysource for a fix for this issue.

https://www.bountysource.com/issues/33063419-2015-apple-bluetooth-keybord-and-magic-mouse-support

Thank you for your guys help!

Oliver

I've got Packet Logger files for the magic mouse 2 if it helps but I am not sure what of the logs should be uploaded. I am a noob to posting here. Can i post a link to the whole file or is that frowned upon?

Created attachment 213591 [details]
OSX Packet Logger Magic Mouse 2

Disregard my question.
So I came into work on Monday to find my right click not working on my OEM mouse. I swapped with like 3 other mice and all had other issues. All hardware related. Amazon Prime Now had a MM2 so i figured i would try it out. Now I'm stuck with no scrolling. Hopefully these packet logs will help whoever can fix this. If i were more versed in this sort of thing I would give it a shot. I will also be throwing in on the bounty to help maybe speed this along.

While the packet captures are otherwise appreciated I took some very tightly scoped ones (roughly 10 captures for each of the 3 devices).

I swapped my mm2 with a friend for a mm1 and it works as described by most posts online. I can't deal with it so i returned it for a trackpad 2. I am on xubuntu 16.04 lts. If anyone wants me to test anything or grab any logs let me know. Currently I have movement a left button click. Nothing else.

Any Progress on this? I hope to have a working touchpad soon.

Michael, have you continued to work on this issue ?

Is there any progress ?

Thanks

Any progress ?

No progress here I guess?

I just tried a new set of apple bluetooth keyboard and magic mouse and problems still persist (scroll, Fn, etc).

Just tested connecting a Magic Keyboard A1644 to Debian 9, kernel 4.9.0-3-amd63, over Bluetooth. It uses the driver hid_generic, not hid_apple, and the "fn" keys don't work.

I added a bounty on bountysource: https://www.bountysource.com/issues/33063419-2015-apple-bluetooth-keybord-and-magic-mouse-support

So what the pid and vid over bluetooth, you should be able to figure this out in bluetoothd logs, e.g: 

bluetoothd[30701]: profiles/input/hog.c:hog_device_accept() name=Arc Touch Mouse SE vendor=0x45E, product=0x7F3, version=0x1

If it helps, from bluetoothctl:


info:
Device
	Name: Magic Mouse 2
	Alias: Magic Mouse 2
	Class: 0x002580
	Icon: input-mouse
	Paired: yes
	Trusted: yes
	Blocked: no
	Connected: yes
	LegacyPairing: no
	UUID: Human Interface Device... (00001124-0000-1000-8000-00805f9b34fb)
	UUID: PnP Information           (00001200-0000-1000-8000-00805f9b34fb)
	Modalias: bluetooth:v004Cp0269d0064
	ManufacturerData Key: 0x004c
	ManufacturerData Value: 0x41
	ManufacturerData Value: 0x75
	ManufacturerData Value: 0x67
	ManufacturerData Value: 0x20
	ManufacturerData Value: 0x33
	ManufacturerData Value: 0x31
	ManufacturerData Value: 0x20
	ManufacturerData Value: 0x32
	ManufacturerData Value: 0x30
	ManufacturerData Value: 0x31
	ManufacturerData Value: 0x35
	ManufacturerData Value: 0x00
	ManufacturerData Value: 0x20
	ManufacturerData Value: 0x31
	ManufacturerData Value: 0x37
	ManufacturerData Value: 0x3a
	ManufacturerData Value: 0x31
	ManufacturerData Value: 0x31
	ManufacturerData Value: 0x3a
	ManufacturerData Value: 0x31
	ManufacturerData Value: 0x37
	ManufacturerData Value: 0x00

Created attachment 257447 [details]
attachment-8611-0.html

I spent a bit of time on this a few months ago and my conclusion was that
it's not just a case of adding the ids to the existing hid-apple driver.
I've got some Wireshark captures from Mac OS that showed Fn key emitting a
keycode that is different from other magic keyboard models.  I also suspect
that Magic Mouse requires a command to be sent to it that enables gestures
as there is no bluetooth tx from the mouse for gestures when running with
hid-generic.  I use both of these devices daily under Ubuntu 17.04 on a MBP
2015.  I'll spend a bit of time and see if I can get the Fn key mapped in
hid-apple.  If I can be of assistance to anyone for testing please let me
know.

On 10 July 2017 at 18:50, <bugzilla-daemon@bugzilla.kernel.org> wrote:

> https://bugzilla.kernel.org/show_bug.cgi?id=99881
>
> --- Comment #33 from Nikos Roussos (comzeradd@fedoraproject.org) ---
> If it helps, from bluetoothctl:
>
>
> info:
> Device
>         Name: Magic Mouse 2
>         Alias: Magic Mouse 2
>         Class: 0x002580
>         Icon: input-mouse
>         Paired: yes
>         Trusted: yes
>         Blocked: no
>         Connected: yes
>         LegacyPairing: no
>         UUID: Human Interface Device... (00001124-0000-1000-8000-
> 00805f9b34fb)
>         UUID: PnP Information           (00001200-0000-1000-8000-
> 00805f9b34fb)
>         Modalias: bluetooth:v004Cp0269d0064
>         ManufacturerData Key: 0x004c
>         ManufacturerData Value: 0x41
>         ManufacturerData Value: 0x75
>         ManufacturerData Value: 0x67
>         ManufacturerData Value: 0x20
>         ManufacturerData Value: 0x33
>         ManufacturerData Value: 0x31
>         ManufacturerData Value: 0x20
>         ManufacturerData Value: 0x32
>         ManufacturerData Value: 0x30
>         ManufacturerData Value: 0x31
>         ManufacturerData Value: 0x35
>         ManufacturerData Value: 0x00
>         ManufacturerData Value: 0x20
>         ManufacturerData Value: 0x31
>         ManufacturerData Value: 0x37
>         ManufacturerData Value: 0x3a
>         ManufacturerData Value: 0x31
>         ManufacturerData Value: 0x31
>         ManufacturerData Value: 0x3a
>         ManufacturerData Value: 0x31
>         ManufacturerData Value: 0x37
>         ManufacturerData Value: 0x00
>
> --
> You are receiving this mail because:
> You are on the CC list for the bug.
>

Either way there should be a hid driver to driver all of this, be it in hid-apple or a new driver has nothing to do with Bluetooth really.

dmesg log when I connect the Magic Keyboard A1644 (mac address replaced with x):

[144510.510615] hid-generic 0005:004C:0267.0006: unknown main item tag 0x0
[144510.510687] input: Magic Keyboard as /devices/pci0000:00/0000:00:14.0/usb1/1-9/1-9:1.0/bluetooth/hci0/hci0:256/0005:004C:0267.0006/input/input23
[144510.510934] hid-generic 0005:004C:0267.0006: input,hidraw1: BLUETOOTH HID v0.66 Keyboard [Magic Keyboard] on xx:xx:xx:xx:xx:xx

Output from bluetoothctl info:

[Magic Keyboard]# info 2C:33:61:E5:1A:F8
Device 2C:33:61:E5:1A:F8
	Name: Magic Keyboard
	Alias: Magic Keyboard
	Class: 0x000540
	Icon: input-keyboard
	Paired: yes
	Trusted: yes
	Blocked: no
	Connected: yes
	LegacyPairing: no
	UUID: Human Interface Device... (00001124-0000-1000-8000-00805f9b34fb)
	UUID: PnP Information           (00001200-0000-1000-8000-00805f9b34fb)
	Modalias: bluetooth:v004Cp0267d0066

---

Let me know if I can do anything to help.

I realize I hid the MAC address from the dmesg output, then forgot to do the same for the bluetoothctl output. I can't find an edit button and it probably doesn't even matter, so what ever :)

(In reply to Tomas Sandven from comment #36)
> [144510.510615] hid-generic 0005:004C:0267.0006: unknown main item tag 0x0
> [144510.510687] input: Magic Keyboard as

I have same keyboard (device Id 0x0267). What is interesting here is that vendor ID is 0x004c when connect via bluetooth, but 0x05ac when connected via USB.

It looks that 0x004c seems to be vendor id for Apple in some bluetooth spec: https://www.bluetooth.com/specifications/assigned-numbers/company-identifiers

A hack to fix it would be something like:

````
diff --git a/drivers/hid/hid-apple.c b/drivers/hid/hid-apple.c
index 25b7bd56ae11..528ec847cc03 100644
--- a/drivers/hid/hid-apple.c
+++ b/drivers/hid/hid-apple.c
@@ -472,6 +472,8 @@ static const struct hid_device_id apple_devices[] = {
                .driver_data = APPLE_NUMLOCK_EMULATION | APPLE_HAS_FN },
        { HID_USB_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_MAGIC_KEYBOARD_ANSI),
                .driver_data = APPLE_HAS_FN },
+       { HID_USB_DEVICE(0x004c, USB_DEVICE_ID_APPLE_MAGIC_KEYBOARD_ANSI),
+               .driver_data = APPLE_HAS_FN },
        { HID_USB_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_WELLSPRING_ANSI),
                .driver_data = APPLE_HAS_FN },
        { HID_USB_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_WELLSPRING_ISO),
````

I suppose we need to define the vendor id 0x004c properly in drivers/hid/hid-ids.h but I don't know how? as BLUETOOTH_VENDOR_ID_APPLE?

(In reply to ncopa from comment #38)
> (In reply to Tomas Sandven from comment #36)
> > [144510.510615] hid-generic 0005:004C:0267.0006: unknown main item tag 0x0
> > [144510.510687] input: Magic Keyboard as
> 
> I have same keyboard (device Id 0x0267). What is interesting here is that
> vendor ID is 0x004c when connect via bluetooth, but 0x05ac when connected
> via USB.
> 
> It looks that 0x004c seems to be vendor id for Apple in some bluetooth spec:
> https://www.bluetooth.com/specifications/assigned-numbers/company-identifiers
> 
> A hack to fix it would be something like:
> 
> ````
> diff --git a/drivers/hid/hid-apple.c b/drivers/hid/hid-apple.c
> index 25b7bd56ae11..528ec847cc03 100644
> --- a/drivers/hid/hid-apple.c
> +++ b/drivers/hid/hid-apple.c
> @@ -472,6 +472,8 @@ static const struct hid_device_id apple_devices[] = {
>                 .driver_data = APPLE_NUMLOCK_EMULATION | APPLE_HAS_FN },
>         { HID_USB_DEVICE(USB_VENDOR_ID_APPLE,
> USB_DEVICE_ID_APPLE_MAGIC_KEYBOARD_ANSI),
>                 .driver_data = APPLE_HAS_FN },
> +       { HID_USB_DEVICE(0x004c, USB_DEVICE_ID_APPLE_MAGIC_KEYBOARD_ANSI),
> +               .driver_data = APPLE_HAS_FN },
>         { HID_USB_DEVICE(USB_VENDOR_ID_APPLE,
> USB_DEVICE_ID_APPLE_WELLSPRING_ANSI),
>                 .driver_data = APPLE_HAS_FN },
>         { HID_USB_DEVICE(USB_VENDOR_ID_APPLE,
> USB_DEVICE_ID_APPLE_WELLSPRING_ISO),
> ````
> 
> I suppose we need to define the vendor id 0x004c properly in
> drivers/hid/hid-ids.h but I don't know how? as BLUETOOTH_VENDOR_ID_APPLE?

I had tried exactly that and had no luck.  I just assumed the keyboard has two completely different modes of operation - one for USB and one for BT.  In USB mode the keyboard operates flawlessly.

Created attachment 258189 [details]
Patch to makes bluetooth Magic Keyboard use hid-apple driver instead of hid-generic

This patch makes the Magic Keyboard (vendor id 0x004c, device id 0x0267) use hid-apple instead of hid-generic.

The Fn key does not work yet though.

Created attachment 258205 [details]
Fix Fn key on bluetooth Apple Magic Keyboard

This patch fixes the Fn Key on Apple Magic Keyboard when connected via bluetooth.

I can confirm the patch from ncopa makes the Fn key and Media/Function keys work on the Apple Magic Keyboard 2 (A1644).

@ncopa What's the significance of usage->hid hack for this model?  I'm not familiar with the hid api and was missing that from my previous attempt at support for this model.

(In reply to Guy Thouret from comment #42)
> I can confirm the patch from ncopa makes the Fn key and Media/Function keys
> work on the Apple Magic Keyboard 2 (A1644).
> 
> @ncopa What's the significance of usage->hid hack for this model?  I'm not
> familiar with the hid api and was missing that from my previous attempt at
> support for this model.

I don't know to be honest and this is my first meeting with hid api too. It appears to be a vendor specific quirk to set up the Fn key mapping. Apparently there was some previous quirk for older models that has been there for long time (according git log).

I discovered it while doing some printf debugging around things that had KEY_FN and analyzing the git log of hid-input.c.

I can also confirm the patch from ncopa makes the Fn key and Media/Function keys work on the Apple Magic Keyboard 2 (A1644).

and when do you think we could have this patch in the mainstream ?

Thank you

I sent the patch upstream:

http://www.spinics.net/lists/linux-input/msg53006.html

It seems this is now broken again, I have this problem in 4.17-0-rc4

My Fn-keys are not working over Bluetooth, they are working over USB.

USB uses driver: usbhid
Bluetooth uses driver: hid-generic
Not sure if this is part of the problem.

Initially I described the problem here:
https://askubuntu.com/questions/1033788/apple-keyboard-fn-not-working-with-bluetooth-but-works-with-usb

I am also noticing the same issue as Stijn Martens. When my keyboard is connected via USB it works as expected, but over bluetooth, hid_apple isn't being used.

John

Patch in comment #41 should fix it. It is not applied to git yet. There have been no comment on the patch i sent either.

I tried manually applying the patch on a newer version and compiling it. But it did not work out. First time trying to change & compile Linux myself.

The patch may need to be updated before it can be accepted.

(In reply to ncopa from comment #49)
> Patch in comment #41 should fix it. It is not applied to git yet. There have
> been no comment on the patch i sent either.

I suppose you should have added hid-apple maintainers to Cc, without it maintainer like Dmitriy just skip this patch. (it's just from my experience)
Also, please, update patch to modern kernel version probably I can write something to maillist too.

This is partially fixed with
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/drivers/hid?id=ee345492437043a79db058a3d4f029ebcb52089a

However, the Fn key still does not work. Apparently it still picks hid-generic driver:

[107893.090949] hid-generic 0005:004C:0267.0007: unknown main item tag 0x0
[107893.091122] input: Magic Keyboard as /devices/pci0000:00/0000:00:14.0/usb1/1-8/1-8:1.0/bluetooth/hci0/hci0:256/0005:004C:0267.0007/input/input24
[107893.091794] hid-generic 0005:004C:0267.0007: input,hidraw3: BLUETOOTH HID v0.67 Keyboard [Magic Keyboard] on a4:34:d9:5a:c7:14

It's working for me:


[   22.260447] apple 0005:004C:0267.0007: unknown main item tag 0x0
[   22.260626] input: FabsKeyboard as /devices/pci0000:00/0000:00:14.0/usb1/1-3/1-3:1.2/bluetooth/hci0/hci0:12/0005:004C:0267.0007/input/input18
[   22.261181] apple 0005:004C:0267.0007: input,hidraw6: BLUETOOTH HID v0.67 Keyboard [FabsKeyboard] on a0:99:9b:19:24:18


Fn key is working, using bluetooth.

Now I'm able to use:

$ cat /etc/modprobe.d/hid_apple.conf
options hid_apple swap_opt_cmd=1
options hid_apple fnmode=2
options hid_apple iso_layout=0


Which before only worked when cabled.

It did not work on 4.14.76 for me.
(In reply to Fabio Falci from comment #53)
> It's working for me:
> 
> 
> [   22.260447] apple 0005:004C:0267.0007: unknown main item tag 0x0
> [   22.260626] input: FabsKeyboard as
> /devices/pci0000:00/0000:00:14.0/usb1/1-3/1-3:1.2/bluetooth/hci0/hci0:12/
> 0005:004C:0267.0007/input/input18
> [   22.261181] apple 0005:004C:0267.0007: input,hidraw6: BLUETOOTH HID v0.67
> Keyboard [FabsKeyboard] on a0:99:9b:19:24:18
> 
> Fn key is working, using bluetooth.

What kernel is it? It did not work for me on 4.14.76.

$ uname -a
Linux fabs 4.18.14-arch1-1-ARCH #1 SMP PREEMPT Sat Oct 13 13:42:37 UTC 2018 x86_64 GNU/Linux

I think an additional patch is needed for kernel pre 4.16.

(In reply to ncopa from comment #56)
> I think an additional patch is needed for kernel pre 4.16.

I mean pre 4.17. So for 4.16 kernels and older.

Is there any news on this? Was the patch for the fn-key merged in a later kernel?

Is this issue still being addressed?

I'm on 4.15.0, and the function key doesn't work over Bluetooth.

Still experiencing this problem over bluetooth on

Linux z 5.3.0-23-generic #25-Ubuntu SMP Tue Nov 12 09:22:33 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux

Was the patch ever merged? Or was it merged for v4.x only?

I can confirm that it has worked, as my previous messages say, at least on 4.18.14.


However it stopped working at some point, unfortunately, I was switching between laptops at the time and don't have access to my previous one where it worked. On my new laptop, it doesn't work. 


So not sure what version it has stopped working but I'd say that it worked on 5.x.

Just an update, it *does* work on another machine with a different keyboard, running:

Linux y 5.3.0-19-generic #20-Ubuntu SMP Fri Oct 18 09:04:39 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux

I downgraded the -23 machine to -19 (debian revision) without change.

Could it be firmware differences in the two keyboards?

I have a version 1.0 identifying Apple Magic Keyboard 2.  From what I can tell (at a quick glance) it's descriptor does NOT match what the kernel is expecting.

The kernel is expecting either 0xFF000003 or 0x00FF0003 for Usage HID.  From my reading of the description it seems like it would be: 0xFF010003.

Seems we need at add another Usage HID variant?

Relevant descriptor dump:

```
     Report Count (1),
        Report Size (1),
        Logical Minimum (0),
        Logical Maximum (1),
        Usage Page (Consumer),                      ; Consumer (0Ch)
        Usage (Eject),                              ; Eject (B8h, one-shot control)
        Input (Variable),
        Report Count (1),
        Report Size (1),
        Usage Page (FF01h),                         ; FF01h, vendor-defined
        Usage (03h),
        Input (Variable),
```

Sorry if this is formatted terribly.  Not matter what I do the Preview looks terrible and this is my first time posting here.

The firmware of my Magic Keyboard (A1644) got updated to version 0x0100 as reported by System Information on macOSX. The keyboard used to work fine in bluetooth mode with fn-mode behavior adjustable through /sys/module/hid_apple/parameters/fnmode. However, after the firmware update, the fn key no longer works as expected. Instead, pressing the fn alone turns the screen brightness low and it can no longer be configured via the fnmode parameter nor does it function in any other way.

Sure enough, reading out the report descriptor, I'm getting an extra key mapped to BrightnessDown:


$ cat /sys/kernel/debug/hid/0005\:004C\:0267.0005/rdesc

...

ff01.0003 ---> Key.BrightnessDown

...

Submitted a patch to the kernel: https://lkml.org/lkml/2020/1/29/61 potentially fixing https://bugzilla.kernel.org/show_bug.cgi?id=99881#c62 , https://bugzilla.kernel.org/show_bug.cgi?id=99881#c63 and https://bugzilla.kernel.org/show_bug.cgi?id=99881#c64 .

(In reply to Mansour Behabadi from comment #66)
> Submitted a patch to the kernel: https://lkml.org/lkml/2020/1/29/61
> potentially fixing https://bugzilla.kernel.org/show_bug.cgi?id=99881#c62 ,
> https://bugzilla.kernel.org/show_bug.cgi?id=99881#c63 and
> https://bugzilla.kernel.org/show_bug.cgi?id=99881#c64 .

Thank u so much. It's save my day.

Linux 5.2.1-050201-generic, this patch fix fn key for MLA22

This patch [1] is now available in mainline, stable and longterm releases.


[1]: https://bugzilla.kernel.org/show_bug.cgi?id=99881#c66

I'm running 5.6.14-300 in Fedora 32.
 
I'm currently having problems with the Magic Keyboard 2, but only when using Bluetooth. It works fine when connected via USB.
 
The problem is that it's not picking up the driver settings. The one I'm trying to set is `iso_layout`, as I'm having the typical <> keys wrongly mapped. No matter what option I set on /parameters/iso_layout, the behaviour is still the same. It's like the driver is not correctly loaded.
 
When I connect it via USB, I can see it reacting to the changes in /parameters/iso_layout correctly in real time, but not when using the Bluetooth connection.
 
dmesg | grep input

```
[   46.012620] input: Magic Keyboard as /devices/pci0000:00/0000:00:14.0/usb1/1-10/1-10:1.0/bluetooth/hci0/hci0:256/0005:004C:0267.0006/input/input27
[   46.013565] input: Magic Keyboard as /devices/pci0000:00/0000:00:14.0/usb1/1-10/1-10:1.0/bluetooth/hci0/hci0:256/0005:004C:0267.0006/input/input28
[   46.013867] hid-generic 0005:004C:0267.0006: input,hidraw5: BLUETOOTH HID v1.00 Keyboard [Magic Keyboard] on 4c:1d:96:7c:7b:70
[   46.067136] input: Magic Keyboard as /devices/pci0000:00/0000:00:14.0/usb1/1-10/1-10:1.0/bluetooth/hci0/hci0:256/0005:004C:0267.0006/input/input29
[   46.067530] apple 0005:004C:0267.0006: input,hidraw5: BLUETOOTH HID v1.00 Keyboard [Magic Keyboard] on 4c:1d:96:7c:7b:70

```