Compatibility with generic pro controllers

[Woazboat]

Hi, I installed the driver via the dkms module but connecting the controller via bluetooth fails with the following error messages in the kernel log

[ 8274.330816] nintendo 0005:057E:2009.0006: unknown main item tag 0x0
[ 8274.331071] nintendo 0005:057E:2009.0006: hidraw2: BLUETOOTH HID v80.01 Gamepad [Pro Controller] on xx:xx:xx:xx:xx:xx
[ 8276.544262] nintendo 0005:057E:2009.0006: controller MAC = 80:A5:03:61:63:25
[ 8277.054077] nintendo 0005:057E:2009.0006: Failed to set home LED dflt; ret=-110
[ 8277.054085] nintendo 0005:057E:2009.0006: Failed to create leds; ret=-110
[ 8277.054400] nintendo 0005:057E:2009.0006: probe - fail = -110
[ 8277.558071] leds 0005:057E:2009.0006:home: Setting an LED's brightness failed (-110)
[ 8278.062004] leds 0005:057E:2009.0006:player4: Setting an LED's brightness failed (-110)
[ 8278.566018] leds 0005:057E:2009.0006:player3: Setting an LED's brightness failed (-110)
[ 8279.070042] leds 0005:057E:2009.0006:player2: Setting an LED's brightness failed (-110)
[ 8279.574077] leds 0005:057E:2009.0006:player1: Setting an LED's brightness failed (-110)

(Same issue as previously reported here: nicman23/dkms-hid-nintendo#4 )

Edit: I'm using a generic pro controller

[rodolpheh]

I also have the same issue with a generic Pro Controller, here is the relevant dmesg messages :

[  175.268848] hid-generic 0005:057E:2009.0008: unknown main item tag 0x0
[  175.268961] input: Pro Controller as /devices/pci0000:00/0000:00:14.0/usb1/1-14/1-14:1.0/bluetooth/hci0/hci0:256/0005:057E:2009.0008/input/input46
[  175.280499] hid-generic 0005:057E:2009.0008: input,hidraw5: BLUETOOTH HID v0.01 Gamepad [Pro Controller] on xx:xx:xx:xx:xx:xx
[  175.341003] nintendo 0005:057E:2009.0008: unknown main item tag 0x0
[  175.341138] nintendo 0005:057E:2009.0008: hidraw5: BLUETOOTH HID v80.01 Gamepad [Pro Controller] on xx:xx:xx:xx:xx:xx
[  177.449401] nintendo 0005:057E:2009.0008: controller MAC = 62:E1:2A:CC:63:25
[  177.976426] nintendo 0005:057E:2009.0008: Failed to set home LED dflt; ret=-110
[  177.976428] nintendo 0005:057E:2009.0008: Failed to create leds; ret=-110
[  177.976641] nintendo 0005:057E:2009.0008: probe - fail = -110
[  178.509678] leds 0005:057E:2009.0008:home: Setting an LED's brightness failed (-110)
[  179.042939] leds 0005:057E:2009.0008:player4: Setting an LED's brightness failed (-110)
[  179.576166] leds 0005:057E:2009.0008:player3: Setting an LED's brightness failed (-110)
[  180.109424] leds 0005:057E:2009.0008:player2: Setting an LED's brightness failed (-110)
[  180.642676] leds 0005:057E:2009.0008:player1: Setting an LED's brightness failed (-110)
[  180.642772] nintendo: probe of 0005:057E:2009.0008 failed with error -110

Let me know if you need more information.

[Woazboat]

Okay, I think I know what the actual problem is. The generic chinese pro controller clone I'm using doesn't actually have an led under the home button unlike the real pro controller.

https://www.aliexpress.com/item/4000571346668.html

[Woazboat]

Connecting to the controller works after simply disabling the home LED in joycon_leds_create(), however none of the inputs seem to be recognized.

/* configure the home LED */
if (false && jc_type_has_right(ctlr)) {
//...

[13580.885490] hid-generic 0005:057E:2009.0003: unknown main item tag 0x0
[13580.885725] input: Pro Controller as /devices/pci0000:00/0000:00:14.0/usb2/2-4/2-4:1.0/bluetooth/hci0/hci0:256/0005:057E:2009.0003/input/input28
[13580.886100] hid-generic 0005:057E:2009.0003: input,hidraw2: BLUETOOTH HID v0.01 Gamepad [Pro Controller] on xx:xx:xx:xx:xx:xx
[13580.891310] hid_nintendo: loading out-of-tree module taints kernel.
[13580.891409] hid_nintendo: module verification failed: signature and/or required key missing - tainting kernel
[13580.944392] nintendo 0005:057E:2009.0003: unknown main item tag 0x0
[13580.944585] nintendo 0005:057E:2009.0003: hidraw2: BLUETOOTH HID v80.01 Gamepad [Pro Controller] on xx:xx:xx:xx:xx:xx
[13583.015320] nintendo 0005:057E:2009.0003: controller MAC = 80:A5:03:61:63:25
[13583.021817] input: Nintendo Switch Pro Controller as /devices/pci0000:00/0000:00:14.0/usb2/2-4/2-4:1.0/bluetooth/hci0/hci0:256/0005:057E:2009.0003/input/input29
[13583.021945] input: Nintendo Switch Pro Controller IMU as /devices/pci0000:00/0000:00:14.0/usb2/2-4/2-4:1.0/bluetooth/hci0/hci0:256/0005:057E:2009.0003/input/input30

[rodolpheh]

I tried the same, with the same result... I also have the same controller than you.

[Woazboat]

Changed the issue name since it seems like this is not a simple fix. I've tried to get my controller to produce HID input reports over bluetooth, but haven't been successful so far.

[dogtopus]

Okay, I think I know what the actual problem is. The generic chinese pro controller clone I'm using doesn't actually have an led under the home button unlike the real pro controller.

https://www.aliexpress.com/item/4000571346668.html

I know this is not hid-sony but IIRC they explicitly stated that they won't make extra effort supporting knockoffs that don't work already since there are just too many of then all with their own quirks.

(Well unless if your knockoff controller used some unsupported official method to select "supported features" like 3rd party PS4 controllers. In that case support definitely sounds possible if the dev is willing to do it. Again seems that hid-sony didn't really got the interest to do this for their supported devices. I think the Xbox controller driver did though.)

[DanielOgorchock]

There is a conversation in the hid-nintendo latest patchset thread on the linux-input mailing list discussing how to handle the generic clones.

I think supporting them will be tricky when they don't support the full interface, but maybe portions of the driver can be tweaked to fail gracefully without exiting the probe. The thing about that which makes me weary is if that ends up masking real errors on the official controllers.

[ludenticus]

After a whole system reinstall, fiddling with dkms modules and changing distro, I found finally found out here that it was a hardware-related issue. Such a shame, provided the quality of the clone.

Thanks anyway, @DanielOgorchock

[SimonDaniel21]

my pro controller clone also has the same issue, but works flawless with my android phone over bluetooth...

[jorants]

To add some information:
I recently bought an off-brand pro controller that works perfectly fine on the switch and android as well.
This one

Similar errors over bluetooth:

[ 5880.300160] nintendo 0005:057E:2009.0005: unknown main item tag 0x0
[ 5880.300363] nintendo 0005:057E:2009.0005: hidraw1: BLUETOOTH HID v80.01 Gamepad [Pro Controller] on e4:70:b8:52:ec:79
[ 5884.359242] nintendo 0005:057E:2009.0005: failed reading SPI flash; ret=-110
[ 5884.359248] nintendo 0005:057E:2009.0005: using factory cal for left stick
[ 5886.385707] nintendo 0005:057E:2009.0005: failed reading SPI flash; ret=-110
[ 5886.385711] nintendo 0005:057E:2009.0005: using factory cal for right stick
[ 5888.412576] nintendo 0005:057E:2009.0005: failed reading SPI flash; ret=-110
[ 5888.412581] nintendo 0005:057E:2009.0005: Failed to read left stick cal, using dflts; e=-110
[ 5890.439234] nintendo 0005:057E:2009.0005: failed reading SPI flash; ret=-110
[ 5890.439241] nintendo 0005:057E:2009.0005: Failed to read right stick cal, using dflts; e=-110
[ 5892.465897] nintendo 0005:057E:2009.0005: failed reading SPI flash; ret=-110
[ 5892.465901] nintendo 0005:057E:2009.0005: using factory cal for IMU
[ 5894.492594] nintendo 0005:057E:2009.0005: failed reading SPI flash; ret=-110
[ 5894.492600] nintendo 0005:057E:2009.0005: Failed to read IMU cal, using defaults; ret=-110
[ 5894.492602] nintendo 0005:057E:2009.0005: Unable to read IMU calibration data
[ 5896.519246] nintendo 0005:057E:2009.0005: Failed to set report mode; ret=-110
[ 5897.052570] nintendo 0005:057E:2009.0005: Failed to enable rumble; ret=-110
[ 5897.052953] nintendo 0005:057E:2009.0005: probe - fail = -110
[ 5897.052978] nintendo: probe of 0005:057E:2009.0005 failed with error -110

When connecting over USB and switching to switch mode (it first pretends to be an XBOX controller) i get :

[ 6542.305622] usb 1-9: new full-speed USB device number 67 using xhci_hcd
[ 6542.428960] usb 1-9: device descriptor read/64, error -71
[ 6542.679072] usb 1-9: New USB device found, idVendor=057e, idProduct=2009, bcdDevice= 2.00
[ 6542.679074] usb 1-9: New USB device strings: Mfr=1, Product=2, SerialNumber=3
[ 6542.679075] usb 1-9: Product: Pro Controller
[ 6542.679076] usb 1-9: Manufacturer: Nintendo Co., Ltd.
[ 6542.679078] usb 1-9: SerialNumber: 000000000001
[ 6542.699657] nintendo 0003:057E:2009.0007: hidraw1: USB HID v81.11 Joystick [Nintendo Co., Ltd. Pro Controller] on usb-0000:00:14.0-9/input0
[ 6546.812296] nintendo 0003:057E:2009.0007: failed reading SPI flash; ret=-110
[ 6546.812297] nintendo 0003:057E:2009.0007: using factory cal for left stick
[ 6548.838972] nintendo 0003:057E:2009.0007: failed reading SPI flash; ret=-110
[ 6548.838974] nintendo 0003:057E:2009.0007: using factory cal for right stick
[ 6550.865644] nintendo 0003:057E:2009.0007: failed reading SPI flash; ret=-110
[ 6550.865645] nintendo 0003:057E:2009.0007: Failed to read left stick cal, using dflts; e=-110
[ 6552.892302] nintendo 0003:057E:2009.0007: failed reading SPI flash; ret=-110
[ 6552.892304] nintendo 0003:057E:2009.0007: Failed to read right stick cal, using dflts; e=-110
[ 6554.918961] nintendo 0003:057E:2009.0007: failed reading SPI flash; ret=-110
[ 6554.918964] nintendo 0003:057E:2009.0007: using factory cal for IMU
[ 6556.945626] nintendo 0003:057E:2009.0007: failed reading SPI flash; ret=-110
[ 6556.945628] nintendo 0003:057E:2009.0007: Failed to read IMU cal, using defaults; ret=-110
[ 6556.945629] nintendo 0003:057E:2009.0007: Unable to read IMU calibration data
[ 6558.972267] nintendo 0003:057E:2009.0007: Failed to set report mode; ret=-110
[ 6559.505642] nintendo 0003:057E:2009.0007: Failed to enable rumble; ret=-110
[ 6559.505822] nintendo 0003:057E:2009.0007: probe - fail = -110
[ 6559.505832] nintendo: probe of 0003:057E:2009.0007 failed with error -110
[ 6559.505858] usbcore: registered new interface driver usbhid
[ 6559.505859] usbhid: USB HID core driver

[hcsaustrup]

I just tested the same controller today on RetroPie, and got the same (kind of) result. The controller pairs up nicely through bluetooth, and is visible with the "jstest" app. The four numbered LEDs are blinking and none of the buttons seem to do anything at all. Again, the build quality seems quite nice, so it would be awesome if it worked :-/

[albfan]

Hi, I'm using a HORIPAD (aliexpress cheapest ones) and everything seems to be mapped correctly using dkms hid-nintendo, expect for:

ABS_RX which looks like only map the half of the movement.

Using generic hid driver, axis are correctly mapped so I supposed there was something incorrect here.

After add some traces on ABS_RX, ABS_RY, ABS_X and ABX_Y:

diff --git c/src/hid-nintendo.c i/src/hid-nintendo.c
index 5d47968..6f77fa6 100644
--- c/src/hid-nintendo.c
+++ i/src/hid-nintendo.c
@@ -1231,9 +1231,12 @@ static void joycon_parse_report(struct joycon_ctlr *ctlr,
                x = joycon_map_stick_val(&ctlr->left_stick_cal_x, raw_x);
                y = -joycon_map_stick_val(&ctlr->left_stick_cal_y, raw_y);
                /* report sticks */
+               hid_dbg(ctlr->hdev, "axis ABS_X: raw_x:%hu x=%ld\n",raw_x, x);
                input_report_abs(dev, ABS_X, x);
+               hid_dbg(ctlr->hdev, "axis ABS_Y: raw_y:%hu y=%ld\n",raw_y, y);
                input_report_abs(dev, ABS_Y, y);
 
+
                /* report buttons */
                input_report_key(dev, BTN_TL, btns & JC_BTN_L);
                input_report_key(dev, BTN_TL2, btns & JC_BTN_ZL);
@@ -1287,7 +1290,10 @@ static void joycon_parse_report(struct joycon_ctlr *ctlr,
                x = joycon_map_stick_val(&ctlr->right_stick_cal_x, raw_x);
                y = -joycon_map_stick_val(&ctlr->right_stick_cal_y, raw_y);
                /* report sticks */
+
+               hid_dbg(ctlr->hdev, "axis ABS_RX: raw_x:%hu x:%ld\n",raw_x, x);
                input_report_abs(dev, ABS_RX, x);
+               hid_dbg(ctlr->hdev, "axis ABS_RY: raw_y:%hu y:%ld\n",raw_y, y);
                input_report_abs(dev, ABS_RY, y);
 
                /* report buttons */

I see no major difference on input, but calibration is different:

[21098.090523] nintendo 0003:057E:2009.002C: using factory cal for left stick
[21098.090528] nintendo 0003:057E:2009.002C: requesting SPI flash data
[21098.113164] nintendo 0003:057E:2009.002C: using user cal for right stick
[21098.113176] nintendo 0003:057E:2009.002C: requesting SPI flash data
[21098.129168] nintendo 0003:057E:2009.002C: requesting SPI flash data
[21098.145237] nintendo 0003:057E:2009.002C: calibration:
               l_x_c=2020 l_x_max=3500 l_x_min=379
               l_y_c=2013 l_y_max=3573 l_y_min=535
               r_x_c=15 r_x_max=1251 r_x_min=-135
               r_y_c=1555 r_y_max=2900 r_y_min=-2336

Not sure where that info comes from, but see right cal comes from user.

I just use the calibration from left joycon and that just works. Well it needs a manual calibration later, but not it is centered.

diff --git i/src/hid-nintendo.c w/src/hid-nintendo.c
index 6f77fa6..800e4bd 100644
--- i/src/hid-nintendo.c
+++ w/src/hid-nintendo.c
@@ -768,7 +768,8 @@ static int joycon_request_calibration(struct joycon_ctlr *ctlr)
        }
 
        /* read the right stick calibration data */
-       ret = joycon_read_stick_calibration(ctlr, right_stick_addr,
+       ret = joycon_read_stick_calibration(ctlr, left_stick_addr,
                                            &ctlr->right_stick_cal_x,
                                            &ctlr->right_stick_cal_y,
                                            false);

Not sure if there's a way to detect this controller and use a different address for user calibration. Any ideas on how to find specs and how to detect this controller would be appreciated to provide a patch.

[DanielOgorchock]

Do you have a nintendo switch you could use to recalibrate the controller? If not, I'd try hardcoding joycon_check_for_cal_magic() to always return 1, making the controller use factory calibrations for both sticks. It sounds like there's bad user calibration data stored for the right stick, and the magic value is present, telling the driver to read the bad data.

[albfan]

@DanielOgorchock Do you think this is general issue? As this is a generic pro controller, my suspiction is that only cheap generic ones has this problems:

I have a pro controller and generic joycons to test (note squared + and - buttons, this are the cheapest ones):

Probably I can check with original ones and see if they behave different.

when the dkms (I'm on kernel 5.4 lts) is not active the axis are reported as 057e:2009 (pro controller), but as no driver deals with that, they are disconnected and reported as 0f0d:00c1 (horypad):

dmesg --follow
[14882.673977] usb 2-1.1: new full-speed USB device number 21 using ehci-pci
[14882.829410] usb 2-1.1: New USB device found, idVendor=057e, idProduct=2009, bcdDevice= 0.20
[14882.829417] usb 2-1.1: New USB device strings: Mfr=1, Product=2, SerialNumber=3
[14882.829421] usb 2-1.1: Product: Pro Controller
[14882.829425] usb 2-1.1: Manufacturer: Nintendo Co., Ltd.
[14882.829428] usb 2-1.1: SerialNumber: 000000000001
[14882.832869] input: Nintendo Co., Ltd. Pro Controller as /devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.1/2-1.1:1.0/0003:057E:2009.000E/input/input36
[14882.833320] hid-generic 0003:057E:2009.000E: input,hidraw0: USB HID v1.11 Joystick [Nintendo Co., Ltd. Pro Controller] on usb-0000:00:1d.0-1.1/input0

[14883.914066] usb 2-1.1: USB disconnect, device number 21

[14884.453983] usb 2-1.1: new full-speed USB device number 22 using ehci-pci
[14884.608810] usb 2-1.1: New USB device found, idVendor=0f0d, idProduct=00c1, bcdDevice= 5.72
[14884.608817] usb 2-1.1: New USB device strings: Mfr=1, Product=2, SerialNumber=0
[14884.608821] usb 2-1.1: Product: Switch Controller
[14884.608824] usb 2-1.1: Manufacturer:  
[14884.611064] input:   Switch Controller as /devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.1/2-1.1:1.0/0003:0F0D:00C1.000F/input/input37
[14884.611606] hid-generic 0003:0F0D:00C1.000F: input,hidraw0: USB HID v1.11 Gamepad [  Switch Controller] on usb-0000:00:1d.0-1.1/input0

from horipad the axis are reported as: ABS_ZZAN from 0-255 while on hid-nintendo they are reported as ABS_RX from -23764
to 32767. After manual calibration:

it works as expected.

Wonder if there's a way that dkms or kernel when it land on 5.10 attends an external config file, so you can calibrate joycons or procontroller for you personal use case. I can work a patch on that.

After that I want to fix the led (not available for joycons but available and not working for procontroller home) and add SL SR buttons (recognized on horipad, but not on hid-nintendo.

I review the @nadiaholmquist improvements and think that SL and SR are covered there so I can help reviewing those.