[Input Device] I2C-HID mouse device cannot work on Chrome OS

[Lak Fu]

Hi everyone, 

I want to re-work a new touchpad via I2C to my Chromebook.

So I installed crouton, and written a kernel module to register it for invoke I2C-HID driver.

Finally, my device is work fine on crouton, it can move, click and support the gesture, but the strange thing is that click function will not work when I switch back to Chrome OS desktop.

I checked the xinput under crouton, it is fine and succeed to drive by I2C HID v1.00 Mouse driver, and back to Chrome OS desktop, it can move and gesture, just no click. 

Does anyone have any ideas about this situation?

I will be very grateful for any suggestion or comment, thanks.

[Daniel Erat]

Please don't repost on the mailing list; you already sent the same message the day before. It would be more productive to post a followup on your original thread supplying more information (link to your module code, post logs, etc.).

--
--
Chromium OS Developers mailing list: chromiu...@chromium.org
View archives, change email options, or unsubscribe:
http://groups.google.com/a/chromium.org/group/chromium-os-dev?hl=en

[Lak Fu]

Sorry it is my mistake, thanks for your suggestion.
I will post more information about this situation later.

[Lak Fu]

More details.

My touchpad is just a normal type like others, and it have not PHY buttons.

The interesting is the input system receive correct input event on crouton environment. 

Below is my xinput shown, my device is "hid-over-i2c 093A:0000 Touchpad          id=15 [slave  pointer  (2)]".

� Virtual core pointer                     id=2 [master pointer  (3)]

�   ��� Virtual core XTEST pointer               id=4 [slave  pointer  (2)]

�   ��� Elan Touchscreen                         id=11 [slave  pointer  (2)]

�   ��� WCOM50C1:00 2D1F:0016                    id=12 [slave  pointer  (2)]

�   ��� hid-over-i2c 093A:0000 Touchpad          id=15 [slave  pointer  (2)]

�   ��� Logitech USB Receiver                    id=10 [slave  pointer  (2)]

� Virtual core keyboard                    id=3 [master keyboard (2)]

    ��� Virtual core XTEST keyboard              id=5 [slave  keyboard (3)]

    ��� Power Button                             id=6 [slave  keyboard (3)]

    ��� Power Button                             id=7 [slave  keyboard (3)]

    ��� Sleep Button                             id=8 [slave  keyboard (3)]

    ��� bxtda7219max Headset Jack                id=13 [slave  keyboard (3)]

    ��� AT Translated Set 2 keyboard             id=14 [slave  keyboard (3)]

    ��� Logitech USB Receiver                    id=9 [slave  keyboard (3)]

and the kernel log,

[   75.067716] input: hid-over-i2c 093A:0000 Touchpad as /devices/pci0000:00/0000:00:17.0/i2c_designware.4/i2c-10/10-002c/0018:093A:0000.0004/input/input14

[   75.068418] hid-multitouch 0018:093A:0000.0004: input,hidraw3: I2C HID v1.00 Mouse [hid-over-i2c 093A:0000] on 10-002c

next, I observer the touch event via "xinput test 15", and the button events are truly exists.

(xenial)lak@localhost:~/workspace/pixart_tp_driver$ xinput test 15

button press   1 

button release 1 

button press   1 

button release 1 

motion a[3]=401 

motion a[3]=396 

motion a[3]=391 

motion a[3]=386 

motion a[3]=380 

motion a[3]=376 

button press   4 

button release 4 

motion a[3]=373 

And it is work fine on crouton, but once I back to Chrome OS desktop, it is become invalid.

[Dmitry Torokhov]

[resending form the right account]

What capabilities does your touchpad report? Use evtest utility - it is easiest - evtest /dev/input/event14 and interrupt when it is done printing device's capabilities. Just to confirm, it is only the physical button underneath the touchpad that is not handled properly, the rest functions well (tapping, tracking multiple fingers, etc), right?

--

[Lepton Wu]

FYI, there is some special handling in chrome side which will try to "identify" device type and use different logic to handle events.

It's possible it identify your device as a "wrong" type and then your event never get handled by chrome.

https://cs.chromium.org/chromium/src/ui/events/ozone/evdev/input_device_factory_evdev.cc?rcl=342377337b75a5bd6178bb4b2189e06c6e0f999e&l=85

[Lak Fu]

Hi,

Thanks for your comments.

 it is only the physical button underneath the touchpad that is not handled properly, the rest functions well (tapping, tracking multiple fingers, etc), right?

Almost correct, all gesture features are well, but I can not "click" with tap the touchpad, not the physical button (I don't try it..).

Below is the evtest result, is any strange?

Input driver version is 1.0.1

Input device ID: bus 0x18 vendor 0x93a product 0x0 version 0x100

Input device name: "hid-over-i2c 093A:0000 Touchpad"

Supported events:

  Event type 0 (EV_SYN)

  Event type 1 (EV_KEY)

    Event code 272 (BTN_LEFT)

    Event code 325 (BTN_TOOL_FINGER)

    Event code 330 (BTN_TOUCH)

    Event code 333 (BTN_TOOL_DOUBLETAP)

    Event code 334 (BTN_TOOL_TRIPLETAP)

    Event code 335 (BTN_TOOL_QUADTAP)

  Event type 3 (EV_ABS)

    Event code 0 (ABS_X)

      Value   1395

      Min        0

      Max     1223

      Resolution      12

    Event code 1 (ABS_Y)

      Value    817

      Min        0

      Max      745

      Resolution      13

    Event code 47 (ABS_MT_SLOT)

      Value      0

      Min        0

      Max        3

    Event code 53 (ABS_MT_POSITION_X)

      Value      0

      Min        0

      Max     1223

      Resolution      12

    Event code 54 (ABS_MT_POSITION_Y)

      Value      0

      Min        0

      Max      745

      Resolution      13

    Event code 57 (ABS_MT_TRACKING_ID)

      Value      0

      Min        0

      Max    65535

Properties:

  Property type 0 (INPUT_PROP_POINTER)

  Property type 2 (INPUT_PROP_BUTTONPAD)

Testing ... (interrupt to exit)

***********************************************

  This device is grabbed by another process.

  No events are available to evtest while the

  other grab is active.

  In most cases, this is caused by an X driver,

  try VT-switching and re-run evtest again.

  Run the following command to see processes with

  an open fd on this device

 "fuser -v /dev/input/event14"

***********************************************

^C(xenial)lak@localhost:~/workspace/pixart_tp_driver$ fuser -v /dev/input/event14

                     USER        PID ACCESS COMMAND

/dev/input/event14:  lak        1168 F.... chrome

                     lak        9476 F.... Xorg

                     lak        9586 f.... hexdump

[Lak Fu]

Thank you, it is useful information, I will try to trace the code....

Oh my, I maybe need to build a own version Chrome OS for debug..

I really hope don't need to do that.... (Is any good suggestion?)

[Lepton Wu]

You can use strace to chrome to see if chrome has read event from /dev/input/event*, if chrome has got event but just don't do thing you are expecting, then I guess it's this issue.

[Lak Fu]

Wow, I am never use this tool before. It looks very powerful, I will study how to use it.

(I guess I need start this tool in crouton then back to ChromeOS desktop, try to use the touchpad and back to crouton, observe it, right?)

Thanks for your hint!!

Lepton Wu於 2018年1月9日星期二 UTC+8下午4時01分35秒寫道：

[Lak Fu]

Hi Lepton,

I tried to find some information via strace tool, but I don't see any log about "event" or "dev" or "input", am I missing anything?

My Chrome pid is 1494, my test command is the following, 

sudo strace -p 1494 2>&1 | grep -i "event"

For test, I am try this command on a common chromebook (without my touchpad).

But it seems not throw any information that I interested...

Would you like give me more clues?? Thank~~

Lepton Wu於 2018年1月9日星期二 UTC+8下午4時01分35秒寫道：

[Dmitry Torokhov]

[Argh, again resending from the right address]

Hi Lak,

Unfortunately your hardware is not Chrome OS  compliant:

"The following parameters must be supplied for each contact:

• ABS_MT_POSITION_X
• ABS_MT_POSITION_Y
• ABS_MT_TOUCH_MAJOR
• ABS_MT_SLOT
• ABS_MT_TRACKING_ID

The following parameters are recommended but not required:

• ABS_MT_PRESSURE
• ABS_MT_TOUCH_MINOR
• ABS_MT_ORIENTATION"

Because your device does not [currently?] report ABS_MT_TOUCH_MAJOR Chrome OS ignores it. If your hardware is capable of reporting the contact size you need to modify your firmware to report it.

Thanks,
Dmitry

[Lepton Wu]

You can use lsof -n -p 1494 or ls /proc/1494/fd -l to find out which fd is connected to /dev/input/event...

And then just pay attention to "read" system call in your strace.

BTW, there is multiple chrome process, so you need to figure out which one is responsible for reading input from event devices.

You can use "lsof" or "ls /proc..." to confirm this.

[Lak Fu]

Hi Dmitry,

Thank you so much! You give me very important information.

But where I can give this rule (spec?)?

"The following parameters must be supplied for each contact:

ABS_MT_POSITION_X

ABS_MT_POSITION_Y

ABS_MT_TOUCH_MAJOR

ABS_MT_SLOT

ABS_MT_TRACKING_ID

And because I am driving my hardware with I2C-HID mouse driver, I will going to check whether I2C-HID mouse driver support "ABS_MT_TOUCH_MAJOR". If it is, yes, I think I need to update my firmware....

One more thing, may I know where the source code about below limitation? 

Because your device does not [currently?] report ABS_MT_TOUCH_MAJOR Chrome OS ignores it.

So once again, thank you very much.

BTW, "[Argh, again resending from the right address]" <-- what it means?? XD.

Dmitry Torokhov於 2018年1月12日星期五 UTC+8下午3時12分31秒寫道：

[Dmitry Torokhov]

On Fri, Jan 12, 2018 at 12:10 AM, Lak Fu <lak4...@gmail.com> wrote:

Hi Dmitry,

Thank you so much! You give me very important information.

But where I can give this rule (spec?)?

Are you trying to get on Chrome OS AVL? You need to get in touch with Partner engineering to get access to hardware specs. There are quite a bit of requirements for touchpads and touchscreens before they can be placed on AVL.

 

"The following parameters must be supplied for each contact:

ABS_MT_POSITION_X

ABS_MT_POSITION_Y

ABS_MT_TOUCH_MAJOR

ABS_MT_SLOT

ABS_MT_TRACKING_ID

And because I am driving my hardware with I2C-HID mouse driver, I will going to check whether I2C-HID mouse driver support "ABS_MT_TOUCH_MAJOR". If it is, yes, I think I need to update my firmware....

Yes, Linux i2c-hid driver supports ABS_MT_TOUCH_MAJOR, but the device's HID descriptor should contain appropriate reports/usages.

 

One more thing, may I know where the source code about below limitation? 

Because your device does not [currently?] report ABS_MT_TOUCH_MAJOR Chrome OS ignores it.

So once again, thank you very much.

BTW, "[Argh, again resending from the right address]" <-- what it means?? XD.

When I compose a reply, by default it uses my @google.com address, but it is not subscribed to the mailing list, so my message is bounced and I have to resend it from my @chromium.org account so not only you see my reply but other folks as well.

Thanks,

Dmitry

[Lak Fu]

Are you trying to get on Chrome OS AVL? You need to get in touch with Partner engineering to get access to hardware specs. There are quite a bit of requirements for touchpads and touchscreens before they can be placed on AVL.

Yes, I am trying that. And thanks, I will follow the official process to get hardware specs.

Yes, Linux i2c-hid driver supports ABS_MT_TOUCH_MAJOR, but the device's HID descriptor should contain appropriate reports/usages.

Got it!!

Because your device does not [currently?] report ABS_MT_TOUCH_MAJOR Chrome OS ignores it.

About this, can I find the source code from "The Chromium Projects"? I am like to trace the source code.. XD

 When I compose a reply, by default it uses my @google.com address, but it is not subscribed to the mailing list, so my message is bounced and I have to resend it from my @chromium.org account so not only you see my reply but other folks as well.

XD 

Believe me, I am truly grateful for your help. I will going to next move, and hope I can done this job well.

 

Dmitry Torokhov於 2018年1月12日星期五 UTC+8下午4時24分34秒寫道：

[Lak Fu]

wow~~ Thanks for your direction, I will try this!! 

Lepton Wu於 2018年1月12日星期五 UTC+8下午4時03分51秒寫道：

[Dmitry Torokhov]

On Fri, Jan 12, 2018 at 12:40 AM, Lak Fu <lak4...@gmail.com> wrote:

Are you trying to get on Chrome OS AVL? You need to get in touch with Partner engineering to get access to hardware specs. There are quite a bit of requirements for touchpads and touchscreens before they can be placed on AVL.

Yes, I am trying that. And thanks, I will follow the official process to get hardware specs.

Yes, Linux i2c-hid driver supports ABS_MT_TOUCH_MAJOR, but the device's HID descriptor should contain appropriate reports/usages.

Got it!!

Because your device does not [currently?] report ABS_MT_TOUCH_MAJOR Chrome OS ignores it.

About this, can I find the source code from "The Chromium Projects"? I am like to trace the source code.. XD

You can see Chrome input handling here: https://cs.chromium.org/chromium/src/ui/events/ozone/evdev/ 

 

 When I compose a reply, by default it uses my @google.com address, but it is not subscribed to the mailing list, so my message is bounced and I have to resend it from my @chromium.org account so not only you see my reply but other folks as well.

XD 

Believe me, I am truly grateful for your help. I will going to next move, and hope I can done this job well.

 

Dmitry Torokhov於 2018年1月12日星期五 UTC+8下午4時24分34秒寫道：

On Fri, Jan 12, 2018 at 12:10 AM, Lak Fu <lak4...@gmail.com> wrote:

Hi Dmitry,

Thank you so much! You give me very important information.

But where I can give this rule (spec?)?

Are you trying to get on Chrome OS AVL? You need to get in touch with Partner engineering to get access to hardware specs. There are quite a bit of requirements for touchpads and touchscreens before they can be placed on AVL.

 

"The following parameters must be supplied for each contact:

ABS_MT_POSITION_X

ABS_MT_POSITION_Y

ABS_MT_TOUCH_MAJOR

ABS_MT_SLOT

ABS_MT_TRACKING_ID

And because I am driving my hardware with I2C-HID mouse driver, I will going to check whether I2C-HID mouse driver support "ABS_MT_TOUCH_MAJOR". If it is, yes, I think I need to update my firmware....

Yes, Linux i2c-hid driver supports ABS_MT_TOUCH_MAJOR, but the device's HID descriptor should contain appropriate reports/usages.

 

One more thing, may I know where the source code about below limitation? 

Because your device does not [currently?] report ABS_MT_TOUCH_MAJOR Chrome OS ignores it.

So once again, thank you very much.

BTW, "[Argh, again resending from the right address]" <-- what it means?? XD.

When I compose a reply, by default it uses my @google.com address, but it is not subscribed to the mailing list, so my message is bounced and I have to resend it from my @chromium.org account so not only you see my reply but other folks as well.

Thanks,

Dmitry

--

[Lak Fu]

Got it!

Dmitry Torokhov於 2018年1月12日星期五 UTC+8下午4時53分59秒寫道：

[Lak Fu]

Hi, 

I finally made the firmware to report ABS_MT_TOUCH_MAJOR event, but the click function is still unavailable.

Is there have anything that I can go to check or I missing?

Below is my current device info. I thinks it is very well, but it is still not work... so sad...

Supported events:
  Event type 0 (EV_SYN)
  Event type 1 (EV_KEY)
    Event code 272 (BTN_LEFT)
    Event code 325 (BTN_TOOL_FINGER)
    Event code 328 (BTN_TOOL_QUINTTAP)
    Event code 330 (BTN_TOUCH)
    Event code 333 (BTN_TOOL_DOUBLETAP)
    Event code 334 (BTN_TOOL_TRIPLETAP)
    Event code 335 (BTN_TOOL_QUADTAP)
  Event type 3 (EV_ABS)
    Event code 0 (ABS_X)
      Value   1298
      Min        0
      Max     1223
      Resolution      12
    Event code 1 (ABS_Y)
      Value    243
      Min        0
      Max      745
      Resolution      12
    Event code 47 (ABS_MT_SLOT)
      Value      1
      Min        0
      Max        4
    Event code 48 (ABS_MT_TOUCH_MAJOR)
      Value      0
      Min        0
      Max      745
      Resolution      12
    Event code 49 (ABS_MT_TOUCH_MINOR)
      Value      0
      Min        0
      Max      745
      Resolution      12
    Event code 52 (ABS_MT_ORIENTATION)
      Value      0
      Min        0
      Max        1
    Event code 53 (ABS_MT_POSITION_X)
      Value      0
      Min        0
      Max     1223
      Resolution      12
    Event code 54 (ABS_MT_POSITION_Y)
      Value      0
      Min        0
      Max      745
      Resolution      12
    Event code 55 (ABS_MT_TOOL_TYPE)
      Value      0
      Min        0
      Max        2
    Event code 57 (ABS_MT_TRACKING_ID)
      Value      0
      Min        0
      Max    65535
  Event type 4 (EV_MSC)
    Event code 5 (MSC_TIMESTAMP)
Properties:
  Property type 0 (INPUT_PROP_POINTER)
  Property type 2 (INPUT_PROP_BUTTONPAD)
Testing ... (interrupt to exit)

And, when I click, the event report below,

Event: time 1516869455.738172, type 3 (EV_ABS), code 57 (ABS_MT_TRACKING_ID), value 118
Event: time 1516869455.738172, type 3 (EV_ABS), code 53 (ABS_MT_POSITION_X), value 1145
Event: time 1516869455.738172, type 3 (EV_ABS), code 54 (ABS_MT_POSITION_Y), value 579
Event: time 1516869455.738172, type 3 (EV_ABS), code 48 (ABS_MT_TOUCH_MAJOR), value 50
Event: time 1516869455.738172, type 3 (EV_ABS), code 49 (ABS_MT_TOUCH_MINOR), value 50
Event: time 1516869455.738172, type 1 (EV_KEY), code 330 (BTN_TOUCH), value 1
Event: time 1516869455.738172, type 1 (EV_KEY), code 325 (BTN_TOOL_FINGER), value 1
Event: time 1516869455.738172, type 3 (EV_ABS), code 0 (ABS_X), value 1145
Event: time 1516869455.738172, type 3 (EV_ABS), code 1 (ABS_Y), value 579
Event: time 1516869455.738172, type 4 (EV_MSC), code 5 (MSC_TIMESTAMP), value 0
Event: time 1516869455.738172, -------------- SYN_REPORT ------------
Event: time 1516869455.741437, type 3 (EV_ABS), code 48 (ABS_MT_TOUCH_MAJOR), value 51
Event: time 1516869455.741437, type 3 (EV_ABS), code 49 (ABS_MT_TOUCH_MINOR), value 51
Event: time 1516869455.741437, type 4 (EV_MSC), code 5 (MSC_TIMESTAMP), value 12900
Event: time 1516869455.741437, -------------- SYN_REPORT ------------
Event: time 1516869455.748971, type 3 (EV_ABS), code 48 (ABS_MT_TOUCH_MAJOR), value 50
Event: time 1516869455.748971, type 3 (EV_ABS), code 49 (ABS_MT_TOUCH_MINOR), value 50
Event: time 1516869455.748971, type 4 (EV_MSC), code 5 (MSC_TIMESTAMP), value 16100
Event: time 1516869455.748971, -------------- SYN_REPORT ------------
Event: time 1516869455.756343, type 3 (EV_ABS), code 48 (ABS_MT_TOUCH_MAJOR), value 44
Event: time 1516869455.756343, type 3 (EV_ABS), code 49 (ABS_MT_TOUCH_MINOR), value 44
Event: time 1516869455.756343, type 4 (EV_MSC), code 5 (MSC_TIMESTAMP), value 23500
Event: time 1516869455.756343, -------------- SYN_REPORT ------------
Event: time 1516869455.763886, type 3 (EV_ABS), code 48 (ABS_MT_TOUCH_MAJOR), value 22
Event: time 1516869455.763886, type 3 (EV_ABS), code 49 (ABS_MT_TOUCH_MINOR), value 22
Event: time 1516869455.763886, type 4 (EV_MSC), code 5 (MSC_TIMESTAMP), value 30700
Event: time 1516869455.763886, -------------- SYN_REPORT ------------
Event: time 1516869455.771316, type 3 (EV_ABS), code 57 (ABS_MT_TRACKING_ID), value -1
Event: time 1516869455.771316, type 1 (EV_KEY), code 330 (BTN_TOUCH), value 0
Event: time 1516869455.771316, type 1 (EV_KEY), code 325 (BTN_TOOL_FINGER), value 0
Event: time 1516869455.771316, type 4 (EV_MSC), code 5 (MSC_TIMESTAMP), value 38100
Event: time 1516869455.771316, -------------- SYN_REPORT ------------

Dmitry Torokhov於 2018年1月12日星期五 UTC+8下午4時53分59秒寫道：

[Michael Spang]

The gestures library has lots of tuning parameters that need to be set for each touchpad. These are installed to /etc/gesture.

See here: https://chromium.googlesource.com/chromiumos/platform/xorg-conf/+/master

If you have reworked a different touchpad in, you won't have the right tuning. It is also possible that the device doesn't meet the library's requirements which would need to be addressed by firmware or hardware changes.

Michael

[Lak Fu]

Hi Michael,

Thanks for those information.

Is it mean that Chrome OS cannot support the simple HID touchpad? 

My device is driving successfully by HID input driver, and the input events are seems correct.

Actually, gesture features are work fine, but only "click" is not.... It really confuse me...

Is there have any debug log to let me know what I missing, why Chrome OS filter out (drop) my touch event?

Anyway, I will check those parameters, thanks~~

[Michael Spang]

It looks like you don't have a pressure axis, you likely need support for that as well.

Some of the parameters you'll probably want to check out if no taps are recognized at all are minimum pressure ("Tap Minimum Pressure") and maximum distance moved ("Tap Maximum Movement").

Michael

[Lak Fu]

Ok, I will add pressure report to my firmware.

As above discussion, Dmitry give me some information so I was thought it is not necessary.

But you are right, I will try it. Thank you.

"The following parameters must be supplied for each contact:

• ABS_MT_POSITION_X
• ABS_MT_POSITION_Y
• ABS_MT_TOUCH_MAJOR
• ABS_MT_SLOT
• ABS_MT_TRACKING_ID

The following parameters are recommended but not required:

• ABS_MT_PRESSURE
• ABS_MT_TOUCH_MINOR
• ABS_MT_ORIENTATION"

[Lak Fu]

Ya!!!!!!!! Michael, you're damned right!

Everything is perfect after I added the pressure info. Finally....

Thank you so much!!!

Michael Spang於 2018年1月26日星期五 UTC+8上午1時43分10秒寫道：