On Thu, 8 Jun 2017 09:13:03 +1000
Peter Hutterer [off-list ref] wrote:

On Wed, Jun 07, 2017 at 09:17:37PM +0200, Michal Suchánek wrote:

quoted

On Wed, 7 Jun 2017 10:16:22 -0700
Dmitry Torokhov [off-list ref] wrote:
  

quoted

On Wed, Jun 07, 2017 at 06:53:51PM +0200, Michal Suchánek wrote:  

quoted

On Sun, 28 May 2017 10:55:40 -0700
Dmitry Torokhov [off-list ref] wrote:
    

quoted

On Sun, May 28, 2017 at 11:47:58AM +0200, Michal Suchanek
wrote:    

quoted

On Tue, 9 May 2017 17:43:27 -0700
Dmitry Torokhov [off-list ref] wrote:

quoted

quoted

quoted

quoted

quoted

quoted

What hardware do you believe would benefit from this and
why?      

Any touchpad hardware where you cannot press two buttons at
once to emulate the third button due to hardware design. And
any touchpad hardware on which some of the buttons are
broken when it comes to it.

It is built into a notebook and works fine for moving the
cursor but due to lack of usable buttons you still need a
mouse to use the notebook.      

Have you tried simply redefining keymap of your keyboard to
emit BTN_RIGHT/BTN_MIDDLE? Both atkbd and HID keyboards
support keymap updates from userspace/udev/hwdb and if there
is a driver that does not support it I will take patches
fixing that.    

Indeed, they do support it. Such keymap update just does not
work as mouse button regardless of sending the BTN_* event. At
least not in X11.

So what is next?    

Teach X11 to handle it properly.

Thanks.
  

That's actually libinputs fault. It marks devices with some random
capabilities and when the event does not match capability set it is
dropped.  

just because it's not immediately apparent doesn't mean it's random.
We use ID_INPUT_* from udev to determine the device capabilities. In
some cases we override it when we have some information udev doesn't
have. e.g. we disable gestures on some touchpads known to be
inaccurate for multi-finger gestures. or on some devices we disable
event codes because the device doesn't actually have them. see my
explanation here:
https://who-t.blogspot.com.au/2015/06/libinput-and-lack-of-device-types.html

the reason we do it this way is so that a) all of the stack can agree
on a device's device type and b) you can override many misdetections
with a hwdb entry or a custom udev rule rather than waiting for a new
libinput release that encodes the new magic.

quoted

Adding the capability with /etc/udev/rules.d/xxx-input.rules

ENV{ID_INPUT_KEYBOARD}=="1" ENV{ID_INPUT_MOUSE}="1"

resolves the problem. It must come very late in rules otherwise
something resets it back.

This is inefficient to enable by default because libinput must
create a second shadow X11 device for devices that generate both
input and keyboard events.   

false. xf86-input-libinput has to do this. libinput doesn't do it,
it's capable of one device having multiple capabilities. due to the
previously mentioned design restrictions in the X server, this is the
most efficient way to work around it. if xf86-input-evdev supported
multi-type devices, it would have to do the same thing.

And that's argument just for the sake of arguing or what?

quoted

⎡ Virtual core pointer                   id=2   [master pointer
(3)] ⎜   ↳ Virtual core XTEST pointer         id=4   [slave
pointer  (2)] ⎜   ↳ DELL Dell USB Entry Keyboard       id=8
[slave  pointer  (2)] ⎜   ↳ PixArt Dell MS116 USB Optical
Mouseid=9   [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)] ↳ DELL
Dell USB Entry Keyboard       id=10  [slave  keyboard (3)] ↳ Power
Button                       id=7   [slave  keyboard (3)]

Presumably it could infer the capabilities from the supported events
rather than hardcoding them in udev. Surely there are devices with
stub/broken features that do not actually generate events but that
is hopefully not the norm.  

how do you think udev decides on the device type? it looks at the
supported events and then picks a type based on that.
https://github.com/systemd/systemd/blob/master/src/udev/udev-builtin-input_id.c
"presumably, it could..." is btw a perfect example of how everything
looks simple when viewed from enough distance...

This is what evtest reports about my keyboard:

Select the device event number [0-12]: 2
Input driver version is 1.0.1
Input device ID: bus 0x3 vendor 0x413c product 0x2107 version 0x111
Input device name: "DELL Dell USB Entry Keyboard"
Supported events:
  Event type 0 (EV_SYN)
  Event type 1 (EV_KEY)
    Event code 1 (KEY_ESC)
    Event code 2 (KEY_1)
    Event code 3 (KEY_2)
    Event code 4 (KEY_3)
...
    Event code 193 (KEY_F23)
    Event code 194 (KEY_F24)
    Event code 240 (KEY_UNKNOWN)
    Event code 272 (BTN_LEFT)
    Event code 273 (BTN_RIGHT)
    Event code 274 (BTN_MIDDLE)
  Event type 4 (EV_MSC)
    Event code 4 (MSC_SCAN)
  Event type 17 (EV_LED)
    Event code 0 (LED_NUML) state 1
    Event code 1 (LED_CAPSL) state 0
    Event code 2 (LED_SCROLLL) state 0
    Event code 3 (LED_COMPOSE) state 0
    Event code 4 (LED_KANA) state 0
Key repeat handling:
  Repeat type 20 (EV_REP)
    Repeat code 0 (REP_DELAY)
      Value    250
    Repeat code 1 (REP_PERIOD)
      Value     33
Properties:

Anyway, I'm done here. If you have any specific technical questions
feel free to ask, but this is enough time wasted arguing. The one

So from the distance of evtest it looks like it supports mouse buttons
yet from the distance of libinput it does not. So maybe libinput needs
to take a step back from the keyboard to see that? 

Like maybe assigning the classes after it went through the hwdb fixups.

question that hasn't been asked yet though: what specific device are
we talking about here? That may put the "broken by design" claims
into a better perspective.

Which part broken by design do you mean, exactly?

There are too many when it comes to input devices to be able to tell.

Thanks

Michal