Some touchpads, e.g. the Cyapa in the Acer c720 have a small axis range
([0, 870], [0, 470]), so the diagonal/magic value yields a hysteresis margin
of 1 device unit. On that device, that's one-tenth of a millimeter, causing
pointer motion just by holding the finger.
For touchpads that provide a physical resolution, set the hysteresis axes to
0.5mm and do away with the magic factor.
https://bugzilla.redhat.com/show_bug.cgi?id=1230441
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Our motion filter takes the last couple of vectors to calculate speed,
provided the direction stays the same and it is within a certain timeout. It
does not take into account lifting the finger, so the velocity on the first
event is off.
Real-world impact is mainly on scrolling. Before commit 289e4675
filter: enforce minimum velocity
the first motion on a scroll was accelerated by a factor of 0 and swallowed.
After 289e4675 the motion was calculated based on the timeout and a fraction
of the expected effect. Now the first scroll motion is based on the real
finger motion since setting the finger down and thus feels a bit more
responsive.
It also makes a couple of test cases using litest_assert_scroll() work again
since the miniumum motion is now as expected.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Tagging a device should occur only once during configure. We do not
have devices that can be changed after they are configured, so there is no
point in having the tagging part in a deferred struct.
Plus, the note saying that we tag with only one of EVDEV_TAG was wrong.
Now that we are chosing when we call each evdev_tag_*, we can also get
rid of the device->seat_caps tests.
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
If a relative device is tagged by udev as ID_INPUT_TOUCHPAD we need to
catch this before we try to dereference device->abs.absinfo_x.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
The current code labels a touch as palm if it started within the typing
timeouts. To move the pointer even after the timeout expires, a user has to
lift the finger which is quite annoying and different to the old synaptics
driver behaviour (which had a simple on/off toggle on whether to let events
through or not).
Be smarter about this: if a touch starts _after_ the last key press event,
release it for pointer motion once the timeout expires. Touches started before
the last key press remain labelled as palms. This makes it possible to rest
the palm on the touchpad while typing without getting interference but also
provides a more responsive UI when moving from typing to using the touchpad
normally.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Tested-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Tested-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Tested-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Tested-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
Check a couple of easy yes/no definitives that cover most Lenovo laptops,
and avoid false positives on Wacoms.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Tested-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
The keyboard test is a simple one, if we have the first row of alphabetic
keys, we assume it's a full keyboard.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Tested-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Tested-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
Alt-tab should not trigger the disable-while-typing timeout, likewise with the
F-keys, multimedia keys, the windows and menu key, etc.
https://bugs.freedesktop.org/show_bug.cgi?id=90613
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Tested-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Tested-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
Unlikely, but there's the odd chance of the first touch coming in with the
same X or Y coordinate the kernel already has internally. This would
generate a bogus delta on the second event when the touch coordinate jumps
from 0/y or x/0 to the real coordinates.
For touchpads with distance support this is a real issue since the default
value for a touch distance is > 0.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Tested-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
The touchpad carries enough state around that calling release_all_keys() isn't
enough to properly suspend it. e.g. a button down after tapping won't be
released by trying to release the physical button for it.
We need to clear the state properly, but that's interface-specific so add a
new hook for it.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Tested-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Add "interface" to the name to make things a bit clearer.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Tested-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Hitting a single key triggers a short timeout, just enough to cover the time
to the next key event. Hitting more than one key triggers the longer timeout.
This should improve responsiveness after single key events when the touchpad is
still the main interaction mode and a key needs to be pressed to advance in
the UI. When typing the hands require physical movement to get back to the
touchpad anyway so a longer timeout is acceptable and more reliable.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
On some touchpads, typing triggers touches in areas of the touchpad that
cannot easily be distinguished from other fingers. Pressure information is
useless too, so we have to go back to a timeout-based handling of touch data.
If we see non-modifier key events, disable the touchpad for a timeout and set
any touches starting during that timeout as palm.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Preparation to add different palm detection types. Not all of them need to be
un-done when leaving the edge area so a boolean is not enough.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
When the device supports true hovering, it reports this
information through ABS_MT_DISTANCE.
When this axis is available, we should rely on it to
(un)hover the touches as BTN_TOUCH is most of the time
unreliable (generated by the mouse emulation in the kernel).
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Causes an error message in the device_disable_release_tap_n_drag test. When
the touchpad is suspended, all touches are ended in tp_clear_state. Since the
hovering support was added, this returns the touches to TOUCH_HOVERING, a
subsequent tp_handle_state() will turn them back into TOUCH_BEGIN based on
BTN_TOUCH and BTN_TOOL_FINGER still being down.
Clear the fake touch buttons as well after ending the touches, this way the
touch points are reset to TOUCH_NONE as intended.
Once we do that we don't need to manually change the tap finger count when
releasing taps, we can just let the count reset naturally.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Those touchpads presents an actual lower resolution that what is
advertised.
We see some jumps from the cursor due to the big steps in X and Y
when we are receiving data.
For instance, we receive:
E: 13.471932 0003 0000 16366 # EV_ABS / ABS_X 16366
E: 13.471932 0003 0001 9591 # EV_ABS / ABS_Y 9591
E: 13.471932 0000 0000 0000 # ------------ SYN_REPORT (0) ----------
E: 13.479924 0003 0000 16316 # EV_ABS / ABS_X 16316
E: 13.479924 0003 0001 9491 # EV_ABS / ABS_Y 9491
E: 13.479924 0000 0000 0000 # ------------ SYN_REPORT (0) ----------
E: 13.487939 0003 0000 16271 # EV_ABS / ABS_X 16271
E: 13.487939 0003 0001 9403 # EV_ABS / ABS_Y 9403
E: 13.487939 0000 0000 0000 # ------------ SYN_REPORT (0) ----------
-> jumps of ~50 in X in each report, and ~100 for Y.
Apply a factor to minimize those jumps at low speed, and try
keeping the same feeling as regular touchpads at high speed.
It still feels slower but it is usable at least
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Tested-by: Vasily Khoruzhick <anarsoul@gmail.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
https://bugzilla.redhat.com/show_bug.cgi?id=1209753 lists a touchpad 76mm wide
that suffers from palm touches
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Touches in the exclusion zone are ignored for palm detection and don't move
the cursor. Tapping however triggers before we know whether something is a
palm or not, so we get erroneous button clickst.
If a tap happens in the top half of the touchpad, within the palm exclusion
zones, ignore it for tap purposes. To avoid further complicating the state
machine simply pretend there was a movement > threshold on that finger. This
advances the tap state machine properly that no button events are sent for
this finger.
https://bugs.freedesktop.org/show_bug.cgi?id=89625
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Internal touchpads with trackpoints are either BUS_I8042 or BUS_I2C, but not
BUS_USB. Lenovo sells external keyboards with a trackpoint built-in, make sure
we don't pair that trackpoint with the internal touchpad.
And likewise, the internal trackpoint should not be paired with e.g. a wacom
touch device.
Lenovo had one external device that has a trackpoint and a touchpad on an
external keyboard. That device won't be covered with this patch, if we have a
user we can re-consider.
https://bugs.freedesktop.org/show_bug.cgi?id=89935
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Less ambiguous since real_touches can be interpreted to "current number of
real touches as opposed to fake touches". Which it isn't, this variable holds
the number of slots.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Up to QUINTTAP, we count fake fingers through the BTN_TOOL_*TAP kernel defines.
Once we exceed QUINTTAP, the nfake_finger count returns to 0 and
tp_unhover_touches terminates all touch sequences. The most visible effect of
this was stopped in 591a41f but the problem remained.
Since we're not using 5 fingers for anything, use that to set the
overflow flag. The kernel gives us either a BTN_TOUCH 0 (all
released) or a lower BTN_TOOL_*TAP to unset the flag when we go below 5
fingers again.
And if we overflow, we can skip the unhovering of touch points since we a)
have a decent touchpad that gives us real touchpoints and b) hovering isn't
supported for 5 touches anyway.
https://bugs.freedesktop.org/show_bug.cgi?id=89800
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
A switch from BTN_TOOL_FINGER to BTN_TOOL_DOUBLETAP sets the former to 0, then
the latter to 1, within the same frame. In the previous code we'd end the
first touchpoint, then start two new ones immediately after when the DOUBLETAP
comes in. This causes bug notices in the edge scrolling code and finger
miscounts in the tapping code (since neither processes the change, there is no
SYN_REPORT between the two).
Only update the state bits when we get the events, handle the fake touch
sequence start/end on SYN_REPORT instead.
https://bugs.freedesktop.org/show_bug.cgi?id=89949
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Change vector_get_direction input to a normalized_coords type rather than
passing in a separate x,y pair, and rename it normalized_get_direction to
match. Since it now depends on the normalized_coords type which gets declared
in libinput-private.h also move it to libinput-private.h .
Note this commit also contains a functional change wrt the get_direction
usuage in the palm detection. The palm-detection code was calling get_direction
on non normalized coordinates, this commits changes the code to normalize
the coordinates first. This is the right thing to do as calling get_direction
on non normalized coordinates may result in a wrong direction getting returned
when the x and y resolution of the touchpad are not identical.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
What we really need is not a specific delta type, but a type which can hold
non discrete device coordinates, this is e.g. also needed for the center
coordinates of gestures. So rename delta_coords to device_float_coords to
properly reflect what we really need.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Change tp_filter_motion to use normalized_coords, rather then having it take
separate x and y values.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Add a normalized_is_zero helper function, and use it where applicable.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
tp_normalize_coords is one of the last functions taking separate x, y
values rather a coordinate pair, this commit cleans this up.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
This reverts commit 0e64837f30.
Rather than a customized touchpad property, let udev handle this and set the
absinfo struct during the normal setup procedures. No need for libinput to
have a custom workaround here.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
This reverts commit d101d43dd0.
This commit was accidentally merged, it was sitting on top of the wip branch.
With the 0e64837f30 commit this isn't necessary
anymore.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
A large range of the Apple touchpads seem to have the same dimensions. Use
that fact to force-set a resolution.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Not all touchpad kernel drivers supply the x/y resolution. Let the udev hwdb
fix this up where possible and read the value from it.
This is intentionally only used on touchpads, touchscreen devices without
resolution should be considered buggy and fixed in the kernel.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
This way the unaccelerated deltas returned by libinput are correct.
To maintain the current behavior we slow down the input speed by the magic
factor and likewise the accelerated output speed. This produces virtually the
same accelerated deltas as the previous code.
The magic factor is applied to the default denominator for guessing a
resolution based on the touchpad diagonal. We can't really get around this
without having a resolution from the touchpad; meanwhile this produces
virtually the same coordinates before/after.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
The struct evdev_device's absinfo_x/y point to the right axis
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
