This enables two-finger scrolling and two- and three-finger tapping on a
single-touch touchpad if BTN_TOOL_DOUBLETAP and BTN_TOOL_TRIPLETAP is set.
These require a bit of special processing:
BTN_TOUCH is set with the first finger down, but somewhat randomly unset and
re-set when switching between the various BTN_TOOL_*TAP values.
BTN_TOOL_<N>TAP is only set for N fingers down, thus a double->triple move
will see a release for DOUBLETAP and a press for TRIPLETAP. This may happen in
the same event, or across two consecutive events.
This patch adds a fake_touches mask to the touchpad struct. The mask is set
for each matching BTN_* event and used to count the number of expected
fake touchpoints. From that we begin/end the number of actual touchpoints
required. Fake touchpoints take their x/y coordinates from the first
touchpoint, which reads ABS_X/ABS_Y.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Touchpads without ABS_MT_SLOT create 5 slots by default (for up to QUINTTAP)
and ABS_X/Y is mapped to the 0-slot touchpoint. This commit adds handling for
a single finger, no BTN_TOOL_DOUBLETAP or similar is being processed yet.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
This is a fairly rough approach, but can be handled more fine-grained later.
Require a minimum of 1 unit to start scrolling and lock the scrolling in the
initial direction, so further scroll events are limited to that direction
only.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
On ClickPads (touchpads without phys. middle/right buttons) it is important to
know whether a physical click is queued up. The finger position or number of
fingers decide which button event to send.
This isn't currently used, we still just send the button number at the moment.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
The tapping state implementation will be in a separate file, so let's make
sure we can access the structs we need.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
If two fingers are down and moving, take the average movement of both fingers
and use that for scrolling.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
This patch is a mixture of an experimental project (libtouchpad) and
evdev-touchpad.c. It adds a new touchpad driver for multi-touch touchpads that
tracks each touchpoint separately. This makes it a lot easier to handle
multi-finger tapping, software button areas, etc.
libtouchpad used a slightly different coding style, this is the attempt to get
closer to the one used in libinput.
Currently sends motion events for single-finger motion, button events only for
physical buttons.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Doesn't do anything but initialize and destroy. This is not a permanent
separate implementation, it's just easier to start this way and then switch
over than to add to the current one.
Temporary measure: LIBINPUT_NEW_TOUCHPAD_DRIVER environment variable can be
used to enable the new driver
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
libinput is supposed to take a close callback in its interface
to allow you to call out to a privileged API to close FDs. But
the FD that libinput passes you is bogus, because
libinput_remove_source closes the FD on which it's passed. This
is really bad, as the libinput_source really doesn't own the FD
which it's passed, so it shouldn't be trying to close() it.
Only one out of the four users of libinput_remove_source actually
wants their FD closed, so move the close() call there.
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Instead of having one touch events representing different types of touch
events by providing a touch type, have one separate event type per touch
type. This means the LIBINPUT_EVENT_TYPE_TOUCH is replaced with
LIBINPUT_EVENT_TYPE_TOUCH_DOWN, LIBINPUT_EVENT_TYPE_TOUCH_MOTION,
LIBINPUT_EVENT_TYPE_TOUCH_UP and LIBINPUT_EVENT_TYPE_TOUCH_CANCEL.
Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Since a Wayland compositor have to represent all touch devices of a seat
as one virtual device, lets make that easier by also providing seat wide
slots with touch events.
Seat wide slots may be accessed using
libinput_event_touch_get_seat_slot().
Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Avoids erroneous timestamps when the system time is reset. This used to a be a
problem with the X.Org synaptics driver where taps, scrolling and a couple of
other things would potentially lock up.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
This gives us the ability to handle SYN_DROPPED transparently to the caller.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Jonas Ådahl <jadahl@gmail.com>
libevdev wraps the various peculiarities of the evdev kernel API into a
type-safe API. It also buffers the device so checking for specific features at
a later time is easier than re-issuing the ioctls. Plus, it gives us almost
free support for SYN_DROPPED events (in the following patch).
This patch switches all the bit checks over to libevdev and leaves the event
processing as-is. Makes it easier to review.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Jonas Ådahl <jadahl@gmail.com>
A caller may have a reference to the device after closing it, make sure that
ref doesn't have a dangling fd so future attempts of reading from/writing to
the device fail.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
X and Y are li_fixed_t, which is 24.8 fixed point real number.
li_fixed_t max is thus ~8388607.
On a touchscreen with a range of 32767 values (like a 3M sensor), and
mapped on monitor with a resolution of 1920x1080, we currently have:
(x - li_fixed_from_int(device->abs.min_x)) * width == 62912640
which is 7 times bigger than li_fixed_t max.
Force a cast to uint64_t to keep the precision of the sensor.
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>
Add a test case and test device that checks if the scale transform can
handle high resolution devices and output monitor resolutions.
The test case is created in a way that it will fail if the coordinate
transform expression will overflow if only 32 bit integer data
containers are used.
Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
The previous log handler wasn't actually hooked up to anything. Add a public
API for the log handler with priority filtering, defaulting to priority
'error' and stderr as output stream.
And to keep the diff down and convenience up, provide a few simple wrappers
for logging. The generic is log_msg(), but let's use log_info, log_error, etc.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Jonas Ådahl <jadahl@gmail.com>
It is unclear what "current" means as events are asynchronous, and
since a slot is associated with a touch event rather than a device,
change the description to reflect this.
Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
With this patch, a user can keep a reference to a libinput_seat
instance, which will cause the seat to never be unlinked from the
libinput context nor destroyed.
Previously, a when the last device of a seat was removed, the seat was
unlinked and if a new device was discovered with a previously empty seat
a new seat instance would always be created, meaning two potential seat
instances with identical physical and logical seat name pairs.
Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
If we don't have capabilities we can deal with, return a different
error so the backends can handle it separately (they already do).
Signe-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Creates an empty context that is not hooked up to a device. Callers can then
add and remove devices to this context using libinput_path_add_device() and
libinput_path_remove_device().
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
This allows multiple devices to share a single libinput context. The new
function returns the newly added device immediately. Unlike the udev seat
where devices may or may not be added - over the lifetime of the seat - a
path-based backend knows immediately if device exists or doesn't exist.
Returning the device is required by callers that have the event processing
separate from adding devices - by the time we have the DEVICE_ADDED event in
the queue we may have other events to process first. And the DEVICE_ADDED
event won't easily link to the path we gave it anyway, so it's hard to figure
out which DEVICE_ADDED event corresponds to the new device.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
The previous path backend created a libinput context attached to a single
device. This is insufficient when we need to use cross-device functionality.
One example of this cross-device functionality include disabling a touchpad
while the trackstick is in use (Lenovo T440 and related models).
This patch merely adds the infrastructure to support multiple devices for a
path backend. Follow-up patches add the function calls to add and remove
devices. This is needed by Xorg input drivers that still make use of the
server's device hotplug mechanisms but want to otherwise use libinput.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
This reverts commit e8c20c7241.
Ooops, bad rebase. This accesses the device after it was already destroyed
which is not the intent of the patch.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
If we don't have capabilities we can deal with, return a different error so
the backends can handle it separately (they already do).
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
