Matching in vendor and product covers many models as possible.
01E8 product is always haptic.
01E0 can be or not haptic, leave it outside this and mantain it per
system model. When the kernel detects haptic touchpads the ones that
cannot be differenciated won't need to have quirk neither.
Signed-off-by: David Santamaría Rogado <howl.nsp@gmail.com>
Part-of: <https://gitlab.freedesktop.org/libinput/libinput/-/merge_requests/1383>
This quirk is a generic one for all the HP laptops with haptic touchpad
so makes more sense here because we are applying it dmi independent
being more difficult to track this change if the touchpad became used in
other vendors.
Signed-off-by: David Santamaría Rogado <howl.nsp@gmail.com>
Part-of: <https://gitlab.freedesktop.org/libinput/libinput/-/merge_requests/1378>
The scrollwheels are similar to the MX Master 3, and need the same
quirks: the horizontal wheel events are inverted (it scrolls "naturally"
by default while the horizontal scrollwheel direction is "traditional"),
and middle-clicking without scrolling is very difficult with high-res
scroll events (from the hid_logitech_hidpp kernel module) enabled.
This adds the device ID seen through Bluetooth, which seems to be the
only one we can add a quirk for:
- When connected using the Bolt receiver, there is no separate device ID
for the mouse (just the same 046d:c548 ID for the receiver already
documented as supporting multiple mice).
- When connected through USB, the mouse charges but does not provide HID
events through USB (it can be used while charging but only by using a
separate Bluetooth or Bolt connection for HID).
Part-of: <https://gitlab.freedesktop.org/libinput/libinput/-/merge_requests/1362>
These devices had the debouncing disabled via a model quirk but
really they are virtual devices and should have all hw-specific
processing disabled - on the assumption that this will be handled
in the host. See also e.g. commit 5d23794d53 ("tablet: disable
smoothing for uinput devices").
Closes#1175
Part-of: <https://gitlab.freedesktop.org/libinput/libinput/-/merge_requests/1308>
HP OmniBook Ultra Flip Laptop 14-fh0xxx manages itself keyboard and
touchpad deactivation when HP's custom Intel ISH firmware is installed
in the system. Without the custom firmware tablet-mode switch isn't
exposed so there is no way we don't need this.
More detailed information in the file comment.
Signed-off-by: David Santamaría Rogado <howl.nsp@gmail.com>
Part-of: <https://gitlab.freedesktop.org/libinput/libinput/-/merge_requests/1311>
The MX Master 3 is difficult, its wheel events are all over the place
and heuristics are tricky to determine. The previous plugin behavior
was seemingly sufficient for the MX Master but not for other devices.
Restore the old behavior if the quirk is set for a device by adding a
fourth state ALWAYS_ACCUMULATE. In this state the min movement is never
updated from the original threshold, causing any wheel motion to
accumulate.
Ref: ca6b82841c ("plugin/wheel: tighten the wheel debouncing code")
Ref: bb05e0d1b5 ("plugin/wheel: don't accumulate for low HID resolution multipliers")
Part-of: <https://gitlab.freedesktop.org/libinput/libinput/-/merge_requests/1262>
As I tested my friend's ThinkPad A485, I found that the default multiplier
resulted in jumpy cursor and a slightly too quick acceleration curve. Upon
checking for Lenovo quirks, I found that since commits 383a60abea
("Better Thinkpad T480 trackpoint multiplier") and a1566e3492 ("quirks:
Thinkpad T470 trackpoint multiplier"), the TrackPoint multiplier for both
T470 and T480 (which shares the same keyboard FRUs with the A485) were set
to 0.4.
However, per my testing, by setting the multiplier to 0.4, the TrackPoint
speed became so painfully slow that it began to hurt my index finger...
I suspect the original commiters have set custom acceleration curves on
their own system, but I might be wrong.
Playing with the multiplier, I found 0.75 to be the most appropriate and,
interestingly, with anything >= 0.8, the TrackPoint began to become jumpy,
with the cursor appearing to have a very low poll rate on the screen (the
higher the multiplier, the worse it gets).
Since, as I mentioned above, the keyboard and TrackPoint parts are shared
between these three models, I'm assuming that this multiplier will work
well for them.
Signed-off-by: Mingcong Bai <jeffbai@aosc.io>
Part-of: <https://gitlab.freedesktop.org/libinput/libinput/-/merge_requests/1149>
