libinput/doc/faqs.dox
Peter Hutterer 64c73d7ae6 timer: change the timer offset warning to a client bug
This looks like a libinput bug (well, it does say "libinput bug" on the
package) but it hasn't been that for a long time. The cause is almost always
insufficient motivation to call libinput_dispatch() by the caller, or at least
not doing it with the celerity libinput demands (and deserves, if I may say
so).

Up-, down- or side-grade it to a client bug, so the outrage can be
directed elsewhere, preferably away from me. And add a section to the
documentation, just in case someone actually reads this stuff.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Olivier Fourdan <ofourdan@redhat.com>
2018-02-06 19:29:26 +10:00

223 lines
9.8 KiB
Text

/**
@page faq FAQs - Frequently Asked Questions
Frequently asked questions about libinput.
@tableofcontents
@section faq_feature Why doesn't libinput support ...?
First, read @ref what_is_libinput If you have a feature that you think
libinput needs to support, please file a bug report. See @ref reporting_bugs
for more details.
@section faq_fast_mouse My mouse moves too fast, even at the slowest setting
This is a symptom of high-dpi mice (greater than 1000dpi). These devices
need a udev hwdb entry to normalize their motion. See @ref
motion_normalization for a detailed explanation.
@section faq_fast_trackpoint My trackpoint moves too slow or too fast
This is a symptom of an invalid trackpoint range. These devices need a udev
hwdb entry to specify the range available so libinput can adjust the pointer
acceleration accordingly. See @ref trackpoint_range for a detailed explanation.
@section faq_enable_tapping Why isn't touchpad tap-to-click enabled by default
See @ref tapping_default
@section faq_touchpad_pressure Why does my touchpad lose track of touches
The most common cause for this is an incorrect pressure threshold range.
See @ref touchpad_pressure for more info.
@section faq_kinetic_scrolling Kinetic scrolling does not work
The X.Org synaptics driver implemented kinetic scrolling in the driver. It
measures the scroll speed and once the finger leaves the touchpad the driver
keeps sending scroll events for a predetermined time. This effectively
provides for kinetic scrolling without client support but triggers an
unfixable [bug](https://bugs.freedesktop.org/show_bug.cgi?id=38909): the
client cannot know that the events are from a kinetic scroll source. Scroll
events in X are always sent to the current cursor position, a movement of the
cursor after lifting the finger will send the kinetic scroll events to the
new client, something the user does not usually expect. A key event during
the kinetic scroll procedure causes side-effects such as triggering zoom.
libinput does not implement kinetic scrolling for touchpads. Instead it
provides the libinput_event_pointer_get_axis_source() function that enables
callers to implement kinetic scrolling on a per-widget basis, see @ref
scroll_sources.
@section faq_gpl Is libinput GPL-licensed?
No, libinput is MIT licensed. The Linux kernel header file linux/input.h in
libinput's tree is provided to ensure the same behavior regardless of which
kernel version libinput is built on. It does not make libinput GPL-licensed.
@section faq_config_options Where is the configuration stored?
libinput does not store configuration options, it is up to the caller to
manage these and decide which configuration option to apply to each device.
This must be done at startup, after a resume and whenever a new device is
detected.
One commonly used way to configure libinput is to have the Wayland
compositor expose a compositor-specific configuration option. For example,
in a GNOME stack, the gnome-control-center modifies dconf entries. These
changes are read by mutter and applied to libinput. Changing these entries
via the gsettings commandline tool has the same effect.
Another commonly used way to configure libinput is to have xorg.conf.d
snippets. When libinput is used with the xf86-input-libinput driver in an
X.Org stack, these options are read on startup and apply to each device.
Changing properties at runtime with the xinput commandline tool has the same
effect.
In both cases, the selection of available options and how they are exposed
depends on the libinput caller (e.g. mutter or xf86-input-libinput).
@dotfile libinput-stack-gnome.gv
This has an effect on the availability of configuration options: if an
option is not exposed by the intermediary, it cannot be configured by the
client. Also some configuration options that are provided by the
intermediary may not be libinput-specific configuration options.
@section faq_configure_wayland How do I configure my device on Wayland?
See @ref faq_config_options Use the configuration tool provided by your
desktop environment (e.g. gnome-control-center) or direct access to your
desktop environment's configuration storage (e.g. gsettings).
@section faq_configure_xorg How do I configure my device on X?
See @ref faq_config_options If your desktop environment does not provide a
graphical configuration tool you can use an
<a href="https://www.x.org/archive/current/doc/man/man5/xorg.conf.5.xhtml">xorg.conf.d snippet</a>.
Usually, such a snippet looks like this:
<pre>
$> cat /etc/X11/xorg.conf.d/99-libinput-custom-config.conf
Section "InputClass"
Identifier "something to identify this snippet"
MatchDriver "libinput"
MatchProduct "substring of the device name"
Option "some option name" "the option value"
EndSection
</pre>
The identifier is merely a human-readable string that shows up in the log
file. The MatchProduct line should contain the device name or a substring of
the device name that the snippet should apply to. For a full list of option
names and permitted values, see the
<a href="https://www.mankier.com/4/libinput">libinput man page</a>.
xorg.conf.d snippets like the above apply to hotplugged devices but can be
overwritten at runtime by desktop tools. Multiple snippets may be placed
into the same file.
For run-time configuration and testing, the
<a href="https://www.x.org/archive/X11R7.5/doc/man/man1/xinput.1.html">xinput</a>
debugging tool can modify a devices' properties. See the
<a href="https://www.mankier.com/4/libinput">libinput man page</a>
for supported property names and values. Usually, an invocation looks like
this:
<pre>
$> xinput set-prop "the device name" "the property name" value [value2] [value3]
</pre>
@note
Changes performed by xinput do not persist across device hotplugs. xinput is
considered a debugging and testing tool only and should not be used for
permanent configurations.
@section faq_configuration Can you add a configuration option for $FEATURE?
No. At least that's going to be the initial answer. Read <a
href="http://who-t.blogspot.com/2016/04/why-libinput-doesnt-have-lot-of-config.html">Why
libinput doesn't have a lot of configuration options</a> first.
Configuration options for most features are a signal that we are incapable
of handling it correctly. To get to that point, we want to be sure we're
truly incapable of doing so. libinput has several features that
are handled automatically (and correctly) that users wanted to have
configuration options for initially.
So the answer to this question will almost always be 'no'. A configuration
option is, in most cases, a cop-out.
@section faq_synclient Why don't synclient and syndaemon work with libinput?
Synclient and syndaemon rely on X input device properties that are specific
to the xf86-input-synaptics X.Org input driver. Both were written when the
synaptics driver was the only commmon touchpad driver in existence. They
assume that if the properties aren't available, no touchpad is available
either. The xf86-input-libinput X.Org input driver does not export these
driver-specific properties, synclient/syndaemon will thus not detect the
touchpad and refuse to work. Other tools that rely on synclient/syndaemon or
those same properties also do not work with xf86-input-libinput.
Most of syndaemon's functionality is built into libinput, see @ref
disable-while-typing. synclient is merely a configuration tool, see
@ref faq_configure_xorg for similar functionality.
See also the blog posts
<a
href="http://who-t.blogspot.com.au/2017/01/the-definitive-guide-to-synclient.html">The
definitive guide to synclient</a> and
<a href="http://who-t.blogspot.com.au/2016/12/the-future-of-xinput-xmodmap-setxkbmap.html">
The future of xinput, xmodmap, setxkbmap, xsetwacom and other tools under
Wayland</a>
@section faq_tablets Does libinput support non-Wacom tablets?
Yes, though unfortunately many non-Wacom tablets suffer from bad firmware
and don't send the required events. But they should all work nonetheless. If
you have a tablet that does not work with libinput, please @ref
reporting_bugs "file a bug".
@section faq_tablet_capabilities My tablet doesn't work
If you see the message
<pre>
libinput bug: device does not meet tablet criteria. Ignoring this device.
</pre>
or the message
<pre>
missing tablet capabilities [...] Ignoring this device.
</pre>
your tablet device does not have the required capabilities to be treated as
a tablet. This is usually a problem with the device and the kernel driver.
See @ref tablet-capabilities for more details.
@section faq_hwdb_changes How to apply hwdb changes
Sometimes users are asked to test updates to the <a
href="https://www.freedesktop.org/software/systemd/man/hwdb.html">udev
hwdb</a> or patches that include a change to the hwdb. See @ref hwdb for
details on the hwdb and how to modify it locally.
@section faq_timer_offset What causes the "timer offset negative" warning?
libinput relies on the caller to call libinput_dispatch() whenever data is
available on the epoll-fd. Doing so will process the state of all devices
and can trigger some timers to be set (e.g. palm detection, tap-to-click,
disable-while-typing, etc.). Internally, libinput's time offsets are always
based on the event time of the triggering event.
For example, a touch event with time T may trigger a timer for the time T +
180ms. When setting a timer, libinput checks the wall clock time to ensure
that this time T + offset is still in the future. If not, the warning is
logged.
When this warning appears, it simply means that too much time has passed
between the event occurring (and the epoll-fd triggering) and the current
time. In almost all cases this is an indication of the caller being
overloaded and not handling events as speedily as required.
The warning has no immediate effect on libinput's behavior but some of the
functionality that relies on the timer may be impeded (e.g. palms are not
detected as they should be).
*/