Commit Graph

239 Commits

Author SHA1 Message Date
Scott Pleb 0bede93daa evdev: Correctly calculate axis range for min values greater than 0.
Axis range was previously calculated as max + abs(min), which relies on the assumption that
min will not exceed 0. For (min, max) values like (0, 255) or (-128, 127), which I assume to
be the most common cases, the range is correctly calculated as 255. However, given (20,
235), the range is erroneously calculated as 255, leading to axis values being normalized
incorrectly.

SDL already handles this case correctly. After changing the range calculation to max - min,
the axis values received from the evdev backend are practically identical to the values
received from the SDL backend.
2017-11-11 02:07:04 -05:00
Michael M 8e6677be90 ControllerInterface: don't call InvokeDevicesChangedCallbacks more than once when refreshing 2017-11-10 13:37:42 -08:00
Michael M fd7cbd633e ControllerInterface: add mutex around callbacks vector 2017-11-10 13:37:42 -08:00
Michael M 1ed7532af8 ControllerInterface: HotplugCallbacks -> DevicesChangedCallbacks 2017-11-10 13:37:42 -08:00
Michael M 7355b5f70d ControllerInterface: invoke callbacks in AddDevice/RemoveDevice
Some backends already cause this to happen, so make it consistent across
systems.
2017-11-10 13:37:41 -08:00
Arthur Carlsson 79a646a67d Prevent multiple HID elements of same usage type on OSX
On OSX, iterate the HID device's elements and only store the last of
each type to accommodate for flaky hardware
2017-09-15 19:19:46 +02:00
Lioncash 696e1b40b5 Common: Move version strings to their own header
Ideally Common.h wouldn't be a header in the Common library, and instead be renamed to something else, like PlatformCompatibility.h or something, but even then, there's still some things in the header that don't really fall under that label

This moves the version strings out to their own version header that doesn't dump a bunch of other unrelated things into scope, like what Common.h was doing.

This also places them into the Common namespace, as opposed to letting them sit in the global namespace.
2017-09-09 19:28:10 -04:00
Nick ef95bf26cb Implement GamePAd and MultiAxisController detection 2017-08-24 00:31:59 -04:00
Nick 3abde44641 formatting cleanup 2017-08-23 00:19:34 -04:00
Nick 3cddb62aea keyboard fixes v2 2017-08-22 12:26:44 -04:00
Nick 0df4f7db4f Depreciate old OSX Keyboard and Mouse 2017-08-21 01:44:03 -04:00
Leo Lam 918b0375e4 Merge pull request #5742 from ToadKing/fix-bad-controllers
Ignore capabilities reported by an XInput device
2017-08-17 03:42:02 +08:00
Michael Lelli 4fb2d1e361 Ignore capabilities reported by an XInput device 2017-08-09 22:51:01 -05:00
Jocelyn Turcotte 200fb26148 Fix axis not working with Switch Pro controller on macOS
Each axis would appear as multiple elements with 0 min/max.
Filter the list of elements using the correct usage page like
done for buttons.
2017-08-01 01:58:30 +02:00
Léo Lam 17ef4c8046 StringUtil: Make SplitString return by value
Simpler usage.
2017-06-11 16:48:20 +02:00
shuffle2 192fec50b9 Merge pull request #5276 from ligfx/macosheadless
Add headless support on macOS
2017-06-05 20:49:02 -07:00
Léo Lam d9fd056803 Fix minor formatting issues
These were not caught by the lint script while it was broken.
2017-06-05 02:32:19 +02:00
Lioncash 5862d1fc41 Device: Provide operator!= counterparts to operator== for DeviceQualifier
Makes comparison logic symmetric
2017-06-03 19:34:35 -04:00
Michael Maltese 5298328cb1 Add headless support on macOS 2017-04-15 19:34:42 -07:00
Steven Newbury 02681bc352 Add missing "functional" include
Building fails on GCC7 without an explicit
2017-02-20 14:25:55 +00:00
Michael Maltese a509f56116 InputCommon: Extract ControlReference from ControllerInterface
Better separation of concerns. Relegates `ControllerInterface` to
enumerating input controls, and the new `ControlReference` deals with
combining inputs and configuration expression parsing.
2017-02-07 22:59:10 -08:00
Michael Maltese 0bc40cacda No longer need to //clang-format off for Windows headers 2017-01-23 16:23:37 -08:00
Michael Maltese 713ec5ffd5 Add includes for building on Windows without PCH 2017-01-23 01:37:41 -08:00
hthh 387164a999 macOS: Add missing includes to fix no-PCH build 2017-01-02 11:45:24 +11:00
Tillmann Karras 57a005ff3e evdev: reduce startup time
GetName() creates a new evdev device which calls tons of ioctls. But the
main culprit is close() which for input devices appears to be a slow
path in the kernel.

This commit reduces PopulateDevices() by 50% on my laptop, but ~730 ms
is still ridiculously slow for something that isn't needed right away.
2016-12-26 01:24:35 +01:00
Tillmann Karras f575902cf3 evdev: set flag in a thread-safe way 2016-12-05 00:47:26 +00:00
Tillmann Karras 3c090a37d4 evdev: fix shutdown hang
Note: This probably means our approach of populating from udev isn't
safe to be called multiple times.
2016-12-05 00:47:26 +00:00
Anthony ffe20705dd Merge pull request #4362 from ligfx/osxhotplug
ControllerInterface: enable hotplugging on macOS
2016-12-01 00:29:56 -06:00
Michael Maltese 7ed8fb95c5 ControllerInterface: enable hotplugging on macOS 2016-11-30 22:27:46 -08:00
Michael Maltese 3e69d066f5 ControllerInterface: replace Reinitialize with RefreshDevices
The SDL backend crashes when you close a joystick after SDL_Quit has
been called. Some backends don't need to be shutdown and
re-initialized everytime, we can just ask to enumerate devices again.
2016-11-30 16:07:55 -08:00
Michael Maltese cebb4d84f2 ControllerInterface: clear devices before shutting down backends
The SDL backend faults if it tries to close a joystick after SDL_Quit
has been called.
2016-11-30 16:07:54 -08:00
Markus Wick b9e4f67d3a Merge pull request #4405 from xrix4096/mac-unknown-joyaxis
Make analog triggers on PS3 controllers accessible on Mac
2016-10-31 13:01:47 +01:00
Markus Wick 2afa877119 Merge pull request #4376 from leoetlino/warning-fix
Warning fixes
2016-10-31 12:57:03 +01:00
Chris Pritchard 9da9ba616b Use the IOHIDElement cookie as a part of the axis name for unknown axis. Previously the 'usage' value was used to identify the axis by name, but this is not unique. For example on a PS3 controller *all* axis other than the well known ones return a usage of '1' so there are 30 or more axis all named "1". This stops things such as analog triggers being usable.
Using the element cookie uniquely identifies each axis and allows them to be assigned successfully as controls
2016-10-30 10:36:53 +00:00
Sintendo f163bd1048 Fix various comment typos 2016-10-24 18:27:49 +02:00
Léo Lam 68156a02ed Fix -Wunused-result warnings 2016-10-23 00:43:08 +02:00
Léo Lam fad17246db InputCommon: Fix formatting issues in Quartz
This wasn't caught by the lint script because it was broken (by me) at
some point :(

If it had been rebased, it wouldn't have passed lint
2016-10-03 18:59:52 +02:00
shuffle2 e38ddfd787 Merge pull request #4111 from EmptyChaos/startup-perf-langs
InputCommon: Fix slow startup on some Windows systems with language packs (Issue 9744)
2016-10-03 04:02:54 -07:00
shuffle2 2747fd3a0d Merge pull request #4102 from ligfx/quartz_input
Add Quartz/CoreGraphics controller interface for default k&m events
2016-10-03 03:13:58 -07:00
Jasper St. Pierre 928d05ec47 InputCommon: Remove the Xlib backend
The XInput2 backend is more performant, so let's default to it and
remove the old, core-only backend.
2016-09-04 10:44:22 -07:00
Jasper St. Pierre 2b640a4f7d XInput2: Only do a round trip when the mouse has moved
This should cut down on round trips and blocking calls dramatically
during gameplay.
2016-09-04 10:44:22 -07:00
EmptyChaos 17e4bd9575 InputCommon: Fix slow startup on some Windows systems
Rewrite GetXInputGUIDS to use SetupAPI instead of WMI Queries. When
using a language pack where the system language and user/program
language differ, Windows starts taking a VERY long time (10+ seconds)
to complete Queries for Win32_PNPEntity objects (it's probably
translating every single string since it transfers every single one
from the WMI server into memory in the program).

Fixes Issue 9744.
2016-08-11 09:30:36 +10:00
Michael Maltese 0f5951e324 And pull KeycodeToName into its own function 2016-08-09 19:59:45 -07:00
Michael Maltese d1594f9529 use a std::map instead of running through the AOS 2016-08-09 19:54:59 -07:00
Michael Maltese a5cc054bd1 use std::to_string instead of sstream 2016-08-09 17:45:41 -07:00
Michael Maltese 6097d30288 clang-format and cleaning up Key::GetName 2016-08-09 17:23:30 -07:00
Michael Maltese 3889e4d4b2 style fixes 2016-08-09 15:52:02 -07:00
Michael Maltese 3dc8136e14 Add Quartz/CoreGraphics controller interface for default k&m events 2016-08-08 09:38:22 -07:00
Léo Lam 77fab5940c InputConfigDiag: Don't prevent closing dialog if device not found
There is no reason to prevent the user from closing the config dialog
if the device is not found. It's not very good UX…

Also fixes ExpressionParser to return NO_DEVICE if the device doesn't
exist instead of SUCCESS.
2016-07-30 10:00:55 +02:00
Léo Lam 135641404a evdev: Add hotplugging support
This adds hotplugging support to the evdev input backend. We use
libudev to monitor changes to input devices in a separate thread.
Removed devices are removed from the devices list, and new devices
are added to the list.

The effect is that controllers are usable immediately after plugging
them without having to manually refresh devices (if they were
configured to be used, of course).
2016-07-29 17:18:40 +02:00
Léo Lam 3926db624d ControllerInterface: Don't block on UpdateInput()
Changes UpdateInput() to skip if we can't lock the mutex, instead of
potentially blocking the CPU thread and causing a short but noticeable
frame drop.
2016-07-29 17:18:40 +02:00
Léo Lam 93f5df4195 ControllerInterface: Add RemoveDevice()
This adds RemoveDevice() to ControllerInterface, fixes ExpressionParser
and some other code to support device removals without crashing,
and adds an IsValid() method to Device, to prepare for hotplugging.
2016-07-29 17:18:37 +02:00
Léo Lam 0d783f0869 ControllerInterface: Add a way to register callbacks
This adds RegisterHotplugCallback() to register a callback which will
be invoked by the input backends' hotplug threads when there is a new
device, so that Core (GCKeyboard, GCPad, Wiimote, Hotkey) can reload
the configuration without adding a dependency to Core from InputCommon.
2016-07-29 17:16:22 +02:00
Léo Lam 788e19f54d ControllerInterface: Make the ID assigning code common
This makes the device ID assigning code common to all backends, by
moving it to AddDevice() instead of copy-pasting or replicating
the logic in the backends.

Also, to prepare for hotplugging, instead of relying on a name usage
count, the new ID assigning system always starts from ID 0 and tries
to assign the first ID that is not used.
2016-07-14 10:50:53 +02:00
Scott Mansell e0839a9478 Strip whitespace from Evdev and OSX controller names. 2016-07-12 14:19:07 +12:00
Léo Lam 6cc4591226 evdev: Clamp axis values to the 0.0-1.0 range
The values are expected to be in the 0.0-1.0 range (as indicated by the
comment), and other parts of Dolphin also expect it to be in that range
since the "full" axis has a -1.0 to 1.0 range.  However, this is not
always the case and fvalue can end up being outside of the range. This
clamps fvalue to always be in the 0.0 and 1.0 range.
2016-07-11 14:00:09 +02:00
Mat M cf50ed82bb Merge pull request #3943 from ShadowsFriend/nopch-fix
ControllerInterface: Include memory to fix building without pch
2016-06-26 02:13:47 -04:00
Lioncash 55916b963d Device: Use std::tie for DeviceQualifier's operator== 2016-06-25 23:45:37 -04:00
Christian Widmer be922730b3 ControllerInterface: Include memory to fix building without pch
All affected files use shared_ptr but do not include memory which breaks
building without precompiled headers.
2016-06-26 05:42:29 +02:00
Léo Lam 8678133e87 ControllerInterface: Switch to std::shared_ptr
Small cleanup by using std::shared_ptr and getting rid of
ciface.Devices() which just returned the m_devices (which defeats the
point of making m_devices protected).

Incidentally, this should make the code safer when we have
different threads accessing devices in the future (for hotplug?).

A lot of code use Device references directly so there is
no easy way to remove FindDevice() and make those unique_ptrs.
2016-06-25 21:46:39 +02:00
Léo Lam 9b075556ea Fix a small formatting issue 2016-06-25 13:46:53 +02:00
Léo Lam d3e2ae35ff ControllerInterface: Add synchronisation
Since we may have to add/access devices from different threads, this
adds synchronisation to anything that touches m_devices.
2016-06-25 13:46:53 +02:00
Léo Lam fd29e5c4cc ControllerInterface: Don't pass m_devices to the backends
Previously, the devices vector would be passed to all backends. They
would then manually push_back to it to add new devices. This was fine
but caused issues when trying to add synchronisation.

Instead, backends now call AddDevice() to fill m_devices so that it is
not accessible from the outside.
2016-06-25 13:46:53 +02:00
Pierre Bourdon 3570c7f03a Reformat all the things. Have fun with merge conflicts. 2016-06-24 10:43:46 +02:00
mimimi085181 fcd5170cff XInput: Apply Rumble/Motor output only on changes (again)
Disclaimer: I can't test if this works on xbox one controllers, i don't have one. But i have conformed that this UpdateMotors() is related to rumble for emulated wiimotes.

This partially reverts commit "XInput: Apply immediately as well" (1958a10b6f) from pr # https://github.com/dolphin-emu/dolphin/pull/1560

Hopefully this fixes the xbox one controller rumble issue:
https://bugs.dolphin-emu.org/issues/9071

And in theory it might reduce the used usb bandwidth, as it was originally intended before pr 1560.

@JMC47: Please do a good amount of testing, to see if this breaks rumble for wiimotes or gamecube controllers emulated with xinput devices.
2016-04-23 13:45:10 +02:00
Ryan Houdek 36af9981c7 Merge pull request #3595 from Sonicadvance1/Android_headless
[Android] Headless support
2016-02-06 20:08:58 -05:00
Nick Tiberi e7ad0fd600 normalize and check upper/lower bounds of hats input on OS X 2016-02-05 16:40:59 -05:00
Ryan Houdek 6305f09467 Use cmake to determine if we are enabling the Android Controller Interface 2016-02-05 10:54:17 -06:00
mathieui 64e2f93f38 [Pipe input] Make the commands locale-unaware
Setting axis values in an automated way should not be locale-aware as
strtod is.
2016-01-20 22:48:22 +01:00
booto b89c79b9f2 Comment change to reflect what code does 2016-01-10 02:42:56 +08:00
Ryan Houdek 74478addeb Move Android JNI bits from DolphinWX to the Android folder. 2016-01-06 15:36:44 -06:00
Phatcat f62054e467 Android: More analog input configurations
Lets the user set the following in intervals of 10 between 10 and 100;
- Stick/Radius (default 100,000000)
- Triggers/Threshold (default 90,000000)
- Tilt/Modifier/Range (default 50,000000) + mapped Tilt/Modifier button
to the configurations for wiimotes & nunchuks
2016-01-04 13:35:04 +01:00
skidau f96f90b334 Ignored the mouse position unless the cursor is over the Dolphin window.
Fixes issue 8673.
2015-12-24 09:44:26 +11:00
Phatcat ca9c9b7135 Wiimote + Extensions for Android 2015-12-02 21:58:54 +01:00
spxtr 4faf958fee Fix some bugs in Pipe input parsing.
If there were two commands in the buffer at once, it would only run the
first because of an error in UpdateInput.

If you sent the command "SET C" it would segfault because of a logic
issue in ParseCommand.
2015-11-03 20:26:23 -08:00
Lioncash e76b1f2d56 Merge pull request #3170 from spxtr/pipes
Implemented GC Controller inputs via named pipes.
2015-10-28 16:39:57 -04:00
spxtr d9d6cf8eda GC controller input using named pipes
Currently only works on unix, but can be extended to other systems. Can
also be extended to do wiimotes.

Searches the Pipes folder for readable named pipes and creates a dolphin
input device out of them. Send controller inputs to the game by writing
to the file. Commands are described in Pipes.h.
2015-10-24 20:20:03 -07:00
catzilla4 867a1beb9f Added some headers to allow Dolphin to compile in my environment. 2015-10-06 19:09:57 -05:00
Lioncash a30681834c InputCommon: Use NSNumber numeric literals in Obj-C code 2015-10-02 13:23:46 -04:00
Lioncash 19ac565e0d Common: Move asserts to their own header 2015-09-26 18:51:27 -04:00
Lioncash c3332604fa ControllerInterface: Fix -Winconsistent-missing-override warnings on OSX 2015-09-17 15:15:15 -04:00
Lioncash 88c9c6d086 ExpressionParser: Make all control paths return a value in operator std::string 2015-09-09 22:41:47 -04:00
Lioncash 22635c1800 Add missing override specifiers 2015-09-05 22:40:19 -04:00
Lioncash 633be0387d General: Remove unimplemented function prototypes 2015-09-05 22:01:07 -04:00
Tillmann Karras b3ff66dc7a evdev: don't pass null path to the kernel 2015-08-15 12:51:34 +02:00
Rohit Nirmal ac54bbe02d Fix building with PCH disabled. 2015-08-04 12:59:34 -05:00
Tillmann Karras 5a05187b3b Fix some warnings
reorder, sign-compare, pessimizing-move
2015-08-04 10:58:24 +02:00
Gabriel Corona 2722f3f337 Stable device identifier for evdev 2015-08-02 01:51:06 +02:00
Lioncash 7ee0e75633 Remove unnecessary virtual keywords 2015-07-30 10:33:08 -04:00
sigmabeta 1cb07b1c48 Android: Wiimote UI Fixup 2015-07-23 23:17:16 -04:00
Ryan Houdek a8227ad9b1 Add Wiimote support to the Android backend.
Not actually wired up to the Android UI for configuration.
2015-07-21 21:33:23 -05:00
Scott Mansell dc25277a2f Merge branch 'stable' into 'master' 2015-07-13 12:40:38 +12:00
Rohit Nirmal c38ae5236e Fix building with PCH disabled. 2015-07-08 13:52:43 -05:00
Scott Mansell 6e969133f5 evdev: Delete rumble effects on the correct file descriptor.
Was using m_id insted of m_fd.
Also re-arrange the code so rumble effects always get deleted instead of
just on stop commands.
2015-07-07 16:59:10 +12:00
Tillmann Karras 64b51df55d evdev: fix -Wunused-private-field warning 2015-07-04 14:14:15 +02:00
Scott Mansell d2578000bb Merge branch 'stable' 2015-07-04 23:43:36 +12:00
Scott Mansell 77183899e0 evdev: Support axis with ranges which extend below zero. 2015-07-01 10:13:05 +12:00
Scott Mansell 2721fdf8a9 Linux: Add an evdev based controller backend, to replace SDL. 2015-07-01 10:13:05 +12:00
comex 700b850acd Fix misc. clang warnings - mostly complaints about inconsistent use of override.
Previously, MacOpenFile only overrode anything on OS X; otherwise it was
just a useless method, which is presumably why it wasn't marked override
in the first place.  Address this more sanely by wrapping it in #ifdef
__APPLE__.
2015-06-21 15:42:49 -04:00
Matthew Parlane 9e1aab663f Revert "SDL: handle SDL_QUIT event" 2015-06-08 13:43:39 +12:00