linux-old/Documentation/input/ff.txt
<<
>>
Prefs
   1Force feedback for Linux.
   2By Johann Deneux <deneux@ifrance.com> on 2001/04/22.
   3
   4----------------------------------------------------------------------------
   5
   60. Introduction
   7~~~~~~~~~~~~~~~
   8This document describes how to use force feedback devices under Linux. The
   9goal is not to support these devices as if they were simple input-only devices
  10(as it is already the case), but to really enable the rendering of force
  11effects.
  12At the moment, only I-Force devices are supported, and not officially. That
  13means I had to find out how the protocol works on my own. Of course, the
  14information I managed to grasp is far from being complete, and I can not
  15guarranty that this driver will work for you.
  16This document only describes the force feedback part of the driver for I-Force
  17devices. Please read joystick.txt before reading further this document.
  18
  192. Instructions to the user
  20~~~~~~~~~~~~~~~~~~~~~~~~~~~
  21Here are instructions on how to compile and use the driver. In fact, this
  22driver is the normal iforce.o, input.o and evdev.o drivers written by Vojtech
  23Pavlik, plus additions to support force feedback.
  24
  25Before you start, let me WARN you that some devices shake violently during the
  26initialisation phase. This happens for example with my "AVB Top Shot Pegasus".
  27To stop this annoying behaviour, move you joystick to its limits. Anyway, you
  28should keep a hand on your device, in order to avoid it to brake down if
  29something goes wrong.
  30
  31At the kernel's compilation:
  32        - Enable IForce/Serial
  33        - Enable Event interface
  34
  35Compile the modules, install them.
  36
  37You also need inputattach.
  38
  39You then need to insert the modules into the following order:
  40% modprobe joydev
  41% modprobe serport
  42% modprobe iforce
  43% modprobe evdev
  44% ./inputattach -ifor $2 &      # Only for serial
  45For convenience, you may use the shell script named "ff" available from
  46the cvs tree of the Linux Console Project at sourceforge. You can also
  47retrieve it from http://www.esil.univ-mrs.fr/~jdeneux/projects/ff/.
  48If you are using USB, you don't need the inputattach step.
  49
  50Please check that you have all the /dev/input entries needed:
  51cd /dev
  52rm js*
  53mkdir input
  54mknod input/js0 c 13 0
  55mknod input/js1 c 13 1
  56mknod input/js2 c 13 2
  57mknod input/js3 c 13 3
  58ln -s input/js0 js0
  59ln -s input/js1 js1
  60ln -s input/js2 js2
  61ln -s input/js3 js3
  62
  63mknod input/event0 c 13 64
  64mknod input/event1 c 13 65
  65mknod input/event2 c 13 66
  66mknod input/event3 c 13 67
  67
  682.1 Does it work ?
  69~~~~~~~~~~~~~~~~~~
  70There is an utility called fftest that will allow you to test the driver.
  71% fftest /dev/eventXX
  72
  733. Instructions to the developper
  74~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
  75  All interactions are done using the event API. That is, you can use ioctl()
  76and write() on /dev/input/eventXX.
  77  This information is subject to change.
  78
  793.1 Querying device capabilities
  80~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
  81#include <linux/input.h>
  82#include <sys/ioctl.h>
  83
  84int ioctl(int file_descriptor, int request, unsigned long *features);
  85
  86"request" must be EVIOCGBIT(EV_FF, sizeof(unsigned long))
  87
  88Returns the features supported by the device. features is a bitfield with the
  89following bits:
  90- FF_X          has an X axis (should allways be the case)
  91- FF_Y          has an Y axis (usually not the case for wheels)
  92- FF_CONSTANT   can render constant force effects
  93- FF_PERIODIC   can render periodic effects (sine, ramp, square...)
  94- FF_SPRING     can simulate the presence of a spring
  95- FF_FRICTION   can simulate friction (aka drag, damper effect...)
  96- FF_RUMBLE     rumble effects (normally the only effect supported by rumble
  97                pads)
  98- 8 bits from FF_N_EFFECTS_0 containing the number of effects that can be
  99                simultaneously played.
 100
 1013.2 Uploading effects to the device
 102~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 103#include <linux/input.h>
 104#include <sys/ioctl.h>
 105 
 106int ioctl(int file_descriptor, int request, struct ff_effect *effect);
 107
 108"request" must be EVIOCSFF.
 109
 110"effect" points to a structure describing the effect to upload. The effect is
 111uploaded, but not played.
 112The content of effect may be modified. In particular, its field "id" is set
 113to the unique id assigned by the driver. This data is required for performing
 114some operations (removing an effect, controlling the playback).
 115See <linux/input.h> for a description of the ff_effect stuct.
 116
 1173.3 Removing an effect from the device
 118~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 119int ioctl(int fd, EVIOCRMFF, effect.id);
 120
 121This makes room for new effects in the device's memory. Please note this won't
 122stop the effect if it was playing.
 123
 1243.4 Controlling the playback of effects
 125~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 126Control of playing is done with write(). Below is an example:
 127
 128#include <linux/input.h>
 129#include <unistd.h>
 130
 131        struct input_event play;
 132        struct input_event stop;
 133        struct ff_effect effect;
 134        int fd;
 135...
 136        fd = open("/dev/input/eventXX", O_RDWR);
 137...
 138        /* Play three times */
 139        play.type = EV_FF;
 140        play.code = effect.id;
 141        play.value = 3;
 142        
 143        write(fd, (const void*) &play, sizeof(play));
 144...
 145        /* Stop an effect */
 146        stop.type = EV_FF;
 147        stop.code = effect.id;
 148        stop.value = 0;
 149        
 150        write(fd, (const void*) &play, sizeof(stop));
 151
 1523.5 Setting the gain
 153~~~~~~~~~~~~~~~~~~~~
 154Not all devices have the same strength. Therefore, users should set a gain
 155factor depending on how strong they want effects to be. This setting is
 156persistent accross access to the driver, so you should not care about it if
 157you are writing games, as another utility probably already set this for you.
 158
 159/* Set the gain of the device
 160int gain;               /* between 0 and 100 */
 161struct input_event ie;  /* structure used to communicate with the driver */
 162
 163ie.type = EV_FF;
 164ie.code = FF_GAIN;
 165ie.value = 0xFFFFUL * gain / 100;
 166
 167if (write(fd, &ie, sizeof(ie)) == -1)
 168        perror("set gain");
 169
 1703.6 Enabling/Disabling autocenter
 171~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 172The autocenter feature quite disturbs the rendering of effects in my opinion,
 173and I think it should be an effect, which computation depends on the game
 174type. But you can enable it if you want.
 175
 176int autocenter;         /* between 0 and 100 */
 177struct input_event ie;
 178
 179ie.type = EV_FF;
 180ie.code = FF_AUTOCENTER;
 181ie.value = 0xFFFFUL * autocenter / 100;
 182
 183if (write(fd, &ie, sizeof(ie)) == -1)
 184        perror("set auto-center");
 185
 186A value of 0 means "no auto-center".
 187
 1883.7 Dynamic update of an effect
 189~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 190This consists in changing some parameters of an effect while it's playing. The
 191driver currently does not support that. You still have the brute-force method,
 192which consists in erasing the effect and uploading the updated version. It
 193actually works pretty well. You don't need to stop-and-start the effect.
 194
 195
lxr.linux.no kindly hosted by Redpill Linpro AS, provider of Linux consulting and operations services since 1995.