Documentation/DocBook/uio-howto.tmpl

   1 <?xml version="1.0" encoding="UTF-8"?>
   2 <!DOCTYPE book PUBLIC "-//OASIS//DTD DocBook XML V4.2//EN"
   3 "http://www.oasis-open.org/docbook/xml/4.2/docbookx.dtd" []>
   4
   5 <book id="index">
   6 <bookinfo>
   7 <title>The Userspace I/O HOWTO</title>
   8
   9 <author>
  10       <firstname>Hans-Jürgen</firstname>
  11       <surname>Koch</surname>
  12       <authorblurb><para>Linux developer, Linutronix</para></authorblurb>
  13         <affiliation>
  14         <orgname>
  15                 <ulink url="http://www.linutronix.de">Linutronix</ulink>
  16         </orgname>
  17
  18         <address>
  19            <email>hjk@linutronix.de</email>
  20         </address>
  21     </affiliation>
  22 </author>
  23
  24 <pubdate>2006-12-11</pubdate>
  25
  26 <abstract>
  27         <para>This HOWTO describes concept and usage of Linux kernel's
  28                 Userspace I/O system.</para>
  29 </abstract>
  30
  31 <revhistory>
  32         <revision>
  33         <revnumber>0.3</revnumber>
  34         <date>2007-04-29</date>
  35         <authorinitials>hjk</authorinitials>
  36         <revremark>Added section about userspace drivers.</revremark>
  37         </revision>
  38         <revision>
  39         <revnumber>0.2</revnumber>
  40         <date>2007-02-13</date>
  41         <authorinitials>hjk</authorinitials>
  42         <revremark>Update after multiple mappings were added.</revremark>
  43         </revision>
  44         <revision>
  45         <revnumber>0.1</revnumber>
  46         <date>2006-12-11</date>
  47         <authorinitials>hjk</authorinitials>
  48         <revremark>First draft.</revremark>
  49         </revision>
  50 </revhistory>
  51 </bookinfo>
  52
  53 <chapter id="aboutthisdoc">
  54 <?dbhtml filename="about.html"?>
  55 <title>About this document</title>
  56
  57 <sect1 id="copyright">
  58 <?dbhtml filename="copyright.html"?>
  59 <title>Copyright and License</title>
  60 <para>
  61       Copyright (c) 2006 by Hans-Jürgen Koch.</para>
  62 <para>
  63 This documentation is Free Software licensed under the terms of the
  64 GPL version 2.
  65 </para>
  66 </sect1>
  67
  68 <sect1 id="translations">
  69 <?dbhtml filename="translations.html"?>
  70 <title>Translations</title>
  71
  72 <para>If you know of any translations for this document, or you are
  73 interested in translating it, please email me
  74 <email>hjk@linutronix.de</email>.
  75 </para>
  76 </sect1>
  77
  78 <sect1 id="preface">
  79 <title>Preface</title>
  80         <para>
  81         For many types of devices, creating a Linux kernel driver is
  82         overkill.  All that is really needed is some way to handle an
  83         interrupt and provide access to the memory space of the
  84         device.  The logic of controlling the device does not
  85         necessarily have to be within the kernel, as the device does
  86         not need to take advantage of any of other resources that the
  87         kernel provides.  One such common class of devices that are
  88         like this are for industrial I/O cards.
  89         </para>
  90         <para>
  91         To address this situation, the userspace I/O system (UIO) was
  92         designed.  For typical industrial I/O cards, only a very small
  93         kernel module is needed. The main part of the driver will run in
  94         user space. This simplifies development and reduces the risk of
  95         serious bugs within a kernel module.
  96         </para>
  97 </sect1>
  98
  99 <sect1 id="thanks">
 100 <title>Acknowledgments</title>
 101         <para>I'd like to thank Thomas Gleixner and Benedikt Spranger of
 102         Linutronix, who have not only written most of the UIO code, but also
 103         helped greatly writing this HOWTO by giving me all kinds of background
 104         information.</para>
 105 </sect1>
 106
 107 <sect1 id="feedback">
 108 <title>Feedback</title>
 109         <para>Find something wrong with this document? (Or perhaps something
 110         right?) I would love to hear from you. Please email me at
 111         <email>hjk@linutronix.de</email>.</para>
 112 </sect1>
 113 </chapter>
 114
 115 <chapter id="about">
 116 <?dbhtml filename="about.html"?>
 117 <title>About UIO</title>
 118
 119 <para>If you use UIO for your card's driver, here's what you get:</para>
 120
 121 <itemizedlist>
 122 <listitem>
 123         <para>only one small kernel module to write and maintain.</para>
 124 </listitem>
 125 <listitem>
 126         <para>develop the main part of your driver in user space,
 127         with all the tools and libraries you're used to.</para>
 128 </listitem>
 129 <listitem>
 130         <para>bugs in your driver won't crash the kernel.</para>
 131 </listitem>
 132 <listitem>
 133         <para>updates of your driver can take place without recompiling
 134         the kernel.</para>
 135 </listitem>
 136 <listitem>
 137         <para>if you need to keep some parts of your driver closed source,
 138         you can do so without violating the GPL license on the kernel.</para>
 139 </listitem>
 140 </itemizedlist>
 141
 142 <sect1 id="how_uio_works">
 143 <title>How UIO works</title>
 144         <para>
 145         Each UIO device is accessed through a device file and several
 146         sysfs attribute files. The device file will be called
 147         <filename>/dev/uio0</filename> for the first device, and
 148         <filename>/dev/uio1</filename>, <filename>/dev/uio2</filename>
 149         and so on for subsequent devices.
 150         </para>
 151
 152         <para><filename>/dev/uioX</filename> is used to access the
 153         address space of the card. Just use
 154         <function>mmap()</function> to access registers or RAM
 155         locations of your card.
 156         </para>
 157
 158         <para>
 159         Interrupts are handled by reading from
 160         <filename>/dev/uioX</filename>. A blocking
 161         <function>read()</function> from
 162         <filename>/dev/uioX</filename> will return as soon as an
 163         interrupt occurs. You can also use
 164         <function>select()</function> on
 165         <filename>/dev/uioX</filename> to wait for an interrupt. The
 166         integer value read from <filename>/dev/uioX</filename>
 167         represents the total interrupt count. You can use this number
 168         to figure out if you missed some interrupts.
 169         </para>
 170
 171         <para>
 172         To handle interrupts properly, your custom kernel module can
 173         provide its own interrupt handler. It will automatically be
 174         called by the built-in handler.
 175         </para>
 176
 177         <para>
 178         For cards that don't generate interrupts but need to be
 179         polled, there is the possibility to set up a timer that
 180         triggers the interrupt handler at configurable time intervals.
 181         See <filename>drivers/uio/uio_dummy.c</filename> for an
 182         example of this technique.
 183         </para>
 184
 185         <para>
 186         Each driver provides attributes that are used to read or write
 187         variables. These attributes are accessible through sysfs
 188         files.  A custom kernel driver module can add its own
 189         attributes to the device owned by the uio driver, but not added
 190         to the UIO device itself at this time.  This might change in the
 191         future if it would be found to be useful.
 192         </para>
 193
 194         <para>
 195         The following standard attributes are provided by the UIO
 196         framework:
 197         </para>
 198 <itemizedlist>
 199 <listitem>
 200         <para>
 201         <filename>name</filename>: The name of your device. It is
 202         recommended to use the name of your kernel module for this.
 203         </para>
 204 </listitem>
 205 <listitem>
 206         <para>
 207         <filename>version</filename>: A version string defined by your
 208         driver. This allows the user space part of your driver to deal
 209         with different versions of the kernel module.
 210         </para>
 211 </listitem>
 212 <listitem>
 213         <para>
 214         <filename>event</filename>: The total number of interrupts
 215         handled by the driver since the last time the device node was
 216         read.
 217         </para>
 218 </listitem>
 219 </itemizedlist>
 220 <para>
 221         These attributes appear under the
 222         <filename>/sys/class/uio/uioX</filename> directory.  Please
 223         note that this directory might be a symlink, and not a real
 224         directory.  Any userspace code that accesses it must be able
 225         to handle this.
 226 </para>
 227 <para>
 228         Each UIO device can make one or more memory regions available for
 229         memory mapping. This is necessary because some industrial I/O cards
 230         require access to more than one PCI memory region in a driver.
 231 </para>
 232 <para>
 233         Each mapping has its own directory in sysfs, the first mapping
 234         appears as <filename>/sys/class/uio/uioX/maps/map0/</filename>.
 235         Subsequent mappings create directories <filename>map1/</filename>,
 236         <filename>map2/</filename>, and so on. These directories will only
 237         appear if the size of the mapping is not 0.
 238 </para>
 239 <para>
 240         Each <filename>mapX/</filename> directory contains two read-only files
 241         that show start address and size of the memory:
 242 </para>
 243 <itemizedlist>
 244 <listitem>
 245         <para>
 246         <filename>addr</filename>: The address of memory that can be mapped.
 247         </para>
 248 </listitem>
 249 <listitem>
 250         <para>
 251         <filename>size</filename>: The size, in bytes, of the memory
 252         pointed to by addr.
 253         </para>
 254 </listitem>
 255 </itemizedlist>
 256
 257 <para>
 258         From userspace, the different mappings are distinguished by adjusting
 259         the <varname>offset</varname> parameter of the
 260         <function>mmap()</function> call. To map the memory of mapping N, you
 261         have to use N times the page size as your offset:
 262 </para>
 263 <programlisting format="linespecific">
 264 offset = N * getpagesize();
 265 </programlisting>
 266
 267 </sect1>
 268 </chapter>
 269
 270 <chapter id="using-uio_dummy" xreflabel="Using uio_dummy">
 271 <?dbhtml filename="using-uio_dummy.html"?>
 272 <title>Using uio_dummy</title>
 273         <para>
 274         Well, there is no real use for uio_dummy. Its only purpose is
 275         to test most parts of the UIO system (everything except
 276         hardware interrupts), and to serve as an example for the
 277         kernel module that you will have to write yourself.
 278         </para>
 279
 280 <sect1 id="what_uio_dummy_does">
 281 <title>What uio_dummy does</title>
 282         <para>
 283         The kernel module <filename>uio_dummy.ko</filename> creates a
 284         device that uses a timer to generate periodic interrupts. The
 285         interrupt handler does nothing but increment a counter. The
 286         driver adds two custom attributes, <varname>count</varname>
 287         and <varname>freq</varname>, that appear under
 288         <filename>/sys/devices/platform/uio_dummy/</filename>.
 289         </para>
 290
 291         <para>
 292         The attribute <varname>count</varname> can be read and
 293         written.  The associated file
 294         <filename>/sys/devices/platform/uio_dummy/count</filename>
 295         appears as a normal text file and contains the total number of
 296         timer interrupts. If you look at it (e.g. using
 297         <function>cat</function>), you'll notice it is slowly counting
 298         up.
 299         </para>
 300
 301         <para>
 302         The attribute <varname>freq</varname> can be read and written.
 303         The content of
 304         <filename>/sys/devices/platform/uio_dummy/freq</filename>
 305         represents the number of system timer ticks between two timer
 306         interrupts. The default value of <varname>freq</varname> is
 307         the value of the kernel variable <varname>HZ</varname>, which
 308         gives you an interval of one second. Lower values will
 309         increase the frequency. Try the following:
 310         </para>
 311 <programlisting format="linespecific">
 312 cd /sys/devices/platform/uio_dummy/
 313 echo 100 > freq
 314 </programlisting>
 315         <para>
 316         Use <function>cat count</function> to see how the interrupt
 317         frequency changes.
 318         </para>
 319 </sect1>
 320 </chapter>
 321
 322 <chapter id="custom_kernel_module" xreflabel="Writing your own kernel module">
 323 <?dbhtml filename="custom_kernel_module.html"?>
 324 <title>Writing your own kernel module</title>
 325         <para>
 326         Please have a look at <filename>uio_dummy.c</filename> as an
 327         example. The following paragraphs explain the different
 328         sections of this file.
 329         </para>
 330
 331 <sect1 id="uio_info">
 332 <title>struct uio_info</title>
 333         <para>
 334         This structure tells the framework the details of your driver,
 335         Some of the members are required, others are optional.
 336         </para>
 337
 338 <itemizedlist>
 339 <listitem><para>
 340 <varname>char *name</varname>: Required. The name of your driver as
 341 it will appear in sysfs. I recommend using the name of your module for this.
 342 </para></listitem>
 343
 344 <listitem><para>
 345 <varname>char *version</varname>: Required. This string appears in
 346 <filename>/sys/class/uio/uioX/version</filename>.
 347 </para></listitem>
 348
 349 <listitem><para>
 350 <varname>struct uio_mem mem[ MAX_UIO_MAPS ]</varname>: Required if you
 351 have memory that can be mapped with <function>mmap()</function>. For each
 352 mapping you need to fill one of the <varname>uio_mem</varname> structures.
 353 See the description below for details.
 354 </para></listitem>
 355
 356 <listitem><para>
 357 <varname>long irq</varname>: Required. If your hardware generates an
 358 interrupt, it's your modules task to determine the irq number during
 359 initialization. If you don't have a hardware generated interrupt but
 360 want to trigger the interrupt handler in some other way, set
 361 <varname>irq</varname> to <varname>UIO_IRQ_CUSTOM</varname>. The
 362 uio_dummy module does this as it triggers the event mechanism in a timer
 363 routine. If you had no interrupt at all, you could set
 364 <varname>irq</varname> to <varname>UIO_IRQ_NONE</varname>, though this
 365 rarely makes sense.
 366 </para></listitem>
 367
 368 <listitem><para>
 369 <varname>unsigned long irq_flags</varname>: Required if you've set
 370 <varname>irq</varname> to a hardware interrupt number. The flags given
 371 here will be used in the call to <function>request_irq()</function>.
 372 </para></listitem>
 373
 374 <listitem><para>
 375 <varname>int (*mmap)(struct uio_info *info, struct vm_area_struct
 376 *vma)</varname>: Optional. If you need a special
 377 <function>mmap()</function> function, you can set it here. If this
 378 pointer is not NULL, your <function>mmap()</function> will be called
 379 instead of the built-in one.
 380 </para></listitem>
 381
 382 <listitem><para>
 383 <varname>int (*open)(struct uio_info *info, struct inode *inode)
 384 </varname>: Optional. You might want to have your own
 385 <function>open()</function>, e.g. to enable interrupts only when your
 386 device is actually used.
 387 </para></listitem>
 388
 389 <listitem><para>
 390 <varname>int (*release)(struct uio_info *info, struct inode *inode)
 391 </varname>: Optional. If you define your own
 392 <function>open()</function>, you will probably also want a custom
 393 <function>release()</function> function.
 394 </para></listitem>
 395 </itemizedlist>
 396
 397 <para>
 398 Usually, your device will have one or more memory regions that can be mapped
 399 to user space. For each region, you have to set up a
 400 <varname>struct uio_mem</varname> in the <varname>mem[]</varname> array.
 401 Here's a description of the fields of <varname>struct uio_mem</varname>:
 402 </para>
 403
 404 <itemizedlist>
 405 <listitem><para>
 406 <varname>int memtype</varname>: Required if the mapping is used. Set this to
 407 <varname>UIO_MEM_PHYS</varname> if you you have physical memory on your
 408 card to be mapped. Use <varname>UIO_MEM_LOGICAL</varname> for logical
 409 memory (e.g. allocated with <function>kmalloc()</function>). There's also
 410 <varname>UIO_MEM_VIRTUAL</varname> for virtual memory.
 411 </para></listitem>
 412
 413 <listitem><para>
 414 <varname>unsigned long addr</varname>: Required if the mapping is used.
 415 Fill in the address of your memory block. This address is the one that
 416 appears in sysfs.
 417 </para></listitem>
 418
 419 <listitem><para>
 420 <varname>unsigned long size</varname>: Fill in the size of the
 421 memory block that <varname>addr</varname> points to. If <varname>size</varname>
 422 is zero, the mapping is considered unused. Note that you
 423 <emphasis>must</emphasis> initialize <varname>size</varname> with zero for
 424 all unused mappings.
 425 </para></listitem>
 426
 427 <listitem><para>
 428 <varname>void *internal_addr</varname>: If you have to access this memory
 429 region from within your kernel module, you will want to map it internally by
 430 using something like <function>ioremap()</function>. Addresses
 431 returned by this function cannot be mapped to user space, so you must not
 432 store it in <varname>addr</varname>. Use <varname>internal_addr</varname>
 433 instead to remember such an address.
 434 </para></listitem>
 435 </itemizedlist>
 436
 437 <para>
 438 Please do not touch the <varname>kobj</varname> element of
 439 <varname>struct uio_mem</varname>! It is used by the UIO framework
 440 to set up sysfs files for this mapping. Simply leave it alone.
 441 </para>
 442 </sect1>
 443
 444 <sect1 id="adding_irq_handler">
 445 <title>Adding an interrupt handler</title>
 446         <para>
 447         What you need to do in your interrupt handler depends on your
 448         hardware and on how you want to handle it. You should try to
 449         keep the amount of code in your kernel interrupt handler low.
 450         If your hardware requires no action that you
 451         <emphasis>have</emphasis> to perform after each interrupt,
 452         then your handler can be empty.</para> <para>If, on the other
 453         hand, your hardware <emphasis>needs</emphasis> some action to
 454         be performed after each interrupt, then you
 455         <emphasis>must</emphasis> do it in your kernel module. Note
 456         that you cannot rely on the userspace part of your driver. Your
 457         userspace program can terminate at any time, possibly leaving
 458         your hardware in a state where proper interrupt handling is
 459         still required.
 460         </para>
 461
 462         <para>
 463         There might also be applications where you want to read data
 464         from your hardware at each interrupt and buffer it in a piece
 465         of kernel memory you've allocated for that purpose.  With this
 466         technique you could avoid loss of data if your userspace
 467         program misses an interrupt.
 468         </para>
 469
 470         <para>
 471         A note on shared interrupts: Your driver should support
 472         interrupt sharing whenever this is possible. It is possible if
 473         and only if your driver can detect whether your hardware has
 474         triggered the interrupt or not. This is usually done by looking
 475         at an interrupt status register. If your driver sees that the
 476         IRQ bit is actually set, it will perform its actions, and the
 477         handler returns IRQ_HANDLED. If the driver detects that it was
 478         not your hardware that caused the interrupt, it will do nothing
 479         and return IRQ_NONE, allowing the kernel to call the next
 480         possible interrupt handler.
 481         </para>
 482
 483         <para>
 484         If you decide not to support shared interrupts, your card
 485         won't work in computers with no free interrupts. As this
 486         frequently happens on the PC platform, you can save yourself a
 487         lot of trouble by supporting interrupt sharing.
 488         </para>
 489 </sect1>
 490
 491 </chapter>
 492
 493 <chapter id="userspace_driver" xreflabel="Writing a driver in user space">
 494 <?dbhtml filename="userspace_driver.html"?>
 495 <title>Writing a driver in userspace</title>
 496         <para>
 497         Once you have a working kernel module for your hardware, you can
 498         write the userspace part of your driver. You don't need any special
 499         libraries, your driver can be written in any reasonable language,
 500         you can use floating point numbers and so on. In short, you can
 501         use all the tools and libraries you'd normally use for writing a
 502         userspace application.
 503         </para>
 504
 505 <sect1 id="getting_uio_information">
 506 <title>Getting information about your UIO device</title>
 507         <para>
 508         Information about all UIO devices is available in sysfs. The
 509         first thing you should do in your driver is check
 510         <varname>name</varname> and <varname>version</varname> to
 511         make sure your talking to the right device and that its kernel
 512         driver has the version you expect.
 513         </para>
 514         <para>
 515         You should also make sure that the memory mapping you need
 516         exists and has the size you expect.
 517         </para>
 518         <para>
 519         There is a tool called <varname>lsuio</varname> that lists
 520         UIO devices and their attributes. It is available here:
 521         </para>
 522         <para>
 523         <ulink url="http://www.osadl.org/projects/downloads/UIO/user/">
 524                 http://www.osadl.org/projects/downloads/UIO/user/</ulink>
 525         </para>
 526         <para>
 527         With <varname>lsuio</varname> you can quickly check if your
 528         kernel module is loaded and which attributes it exports.
 529         Have a look at the manpage for details.
 530         </para>
 531         <para>
 532         The source code of <varname>lsuio</varname> can serve as an
 533         example for getting information about an UIO device.
 534         The file <filename>uio_helper.c</filename> contains a lot of
 535         functions you could use in your userspace driver code.
 536         </para>
 537 </sect1>
 538
 539 <sect1 id="mmap_device_memory">
 540 <title>mmap() device memory</title>
 541         <para>
 542         After you made sure you've got the right device with the
 543         memory mappings you need, all you have to do is to call
 544         <function>mmap()</function> to map the device's memory
 545         to userspace.
 546         </para>
 547         <para>
 548         The parameter <varname>offset</varname> of the
 549         <function>mmap()</function> call has a special meaning
 550         for UIO devices: It is used to select which mapping of
 551         your device you want to map. To map the memory of
 552         mapping N, you have to use N times the page size as
 553         your offset:
 554         </para>
 555 <programlisting format="linespecific">
 556         offset = N * getpagesize();
 557 </programlisting>
 558         <para>
 559         N starts from zero, so if you've got only one memory
 560         range to map, set <varname>offset = 0</varname>.
 561         A drawback of this technique is that memory is always
 562         mapped beginning with its start address.
 563         </para>
 564 </sect1>
 565
 566 <sect1 id="wait_for_interrupts">
 567 <title>Waiting for interrupts</title>
 568         <para>
 569         After you successfully mapped your devices memory, you
 570         can access it like an ordinary array. Usually, you will
 571         perform some initialization. After that, your hardware
 572         starts working and will generate an interrupt as soon
 573         as it's finished, has some data available, or needs your
 574         attention because an error occured.
 575         </para>
 576         <para>
 577         <filename>/dev/uioX</filename> is a read-only file. A
 578         <function>read()</function> will always block until an
 579         interrupt occurs. There is only one legal value for the
 580         <varname>count</varname> parameter of
 581         <function>read()</function>, and that is the size of a
 582         signed 32 bit integer (4). Any other value for
 583         <varname>count</varname> causes <function>read()</function>
 584         to fail. The signed 32 bit integer read is the interrupt
 585         count of your device. If the value is one more than the value
 586         you read the last time, everything is OK. If the difference
 587         is greater than one, you missed interrupts.
 588         </para>
 589         <para>
 590         You can also use <function>select()</function> on
 591         <filename>/dev/uioX</filename>.
 592         </para>
 593 </sect1>
 594
 595 </chapter>
 596
 597 <appendix id="app1">
 598 <title>Further information</title>
 599 <itemizedlist>
 600         <listitem><para>
 601                         <ulink url="http://www.osadl.org">
 602                                 OSADL homepage.</ulink>
 603                 </para></listitem>
 604         <listitem><para>
 605                 <ulink url="http://www.linutronix.de">
 606                  Linutronix homepage.</ulink>
 607                 </para></listitem>
 608 </itemizedlist>
 609 </appendix>
 610
 611 </book>