1 <?xml version="1.0" encoding="UTF-8"?>
2 <!DOCTYPE book PUBLIC "-//OASIS//DTD DocBook XML V4.1.2//EN"
3 "http://www.oasis-open.org/docbook/xml/4.1.2/docbookx.dtd" []>
5 <book id="USB-Gadget-API">
7 <title>USB Gadget API for Linux</title>
8 <date>20 August 2004</date>
9 <edition>20 August 2004</edition>
13 This documentation is free software; you can redistribute
14 it and/or modify it under the terms of the GNU General Public
15 License as published by the Free Software Foundation; either
16 version 2 of the License, or (at your option) any later
21 This program is distributed in the hope that it will be
22 useful, but WITHOUT ANY WARRANTY; without even the implied
23 warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
24 See the GNU General Public License for more details.
28 You should have received a copy of the GNU General Public
29 License along with this program; if not, write to the Free
30 Software Foundation, Inc., 59 Temple Place, Suite 330, Boston,
35 For more details see the file COPYING in the source
36 distribution of Linux.
40 <year>2003-2004</year>
41 <holder>David Brownell</holder>
45 <firstname>David</firstname>
46 <surname>Brownell</surname>
48 <address><email>dbrownell@users.sourceforge.net</email></address>
55 <chapter id="intro"><title>Introduction</title>
57 <para>This document presents a Linux-USB "Gadget"
59 API, for use within peripherals and other USB devices
61 It provides an overview of the API structure,
62 and shows how that fits into a system development project.
63 This is the first such API released on Linux to address
64 a number of important problems, including: </para>
67 <listitem><para>Supports USB 2.0, for high speed devices which
68 can stream data at several dozen megabytes per second.
70 <listitem><para>Handles devices with dozens of endpoints just as
71 well as ones with just two fixed-function ones. Gadget drivers
72 can be written so they're easy to port to new hardware.
74 <listitem><para>Flexible enough to expose more complex USB device
75 capabilities such as multiple configurations, multiple interfaces,
77 and alternate interface settings.
79 <listitem><para>USB "On-The-Go" (OTG) support, in conjunction
80 with updates to the Linux-USB host side.
82 <listitem><para>Sharing data structures and API models with the
83 Linux-USB host side API. This helps the OTG support, and
84 looks forward to more-symmetric frameworks (where the same
85 I/O model is used by both host and device side drivers).
87 <listitem><para>Minimalist, so it's easier to support new device
88 controller hardware. I/O processing doesn't imply large
89 demands for memory or CPU resources.
94 <para>Most Linux developers will not be able to use this API, since they
95 have USB "host" hardware in a PC, workstation, or server.
96 Linux users with embedded systems are more likely to
97 have USB peripheral hardware.
98 To distinguish drivers running inside such hardware from the
99 more familiar Linux "USB device drivers",
100 which are host side proxies for the real USB devices,
101 a different term is used:
102 the drivers inside the peripherals are "USB gadget drivers".
103 In USB protocol interactions, the device driver is the master
105 and the gadget driver is the slave (or "function driver").
108 <para>The gadget API resembles the host side Linux-USB API in that both
109 use queues of request objects to package I/O buffers, and those requests
110 may be submitted or canceled.
111 They share common definitions for the standard USB
112 <emphasis>Chapter 9</emphasis> messages, structures, and constants.
113 Also, both APIs bind and unbind drivers to devices.
114 The APIs differ in detail, since the host side's current
115 URB framework exposes a number of implementation details
116 and assumptions that are inappropriate for a gadget API.
117 While the model for control transfers and configuration
118 management is necessarily different (one side is a hardware-neutral master,
119 the other is a hardware-aware slave), the endpoint I/0 API used here
120 should also be usable for an overhead-reduced host side API.
125 <chapter id="structure"><title>Structure of Gadget Drivers</title>
127 <para>A system running inside a USB peripheral
128 normally has at least three layers inside the kernel to handle
129 USB protocol processing, and may have additional layers in
131 The "gadget" API is used by the middle layer to interact
132 with the lowest level (which directly handles hardware).
135 <para>In Linux, from the bottom up, these layers are:
141 <term><emphasis>USB Controller Driver</emphasis></term>
144 <para>This is the lowest software level.
145 It is the only layer that talks to hardware,
146 through registers, fifos, dma, irqs, and the like.
147 The <filename><linux/usb/gadget.h></filename> API abstracts
148 the peripheral controller endpoint hardware.
149 That hardware is exposed through endpoint objects, which accept
150 streams of IN/OUT buffers, and through callbacks that interact
152 Since normal USB devices only have one upstream
153 port, they only have one of these drivers.
154 The controller driver can support any number of different
155 gadget drivers, but only one of them can be used at a time.
158 <para>Examples of such controller hardware include
159 the PCI-based NetChip 2280 USB 2.0 high speed controller,
160 the SA-11x0 or PXA-25x UDC (found within many PDAs),
161 and a variety of other products.
164 </listitem></varlistentry>
167 <term><emphasis>Gadget Driver</emphasis></term>
170 <para>The lower boundary of this driver implements hardware-neutral
171 USB functions, using calls to the controller driver.
172 Because such hardware varies widely in capabilities and restrictions,
173 and is used in embedded environments where space is at a premium,
174 the gadget driver is often configured at compile time
175 to work with endpoints supported by one particular controller.
176 Gadget drivers may be portable to several different controllers,
177 using conditional compilation.
178 (Recent kernels substantially simplify the work involved in
179 supporting new hardware, by <emphasis>autoconfiguring</emphasis>
180 endpoints automatically for many bulk-oriented drivers.)
181 Gadget driver responsibilities include:
184 <listitem><para>handling setup requests (ep0 protocol responses)
185 possibly including class-specific functionality
187 <listitem><para>returning configuration and string descriptors
189 <listitem><para>(re)setting configurations and interface
190 altsettings, including enabling and configuring endpoints
192 <listitem><para>handling life cycle events, such as managing
193 bindings to hardware,
194 USB suspend/resume, remote wakeup,
195 and disconnection from the USB host.
197 <listitem><para>managing IN and OUT transfers on all currently
203 Such drivers may be modules of proprietary code, although
204 that approach is discouraged in the Linux community.
206 </listitem></varlistentry>
209 <term><emphasis>Upper Level</emphasis></term>
212 <para>Most gadget drivers have an upper boundary that connects
213 to some Linux driver or framework in Linux.
214 Through that boundary flows the data which the gadget driver
215 produces and/or consumes through protocol transfers over USB.
219 <listitem><para>user mode code, using generic (gadgetfs)
220 or application specific files in
221 <filename>/dev</filename>
223 <listitem><para>networking subsystem (for network gadgets,
224 like the CDC Ethernet Model gadget driver)
226 <listitem><para>data capture drivers, perhaps video4Linux or
227 a scanner driver; or test and measurement hardware.
229 <listitem><para>input subsystem (for HID gadgets)
231 <listitem><para>sound subsystem (for audio gadgets)
233 <listitem><para>file system (for PTP gadgets)
235 <listitem><para>block i/o subsystem (for usb-storage gadgets)
237 <listitem><para>... and more </para></listitem>
239 </listitem></varlistentry>
242 <term><emphasis>Additional Layers</emphasis></term>
245 <para>Other layers may exist.
246 These could include kernel layers, such as network protocol stacks,
247 as well as user mode applications building on standard POSIX
248 system call APIs such as
249 <emphasis>open()</emphasis>, <emphasis>close()</emphasis>,
250 <emphasis>read()</emphasis> and <emphasis>write()</emphasis>.
251 On newer systems, POSIX Async I/O calls may be an option.
252 Such user mode code will not necessarily be subject to
253 the GNU General Public License (GPL).
255 </listitem></varlistentry>
260 <para>OTG-capable systems will also need to include a standard Linux-USB
262 with <emphasis>usbcore</emphasis>,
263 one or more <emphasis>Host Controller Drivers</emphasis> (HCDs),
264 <emphasis>USB Device Drivers</emphasis> to support
265 the OTG "Targeted Peripheral List",
267 There will also be an <emphasis>OTG Controller Driver</emphasis>,
268 which is visible to gadget and device driver developers only indirectly.
269 That helps the host and device side USB controllers implement the
270 two new OTG protocols (HNP and SRP).
271 Roles switch (host to peripheral, or vice versa) using HNP
272 during USB suspend processing, and SRP can be viewed as a
273 more battery-friendly kind of device wakeup protocol.
276 <para>Over time, reusable utilities are evolving to help make some
277 gadget driver tasks simpler.
278 For example, building configuration descriptors from vectors of
279 descriptors for the configurations interfaces and endpoints is
280 now automated, and many drivers now use autoconfiguration to
281 choose hardware endpoints and initialize their descriptors.
283 A potential example of particular interest
284 is code implementing standard USB-IF protocols for
285 HID, networking, storage, or audio classes.
286 Some developers are interested in KDB or KGDB hooks, to let
287 target hardware be remotely debugged.
288 Most such USB protocol code doesn't need to be hardware-specific,
289 any more than network protocols like X11, HTTP, or NFS are.
290 Such gadget-side interface drivers should eventually be combined,
291 to implement composite devices.
297 <chapter id="api"><title>Kernel Mode Gadget API</title>
299 <para>Gadget drivers declare themselves through a
300 <emphasis>struct usb_gadget_driver</emphasis>, which is responsible for
301 most parts of enumeration for a <emphasis>struct usb_gadget</emphasis>.
302 The response to a set_configuration usually involves
303 enabling one or more of the <emphasis>struct usb_ep</emphasis> objects
304 exposed by the gadget, and submitting one or more
305 <emphasis>struct usb_request</emphasis> buffers to transfer data.
306 Understand those four data types, and their operations, and
307 you will understand how this API works.
310 <note><title>Incomplete Data Type Descriptions</title>
312 <para>This documentation was prepared using the standard Linux
313 kernel <filename>docproc</filename> tool, which turns text
314 and in-code comments into SGML DocBook and then into usable
315 formats such as HTML or PDF.
316 Other than the "Chapter 9" data types, most of the significant
317 data types and functions are described here.
320 <para>However, docproc does not understand all the C constructs
321 that are used, so some relevant information is likely omitted from
322 what you are reading.
323 One example of such information is endpoint autoconfiguration.
324 You'll have to read the header file, and use example source
325 code (such as that for "Gadget Zero"), to fully understand the API.
328 <para>The part of the API implementing some basic
329 driver capabilities is specific to the version of the
330 Linux kernel that's in use.
331 The 2.6 kernel includes a <emphasis>driver model</emphasis>
332 framework that has no analogue on earlier kernels;
333 so those parts of the gadget API are not fully portable.
334 (They are implemented on 2.4 kernels, but in a different way.)
335 The driver model state is another part of this API that is
336 ignored by the kerneldoc tools.
340 <para>The core API does not expose
341 every possible hardware feature, only the most widely available ones.
342 There are significant hardware features, such as device-to-device DMA
343 (without temporary storage in a memory buffer)
344 that would be added using hardware-specific APIs.
347 <para>This API allows drivers to use conditional compilation to handle
348 endpoint capabilities of different hardware, but doesn't require that.
349 Hardware tends to have arbitrary restrictions, relating to
350 transfer types, addressing, packet sizes, buffering, and availability.
351 As a rule, such differences only matter for "endpoint zero" logic
352 that handles device configuration and management.
353 The API supports limited run-time
354 detection of capabilities, through naming conventions for endpoints.
355 Many drivers will be able to at least partially autoconfigure
357 In particular, driver init sections will often have endpoint
358 autoconfiguration logic that scans the hardware's list of endpoints
359 to find ones matching the driver requirements
360 (relying on those conventions), to eliminate some of the most
361 common reasons for conditional compilation.
364 <para>Like the Linux-USB host side API, this API exposes
365 the "chunky" nature of USB messages: I/O requests are in terms
366 of one or more "packets", and packet boundaries are visible to drivers.
367 Compared to RS-232 serial protocols, USB resembles
368 synchronous protocols like HDLC
369 (N bytes per frame, multipoint addressing, host as the primary
370 station and devices as secondary stations)
371 more than asynchronous ones
372 (tty style: 8 data bits per frame, no parity, one stop bit).
373 So for example the controller drivers won't buffer
374 two single byte writes into a single two-byte USB IN packet,
375 although gadget drivers may do so when they implement
376 protocols where packet boundaries (and "short packets")
380 <sect1 id="lifecycle"><title>Driver Life Cycle</title>
382 <para>Gadget drivers make endpoint I/O requests to hardware without
383 needing to know many details of the hardware, but driver
384 setup/configuration code needs to handle some differences.
385 Use the API like this:
388 <orderedlist numeration='arabic'>
390 <listitem><para>Register a driver for the particular device side
391 usb controller hardware,
392 such as the net2280 on PCI (USB 2.0),
393 sa11x0 or pxa25x as found in Linux PDAs,
395 At this point the device is logically in the USB ch9 initial state
396 ("attached"), drawing no power and not usable
397 (since it does not yet support enumeration).
398 Any host should not see the device, since it's not
399 activated the data line pullup used by the host to
400 detect a device, even if VBUS power is available.
403 <listitem><para>Register a gadget driver that implements some higher level
404 device function. That will then bind() to a usb_gadget, which
405 activates the data line pullup sometime after detecting VBUS.
408 <listitem><para>The hardware driver can now start enumerating.
409 The steps it handles are to accept USB power and set_address requests.
410 Other steps are handled by the gadget driver.
411 If the gadget driver module is unloaded before the host starts to
412 enumerate, steps before step 7 are skipped.
415 <listitem><para>The gadget driver's setup() call returns usb descriptors,
416 based both on what the bus interface hardware provides and on the
417 functionality being implemented.
418 That can involve alternate settings or configurations,
419 unless the hardware prevents such operation.
420 For OTG devices, each configuration descriptor includes
424 <listitem><para>The gadget driver handles the last step of enumeration,
425 when the USB host issues a set_configuration call.
426 It enables all endpoints used in that configuration,
427 with all interfaces in their default settings.
428 That involves using a list of the hardware's endpoints, enabling each
429 endpoint according to its descriptor.
430 It may also involve using <function>usb_gadget_vbus_draw</function>
431 to let more power be drawn from VBUS, as allowed by that configuration.
432 For OTG devices, setting a configuration may also involve reporting
433 HNP capabilities through a user interface.
436 <listitem><para>Do real work and perform data transfers, possibly involving
437 changes to interface settings or switching to new configurations, until the
438 device is disconnect()ed from the host.
439 Queue any number of transfer requests to each endpoint.
440 It may be suspended and resumed several times before being disconnected.
441 On disconnect, the drivers go back to step 3 (above).
444 <listitem><para>When the gadget driver module is being unloaded,
445 the driver unbind() callback is issued. That lets the controller
451 <para>Drivers will normally be arranged so that just loading the
452 gadget driver module (or statically linking it into a Linux kernel)
453 allows the peripheral device to be enumerated, but some drivers
454 will defer enumeration until some higher level component (like
455 a user mode daemon) enables it.
456 Note that at this lowest level there are no policies about how
457 ep0 configuration logic is implemented,
458 except that it should obey USB specifications.
459 Such issues are in the domain of gadget drivers,
460 including knowing about implementation constraints
461 imposed by some USB controllers
462 or understanding that composite devices might happen to
463 be built by integrating reusable components.
466 <para>Note that the lifecycle above can be slightly different
468 Other than providing an additional OTG descriptor in each
469 configuration, only the HNP-related differences are particularly
470 visible to driver code.
471 They involve reporting requirements during the SET_CONFIGURATION
472 request, and the option to invoke HNP during some suspend callbacks.
473 Also, SRP changes the semantics of
474 <function>usb_gadget_wakeup</function>
480 <sect1 id="ch9"><title>USB 2.0 Chapter 9 Types and Constants</title>
483 rely on common USB structures and constants
485 <filename><linux/usb/ch9.h></filename>
486 header file, which is standard in Linux 2.6 kernels.
487 These are the same types and constants used by host
488 side drivers (and usbcore).
491 !Iinclude/linux/usb/ch9.h
494 <sect1 id="core"><title>Core Objects and Methods</title>
496 <para>These are declared in
497 <filename><linux/usb/gadget.h></filename>,
498 and are used by gadget drivers to interact with
499 USB peripheral controller drivers.
502 <!-- yeech, this is ugly in nsgmls PDF output.
504 the PDF bookmark and refentry output nesting is wrong,
505 and the member/argument documentation indents ugly.
507 plus something (docproc?) adds whitespace before the
508 descriptive paragraph text, so it can't line up right
509 unless the explanations are trivial.
512 !Iinclude/linux/usb/gadget.h
515 <sect1 id="utils"><title>Optional Utilities</title>
517 <para>The core API is sufficient for writing a USB Gadget Driver,
518 but some optional utilities are provided to simplify common tasks.
519 These utilities include endpoint autoconfiguration.
522 !Edrivers/usb/gadget/usbstring.c
523 !Edrivers/usb/gadget/config.c
524 <!-- !Edrivers/usb/gadget/epautoconf.c -->
527 <sect1 id="composite"><title>Composite Device Framework</title>
529 <para>The core API is sufficient for writing drivers for composite
530 USB devices (with more than one function in a given configuration),
531 and also multi-configuration devices (also more than one function,
532 but not necessarily sharing a given configuration).
533 There is however an optional framework which makes it easier to
534 reuse and combine functions.
537 <para>Devices using this framework provide a <emphasis>struct
538 usb_composite_driver</emphasis>, which in turn provides one or
539 more <emphasis>struct usb_configuration</emphasis> instances.
540 Each such configuration includes at least one
541 <emphasis>struct usb_function</emphasis>, which packages a user
542 visible role such as "network link" or "mass storage device".
543 Management functions may also exist, such as "Device Firmware
547 !Iinclude/linux/usb/composite.h
548 !Edrivers/usb/gadget/composite.c
552 <sect1 id="functions"><title>Composite Device Functions</title>
554 <para>At this writing, a few of the current gadget drivers have
555 been converted to this framework.
556 Near-term plans include converting all of them, except for "gadgetfs".
564 <chapter id="controllers"><title>Peripheral Controller Drivers</title>
566 <para>The first hardware supporting this API was the NetChip 2280
567 controller, which supports USB 2.0 high speed and is based on PCI.
568 This is the <filename>net2280</filename> driver module.
569 The driver supports Linux kernel versions 2.4 and 2.6;
570 contact NetChip Technologies for development boards and product
574 <para>Other hardware working in the "gadget" framework includes:
575 Intel's PXA 25x and IXP42x series processors
576 (<filename>pxa2xx_udc</filename>),
577 Toshiba TC86c001 "Goku-S" (<filename>goku_udc</filename>),
578 Renesas SH7705/7727 (<filename>sh_udc</filename>),
579 MediaQ 11xx (<filename>mq11xx_udc</filename>),
580 Hynix HMS30C7202 (<filename>h7202_udc</filename>),
581 National 9303/4 (<filename>n9604_udc</filename>),
582 Texas Instruments OMAP (<filename>omap_udc</filename>),
583 Sharp LH7A40x (<filename>lh7a40x_udc</filename>),
585 Most of those are full speed controllers.
588 <para>At this writing, there are people at work on drivers in
589 this framework for several other USB device controllers,
590 with plans to make many of them be widely available.
593 <!-- !Edrivers/usb/gadget/net2280.c -->
595 <para>A partial USB simulator,
596 the <filename>dummy_hcd</filename> driver, is available.
597 It can act like a net2280, a pxa25x, or an sa11x0 in terms
598 of available endpoints and device speeds; and it simulates
599 control, bulk, and to some extent interrupt transfers.
600 That lets you develop some parts of a gadget driver on a normal PC,
601 without any special hardware, and perhaps with the assistance
602 of tools such as GDB running with User Mode Linux.
603 At least one person has expressed interest in adapting that
604 approach, hooking it up to a simulator for a microcontroller.
605 Such simulators can help debug subsystems where the runtime hardware
606 is unfriendly to software development, or is not yet available.
609 <para>Support for other controllers is expected to be developed
611 over time, as this driver framework evolves.
616 <chapter id="gadget"><title>Gadget Drivers</title>
618 <para>In addition to <emphasis>Gadget Zero</emphasis>
619 (used primarily for testing and development with drivers
620 for usb controller hardware), other gadget drivers exist.
623 <para>There's an <emphasis>ethernet</emphasis> gadget
624 driver, which implements one of the most useful
625 <emphasis>Communications Device Class</emphasis> (CDC) models.
626 One of the standards for cable modem interoperability even
627 specifies the use of this ethernet model as one of two
629 Gadgets using this code look to a USB host as if they're
631 It provides access to a network where the gadget's CPU is one host,
632 which could easily be bridging, routing, or firewalling
633 access to other networks.
634 Since some hardware can't fully implement the CDC Ethernet
635 requirements, this driver also implements a "good parts only"
636 subset of CDC Ethernet.
637 (That subset doesn't advertise itself as CDC Ethernet,
638 to avoid creating problems.)
641 <para>Support for Microsoft's <emphasis>RNDIS</emphasis>
642 protocol has been contributed by Pengutronix and Auerswald GmbH.
643 This is like CDC Ethernet, but it runs on more slightly USB hardware
644 (but less than the CDC subset).
645 However, its main claim to fame is being able to connect directly to
646 recent versions of Windows, using drivers that Microsoft bundles
647 and supports, making it much simpler to network with Windows.
650 <para>There is also support for user mode gadget drivers,
651 using <emphasis>gadgetfs</emphasis>.
652 This provides a <emphasis>User Mode API</emphasis> that presents
653 each endpoint as a single file descriptor. I/O is done using
654 normal <emphasis>read()</emphasis> and <emphasis>read()</emphasis> calls.
655 Familiar tools like GDB and pthreads can be used to
656 develop and debug user mode drivers, so that once a robust
657 controller driver is available many applications for it
658 won't require new kernel mode software.
659 Linux 2.6 <emphasis>Async I/O (AIO)</emphasis>
660 support is available, so that user mode software
661 can stream data with only slightly more overhead
662 than a kernel driver.
665 <para>There's a USB Mass Storage class driver, which provides
666 a different solution for interoperability with systems such
667 as MS-Windows and MacOS.
668 That <emphasis>File-backed Storage</emphasis> driver uses a
669 file or block device as backing store for a drive,
670 like the <filename>loop</filename> driver.
671 The USB host uses the BBB, CB, or CBI versions of the mass
672 storage class specification, using transparent SCSI commands
673 to access the data from the backing store.
676 <para>There's a "serial line" driver, useful for TTY style
678 The latest version of that driver supports CDC ACM style
679 operation, like a USB modem, and so on most hardware it can
680 interoperate easily with MS-Windows.
681 One interesting use of that driver is in boot firmware (like a BIOS),
682 which can sometimes use that model with very small systems without
686 <para>Support for other kinds of gadget is expected to
687 be developed and contributed
688 over time, as this driver framework evolves.
693 <chapter id="otg"><title>USB On-The-GO (OTG)</title>
695 <para>USB OTG support on Linux 2.6 was initially developed
696 by Texas Instruments for
697 <ulink url="http://www.omap.com">OMAP</ulink> 16xx and 17xx
699 Other OTG systems should work in similar ways, but the
700 hardware level details could be very different.
703 <para>Systems need specialized hardware support to implement OTG,
704 notably including a special <emphasis>Mini-AB</emphasis> jack
705 and associated transciever to support <emphasis>Dual-Role</emphasis>
707 they can act either as a host, using the standard
708 Linux-USB host side driver stack,
709 or as a peripheral, using this "gadget" framework.
710 To do that, the system software relies on small additions
711 to those programming interfaces,
712 and on a new internal component (here called an "OTG Controller")
713 affecting which driver stack connects to the OTG port.
714 In each role, the system can re-use the existing pool of
715 hardware-neutral drivers, layered on top of the controller
716 driver interfaces (<emphasis>usb_bus</emphasis> or
717 <emphasis>usb_gadget</emphasis>).
718 Such drivers need at most minor changes, and most of the calls
719 added to support OTG can also benefit non-OTG products.
723 <listitem><para>Gadget drivers test the <emphasis>is_otg</emphasis>
724 flag, and use it to determine whether or not to include
725 an OTG descriptor in each of their configurations.
727 <listitem><para>Gadget drivers may need changes to support the
728 two new OTG protocols, exposed in new gadget attributes
729 such as <emphasis>b_hnp_enable</emphasis> flag.
730 HNP support should be reported through a user interface
731 (two LEDs could suffice), and is triggered in some cases
732 when the host suspends the peripheral.
733 SRP support can be user-initiated just like remote wakeup,
734 probably by pressing the same button.
736 <listitem><para>On the host side, USB device drivers need
737 to be taught to trigger HNP at appropriate moments, using
738 <function>usb_suspend_device()</function>.
739 That also conserves battery power, which is useful even
740 for non-OTG configurations.
742 <listitem><para>Also on the host side, a driver must support the
743 OTG "Targeted Peripheral List". That's just a whitelist,
744 used to reject peripherals not supported with a given
746 <emphasis>This whitelist is product-specific;
747 each product must modify <filename>otg_whitelist.h</filename>
748 to match its interoperability specification.
751 <para>Non-OTG Linux hosts, like PCs and workstations,
752 normally have some solution for adding drivers, so that
753 peripherals that aren't recognized can eventually be supported.
754 That approach is unreasonable for consumer products that may
755 never have their firmware upgraded, and where it's usually
756 unrealistic to expect traditional PC/workstation/server kinds
757 of support model to work.
758 For example, it's often impractical to change device firmware
759 once the product has been distributed, so driver bugs can't
760 normally be fixed if they're found after shipment.
765 Additional changes are needed below those hardware-neutral
766 <emphasis>usb_bus</emphasis> and <emphasis>usb_gadget</emphasis>
767 driver interfaces; those aren't discussed here in any detail.
768 Those affect the hardware-specific code for each USB Host or Peripheral
769 controller, and how the HCD initializes (since OTG can be active only
771 They also involve what may be called an <emphasis>OTG Controller
772 Driver</emphasis>, managing the OTG transceiver and the OTG state
773 machine logic as well as much of the root hub behavior for the
775 The OTG controller driver needs to activate and deactivate USB
776 controllers depending on the relevant device role.
777 Some related changes were needed inside usbcore, so that it
778 can identify OTG-capable devices and respond appropriately
779 to HNP or SRP protocols.