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="drmDevelopersGuide">
7 <title>Linux DRM Developer's Guide</title>
11 <firstname>Jesse</firstname>
12 <surname>Barnes</surname>
13 <contrib>Initial version</contrib>
15 <orgname>Intel Corporation</orgname>
17 <email>jesse.barnes@intel.com</email>
22 <firstname>Laurent</firstname>
23 <surname>Pinchart</surname>
24 <contrib>Driver internals</contrib>
26 <orgname>Ideas on board SPRL</orgname>
28 <email>laurent.pinchart@ideasonboard.com</email>
35 <year>2008-2009</year>
37 <holder>Intel Corporation</holder>
38 <holder>Laurent Pinchart</holder>
43 The contents of this file may be used under the terms of the GNU
44 General Public License version 2 (the "GPL") as distributed in
45 the kernel source COPYING file.
50 <!-- Put document revisions here, newest first. -->
52 <revnumber>1.0</revnumber>
53 <date>2012-07-13</date>
54 <authorinitials>LP</authorinitials>
55 <revremark>Added extensive documentation about driver internals.
64 <title>DRM Core</title>
67 This first part of the DRM Developer's Guide documents core DRM code,
68 helper libraries for writting drivers and generic userspace interfaces
69 exposed by DRM drivers.
73 <chapter id="drmIntroduction">
74 <title>Introduction</title>
76 The Linux DRM layer contains code intended to support the needs
77 of complex graphics devices, usually containing programmable
78 pipelines well suited to 3D graphics acceleration. Graphics
79 drivers in the kernel may make use of DRM functions to make
80 tasks like memory management, interrupt handling and DMA easier,
81 and provide a uniform interface to applications.
84 A note on versions: this guide covers features found in the DRM
85 tree, including the TTM memory manager, output configuration and
86 mode setting, and the new vblank internals, in addition to all
87 the regular features found in current kernels.
90 [Insert diagram of typical DRM stack here]
96 <chapter id="drmInternals">
97 <title>DRM Internals</title>
99 This chapter documents DRM internals relevant to driver authors
100 and developers working to add support for the latest features to
104 First, we go over some typical driver initialization
105 requirements, like setting up command buffers, creating an
106 initial output configuration, and initializing core services.
107 Subsequent sections cover core internals in more detail,
108 providing implementation notes and examples.
111 The DRM layer provides several services to graphics drivers,
112 many of them driven by the application interfaces it provides
113 through libdrm, the library that wraps most of the DRM ioctls.
114 These include vblank event handling, memory
115 management, output management, framebuffer management, command
116 submission & fencing, suspend/resume support, and DMA
120 <!-- Internals: driver init -->
123 <title>Driver Initialization</title>
125 At the core of every DRM driver is a <structname>drm_driver</structname>
126 structure. Drivers typically statically initialize a drm_driver structure,
127 and then pass it to one of the <function>drm_*_init()</function> functions
128 to register it with the DRM subsystem.
131 The <structname>drm_driver</structname> structure contains static
132 information that describes the driver and features it supports, and
133 pointers to methods that the DRM core will call to implement the DRM API.
134 We will first go through the <structname>drm_driver</structname> static
135 information fields, and will then describe individual operations in
136 details as they get used in later sections.
139 <title>Driver Information</title>
141 <title>Driver Features</title>
143 Drivers inform the DRM core about their requirements and supported
144 features by setting appropriate flags in the
145 <structfield>driver_features</structfield> field. Since those flags
146 influence the DRM core behaviour since registration time, most of them
147 must be set to registering the <structname>drm_driver</structname>
150 <synopsis>u32 driver_features;</synopsis>
152 <title>Driver Feature Flags</title>
154 <term>DRIVER_USE_AGP</term>
156 Driver uses AGP interface, the DRM core will manage AGP resources.
160 <term>DRIVER_REQUIRE_AGP</term>
162 Driver needs AGP interface to function. AGP initialization failure
163 will become a fatal error.
167 <term>DRIVER_PCI_DMA</term>
169 Driver is capable of PCI DMA, mapping of PCI DMA buffers to
170 userspace will be enabled. Deprecated.
174 <term>DRIVER_SG</term>
176 Driver can perform scatter/gather DMA, allocation and mapping of
177 scatter/gather buffers will be enabled. Deprecated.
181 <term>DRIVER_HAVE_DMA</term>
183 Driver supports DMA, the userspace DMA API will be supported.
188 <term>DRIVER_HAVE_IRQ</term><term>DRIVER_IRQ_SHARED</term>
190 DRIVER_HAVE_IRQ indicates whether the driver has an IRQ handler
191 managed by the DRM Core. The core will support simple IRQ handler
192 installation when the flag is set. The installation process is
193 described in <xref linkend="drm-irq-registration"/>.</para>
194 <para>DRIVER_IRQ_SHARED indicates whether the device & handler
195 support shared IRQs (note that this is required of PCI drivers).
199 <term>DRIVER_GEM</term>
201 Driver use the GEM memory manager.
205 <term>DRIVER_MODESET</term>
207 Driver supports mode setting interfaces (KMS).
211 <term>DRIVER_PRIME</term>
213 Driver implements DRM PRIME buffer sharing.
217 <term>DRIVER_RENDER</term>
219 Driver supports dedicated render nodes.
225 <title>Major, Minor and Patchlevel</title>
228 int patchlevel;</synopsis>
230 The DRM core identifies driver versions by a major, minor and patch
231 level triplet. The information is printed to the kernel log at
232 initialization time and passed to userspace through the
233 DRM_IOCTL_VERSION ioctl.
236 The major and minor numbers are also used to verify the requested driver
237 API version passed to DRM_IOCTL_SET_VERSION. When the driver API changes
238 between minor versions, applications can call DRM_IOCTL_SET_VERSION to
239 select a specific version of the API. If the requested major isn't equal
240 to the driver major, or the requested minor is larger than the driver
241 minor, the DRM_IOCTL_SET_VERSION call will return an error. Otherwise
242 the driver's set_version() method will be called with the requested
247 <title>Name, Description and Date</title>
248 <synopsis>char *name;
250 char *date;</synopsis>
252 The driver name is printed to the kernel log at initialization time,
253 used for IRQ registration and passed to userspace through
257 The driver description is a purely informative string passed to
258 userspace through the DRM_IOCTL_VERSION ioctl and otherwise unused by
262 The driver date, formatted as YYYYMMDD, is meant to identify the date of
263 the latest modification to the driver. However, as most drivers fail to
264 update it, its value is mostly useless. The DRM core prints it to the
265 kernel log at initialization time and passes it to userspace through the
266 DRM_IOCTL_VERSION ioctl.
271 <title>Driver Load</title>
273 The <methodname>load</methodname> method is the driver and device
274 initialization entry point. The method is responsible for allocating and
275 initializing driver private data, performing resource allocation and
276 mapping (e.g. acquiring
277 clocks, mapping registers or allocating command buffers), initializing
278 the memory manager (<xref linkend="drm-memory-management"/>), installing
279 the IRQ handler (<xref linkend="drm-irq-registration"/>), setting up
280 vertical blanking handling (<xref linkend="drm-vertical-blank"/>), mode
281 setting (<xref linkend="drm-mode-setting"/>) and initial output
282 configuration (<xref linkend="drm-kms-init"/>).
285 If compatibility is a concern (e.g. with drivers converted over from
286 User Mode Setting to Kernel Mode Setting), care must be taken to prevent
287 device initialization and control that is incompatible with currently
288 active userspace drivers. For instance, if user level mode setting
289 drivers are in use, it would be problematic to perform output discovery
290 & configuration at load time. Likewise, if user-level drivers
291 unaware of memory management are in use, memory management and command
292 buffer setup may need to be omitted. These requirements are
293 driver-specific, and care needs to be taken to keep both old and new
294 applications and libraries working.
296 <synopsis>int (*load) (struct drm_device *, unsigned long flags);</synopsis>
298 The method takes two arguments, a pointer to the newly created
299 <structname>drm_device</structname> and flags. The flags are used to
300 pass the <structfield>driver_data</structfield> field of the device id
301 corresponding to the device passed to <function>drm_*_init()</function>.
302 Only PCI devices currently use this, USB and platform DRM drivers have
303 their <methodname>load</methodname> method called with flags to 0.
306 <title>Driver Private Data</title>
308 The driver private hangs off the main
309 <structname>drm_device</structname> structure and can be used for
310 tracking various device-specific bits of information, like register
311 offsets, command buffer status, register state for suspend/resume, etc.
312 At load time, a driver may simply allocate one and set
313 <structname>drm_device</structname>.<structfield>dev_priv</structfield>
314 appropriately; it should be freed and
315 <structname>drm_device</structname>.<structfield>dev_priv</structfield>
316 set to NULL when the driver is unloaded.
319 <sect3 id="drm-irq-registration">
320 <title>IRQ Registration</title>
322 The DRM core tries to facilitate IRQ handler registration and
323 unregistration by providing <function>drm_irq_install</function> and
324 <function>drm_irq_uninstall</function> functions. Those functions only
325 support a single interrupt per device, devices that use more than one
326 IRQs need to be handled manually.
329 <title>Managed IRQ Registration</title>
331 Both the <function>drm_irq_install</function> and
332 <function>drm_irq_uninstall</function> functions get the device IRQ by
333 calling <function>drm_dev_to_irq</function>. This inline function will
334 call a bus-specific operation to retrieve the IRQ number. For platform
335 devices, <function>platform_get_irq</function>(..., 0) is used to
336 retrieve the IRQ number.
339 <function>drm_irq_install</function> starts by calling the
340 <methodname>irq_preinstall</methodname> driver operation. The operation
341 is optional and must make sure that the interrupt will not get fired by
342 clearing all pending interrupt flags or disabling the interrupt.
345 The IRQ will then be requested by a call to
346 <function>request_irq</function>. If the DRIVER_IRQ_SHARED driver
347 feature flag is set, a shared (IRQF_SHARED) IRQ handler will be
351 The IRQ handler function must be provided as the mandatory irq_handler
352 driver operation. It will get passed directly to
353 <function>request_irq</function> and thus has the same prototype as all
354 IRQ handlers. It will get called with a pointer to the DRM device as the
358 Finally the function calls the optional
359 <methodname>irq_postinstall</methodname> driver operation. The operation
360 usually enables interrupts (excluding the vblank interrupt, which is
361 enabled separately), but drivers may choose to enable/disable interrupts
365 <function>drm_irq_uninstall</function> is similarly used to uninstall an
366 IRQ handler. It starts by waking up all processes waiting on a vblank
367 interrupt to make sure they don't hang, and then calls the optional
368 <methodname>irq_uninstall</methodname> driver operation. The operation
369 must disable all hardware interrupts. Finally the function frees the IRQ
370 by calling <function>free_irq</function>.
374 <title>Manual IRQ Registration</title>
376 Drivers that require multiple interrupt handlers can't use the managed
377 IRQ registration functions. In that case IRQs must be registered and
378 unregistered manually (usually with the <function>request_irq</function>
379 and <function>free_irq</function> functions, or their devm_* equivalent).
382 When manually registering IRQs, drivers must not set the DRIVER_HAVE_IRQ
383 driver feature flag, and must not provide the
384 <methodname>irq_handler</methodname> driver operation. They must set the
385 <structname>drm_device</structname> <structfield>irq_enabled</structfield>
386 field to 1 upon registration of the IRQs, and clear it to 0 after
387 unregistering the IRQs.
392 <title>Memory Manager Initialization</title>
394 Every DRM driver requires a memory manager which must be initialized at
395 load time. DRM currently contains two memory managers, the Translation
396 Table Manager (TTM) and the Graphics Execution Manager (GEM).
397 This document describes the use of the GEM memory manager only. See
398 <xref linkend="drm-memory-management"/> for details.
402 <title>Miscellaneous Device Configuration</title>
404 Another task that may be necessary for PCI devices during configuration
405 is mapping the video BIOS. On many devices, the VBIOS describes device
406 configuration, LCD panel timings (if any), and contains flags indicating
407 device state. Mapping the BIOS can be done using the pci_map_rom() call,
408 a convenience function that takes care of mapping the actual ROM,
409 whether it has been shadowed into memory (typically at address 0xc0000)
410 or exists on the PCI device in the ROM BAR. Note that after the ROM has
411 been mapped and any necessary information has been extracted, it should
412 be unmapped; on many devices, the ROM address decoder is shared with
413 other BARs, so leaving it mapped could cause undesired behaviour like
414 hangs or memory corruption.
415 <!--!Fdrivers/pci/rom.c pci_map_rom-->
421 <!-- Internals: memory management -->
423 <sect1 id="drm-memory-management">
424 <title>Memory management</title>
426 Modern Linux systems require large amount of graphics memory to store
427 frame buffers, textures, vertices and other graphics-related data. Given
428 the very dynamic nature of many of that data, managing graphics memory
429 efficiently is thus crucial for the graphics stack and plays a central
430 role in the DRM infrastructure.
433 The DRM core includes two memory managers, namely Translation Table Maps
434 (TTM) and Graphics Execution Manager (GEM). TTM was the first DRM memory
435 manager to be developed and tried to be a one-size-fits-them all
436 solution. It provides a single userspace API to accommodate the need of
437 all hardware, supporting both Unified Memory Architecture (UMA) devices
438 and devices with dedicated video RAM (i.e. most discrete video cards).
439 This resulted in a large, complex piece of code that turned out to be
440 hard to use for driver development.
443 GEM started as an Intel-sponsored project in reaction to TTM's
444 complexity. Its design philosophy is completely different: instead of
445 providing a solution to every graphics memory-related problems, GEM
446 identified common code between drivers and created a support library to
447 share it. GEM has simpler initialization and execution requirements than
448 TTM, but has no video RAM management capabitilies and is thus limited to
452 <title>The Translation Table Manager (TTM)</title>
454 TTM design background and information belongs here.
457 <title>TTM initialization</title>
458 <warning><para>This section is outdated.</para></warning>
460 Drivers wishing to support TTM must fill out a drm_bo_driver
461 structure. The structure contains several fields with function
462 pointers for initializing the TTM, allocating and freeing memory,
463 waiting for command completion and fence synchronization, and memory
464 migration. See the radeon_ttm.c file for an example of usage.
467 The ttm_global_reference structure is made up of several fields:
470 struct ttm_global_reference {
471 enum ttm_global_types global_type;
474 int (*init) (struct ttm_global_reference *);
475 void (*release) (struct ttm_global_reference *);
479 There should be one global reference structure for your memory
480 manager as a whole, and there will be others for each object
481 created by the memory manager at runtime. Your global TTM should
482 have a type of TTM_GLOBAL_TTM_MEM. The size field for the global
483 object should be sizeof(struct ttm_mem_global), and the init and
484 release hooks should point at your driver-specific init and
485 release routines, which probably eventually call
486 ttm_mem_global_init and ttm_mem_global_release, respectively.
489 Once your global TTM accounting structure is set up and initialized
490 by calling ttm_global_item_ref() on it,
491 you need to create a buffer object TTM to
492 provide a pool for buffer object allocation by clients and the
493 kernel itself. The type of this object should be TTM_GLOBAL_TTM_BO,
494 and its size should be sizeof(struct ttm_bo_global). Again,
495 driver-specific init and release functions may be provided,
496 likely eventually calling ttm_bo_global_init() and
497 ttm_bo_global_release(), respectively. Also, like the previous
498 object, ttm_global_item_ref() is used to create an initial reference
499 count for the TTM, which will call your initialization function.
504 <title>The Graphics Execution Manager (GEM)</title>
506 The GEM design approach has resulted in a memory manager that doesn't
507 provide full coverage of all (or even all common) use cases in its
508 userspace or kernel API. GEM exposes a set of standard memory-related
509 operations to userspace and a set of helper functions to drivers, and let
510 drivers implement hardware-specific operations with their own private API.
513 The GEM userspace API is described in the
514 <ulink url="http://lwn.net/Articles/283798/"><citetitle>GEM - the Graphics
515 Execution Manager</citetitle></ulink> article on LWN. While slightly
516 outdated, the document provides a good overview of the GEM API principles.
517 Buffer allocation and read and write operations, described as part of the
518 common GEM API, are currently implemented using driver-specific ioctls.
521 GEM is data-agnostic. It manages abstract buffer objects without knowing
522 what individual buffers contain. APIs that require knowledge of buffer
523 contents or purpose, such as buffer allocation or synchronization
524 primitives, are thus outside of the scope of GEM and must be implemented
525 using driver-specific ioctls.
528 On a fundamental level, GEM involves several operations:
530 <listitem>Memory allocation and freeing</listitem>
531 <listitem>Command execution</listitem>
532 <listitem>Aperture management at command execution time</listitem>
534 Buffer object allocation is relatively straightforward and largely
535 provided by Linux's shmem layer, which provides memory to back each
539 Device-specific operations, such as command execution, pinning, buffer
540 read & write, mapping, and domain ownership transfers are left to
541 driver-specific ioctls.
544 <title>GEM Initialization</title>
546 Drivers that use GEM must set the DRIVER_GEM bit in the struct
547 <structname>drm_driver</structname>
548 <structfield>driver_features</structfield> field. The DRM core will
549 then automatically initialize the GEM core before calling the
550 <methodname>load</methodname> operation. Behind the scene, this will
551 create a DRM Memory Manager object which provides an address space
552 pool for object allocation.
555 In a KMS configuration, drivers need to allocate and initialize a
556 command ring buffer following core GEM initialization if required by
557 the hardware. UMA devices usually have what is called a "stolen"
558 memory region, which provides space for the initial framebuffer and
559 large, contiguous memory regions required by the device. This space is
560 typically not managed by GEM, and must be initialized separately into
561 its own DRM MM object.
565 <title>GEM Objects Creation</title>
567 GEM splits creation of GEM objects and allocation of the memory that
568 backs them in two distinct operations.
571 GEM objects are represented by an instance of struct
572 <structname>drm_gem_object</structname>. Drivers usually need to extend
573 GEM objects with private information and thus create a driver-specific
574 GEM object structure type that embeds an instance of struct
575 <structname>drm_gem_object</structname>.
578 To create a GEM object, a driver allocates memory for an instance of its
579 specific GEM object type and initializes the embedded struct
580 <structname>drm_gem_object</structname> with a call to
581 <function>drm_gem_object_init</function>. The function takes a pointer to
582 the DRM device, a pointer to the GEM object and the buffer object size
586 GEM uses shmem to allocate anonymous pageable memory.
587 <function>drm_gem_object_init</function> will create an shmfs file of
588 the requested size and store it into the struct
589 <structname>drm_gem_object</structname> <structfield>filp</structfield>
590 field. The memory is used as either main storage for the object when the
591 graphics hardware uses system memory directly or as a backing store
595 Drivers are responsible for the actual physical pages allocation by
596 calling <function>shmem_read_mapping_page_gfp</function> for each page.
597 Note that they can decide to allocate pages when initializing the GEM
598 object, or to delay allocation until the memory is needed (for instance
599 when a page fault occurs as a result of a userspace memory access or
600 when the driver needs to start a DMA transfer involving the memory).
603 Anonymous pageable memory allocation is not always desired, for instance
604 when the hardware requires physically contiguous system memory as is
605 often the case in embedded devices. Drivers can create GEM objects with
606 no shmfs backing (called private GEM objects) by initializing them with
607 a call to <function>drm_gem_private_object_init</function> instead of
608 <function>drm_gem_object_init</function>. Storage for private GEM
609 objects must be managed by drivers.
612 Drivers that do not need to extend GEM objects with private information
613 can call the <function>drm_gem_object_alloc</function> function to
614 allocate and initialize a struct <structname>drm_gem_object</structname>
615 instance. The GEM core will call the optional driver
616 <methodname>gem_init_object</methodname> operation after initializing
617 the GEM object with <function>drm_gem_object_init</function>.
618 <synopsis>int (*gem_init_object) (struct drm_gem_object *obj);</synopsis>
621 No alloc-and-init function exists for private GEM objects.
625 <title>GEM Objects Lifetime</title>
627 All GEM objects are reference-counted by the GEM core. References can be
628 acquired and release by <function>calling drm_gem_object_reference</function>
629 and <function>drm_gem_object_unreference</function> respectively. The
630 caller must hold the <structname>drm_device</structname>
631 <structfield>struct_mutex</structfield> lock. As a convenience, GEM
632 provides the <function>drm_gem_object_reference_unlocked</function> and
633 <function>drm_gem_object_unreference_unlocked</function> functions that
634 can be called without holding the lock.
637 When the last reference to a GEM object is released the GEM core calls
638 the <structname>drm_driver</structname>
639 <methodname>gem_free_object</methodname> operation. That operation is
640 mandatory for GEM-enabled drivers and must free the GEM object and all
641 associated resources.
644 <synopsis>void (*gem_free_object) (struct drm_gem_object *obj);</synopsis>
645 Drivers are responsible for freeing all GEM object resources, including
646 the resources created by the GEM core. If an mmap offset has been
647 created for the object (in which case
648 <structname>drm_gem_object</structname>::<structfield>map_list</structfield>::<structfield>map</structfield>
649 is not NULL) it must be freed by a call to
650 <function>drm_gem_free_mmap_offset</function>. The shmfs backing store
651 must be released by calling <function>drm_gem_object_release</function>
652 (that function can safely be called if no shmfs backing store has been
657 <title>GEM Objects Naming</title>
659 Communication between userspace and the kernel refers to GEM objects
660 using local handles, global names or, more recently, file descriptors.
661 All of those are 32-bit integer values; the usual Linux kernel limits
662 apply to the file descriptors.
665 GEM handles are local to a DRM file. Applications get a handle to a GEM
666 object through a driver-specific ioctl, and can use that handle to refer
667 to the GEM object in other standard or driver-specific ioctls. Closing a
668 DRM file handle frees all its GEM handles and dereferences the
669 associated GEM objects.
672 To create a handle for a GEM object drivers call
673 <function>drm_gem_handle_create</function>. The function takes a pointer
674 to the DRM file and the GEM object and returns a locally unique handle.
675 When the handle is no longer needed drivers delete it with a call to
676 <function>drm_gem_handle_delete</function>. Finally the GEM object
677 associated with a handle can be retrieved by a call to
678 <function>drm_gem_object_lookup</function>.
681 Handles don't take ownership of GEM objects, they only take a reference
682 to the object that will be dropped when the handle is destroyed. To
683 avoid leaking GEM objects, drivers must make sure they drop the
684 reference(s) they own (such as the initial reference taken at object
685 creation time) as appropriate, without any special consideration for the
686 handle. For example, in the particular case of combined GEM object and
687 handle creation in the implementation of the
688 <methodname>dumb_create</methodname> operation, drivers must drop the
689 initial reference to the GEM object before returning the handle.
692 GEM names are similar in purpose to handles but are not local to DRM
693 files. They can be passed between processes to reference a GEM object
694 globally. Names can't be used directly to refer to objects in the DRM
695 API, applications must convert handles to names and names to handles
696 using the DRM_IOCTL_GEM_FLINK and DRM_IOCTL_GEM_OPEN ioctls
697 respectively. The conversion is handled by the DRM core without any
698 driver-specific support.
701 Similar to global names, GEM file descriptors are also used to share GEM
702 objects across processes. They offer additional security: as file
703 descriptors must be explicitly sent over UNIX domain sockets to be shared
704 between applications, they can't be guessed like the globally unique GEM
708 Drivers that support GEM file descriptors, also known as the DRM PRIME
709 API, must set the DRIVER_PRIME bit in the struct
710 <structname>drm_driver</structname>
711 <structfield>driver_features</structfield> field, and implement the
712 <methodname>prime_handle_to_fd</methodname> and
713 <methodname>prime_fd_to_handle</methodname> operations.
716 <synopsis>int (*prime_handle_to_fd)(struct drm_device *dev,
717 struct drm_file *file_priv, uint32_t handle,
718 uint32_t flags, int *prime_fd);
719 int (*prime_fd_to_handle)(struct drm_device *dev,
720 struct drm_file *file_priv, int prime_fd,
721 uint32_t *handle);</synopsis>
722 Those two operations convert a handle to a PRIME file descriptor and
723 vice versa. Drivers must use the kernel dma-buf buffer sharing framework
724 to manage the PRIME file descriptors.
727 While non-GEM drivers must implement the operations themselves, GEM
728 drivers must use the <function>drm_gem_prime_handle_to_fd</function>
729 and <function>drm_gem_prime_fd_to_handle</function> helper functions.
730 Those helpers rely on the driver
731 <methodname>gem_prime_export</methodname> and
732 <methodname>gem_prime_import</methodname> operations to create a dma-buf
733 instance from a GEM object (dma-buf exporter role) and to create a GEM
734 object from a dma-buf instance (dma-buf importer role).
737 <synopsis>struct dma_buf * (*gem_prime_export)(struct drm_device *dev,
738 struct drm_gem_object *obj,
740 struct drm_gem_object * (*gem_prime_import)(struct drm_device *dev,
741 struct dma_buf *dma_buf);</synopsis>
742 These two operations are mandatory for GEM drivers that support DRM
746 <title>DRM PRIME Helper Functions Reference</title>
747 !Pdrivers/gpu/drm/drm_prime.c PRIME Helpers
750 <sect3 id="drm-gem-objects-mapping">
751 <title>GEM Objects Mapping</title>
753 Because mapping operations are fairly heavyweight GEM favours
754 read/write-like access to buffers, implemented through driver-specific
755 ioctls, over mapping buffers to userspace. However, when random access
756 to the buffer is needed (to perform software rendering for instance),
757 direct access to the object can be more efficient.
760 The mmap system call can't be used directly to map GEM objects, as they
761 don't have their own file handle. Two alternative methods currently
762 co-exist to map GEM objects to userspace. The first method uses a
763 driver-specific ioctl to perform the mapping operation, calling
764 <function>do_mmap</function> under the hood. This is often considered
765 dubious, seems to be discouraged for new GEM-enabled drivers, and will
766 thus not be described here.
769 The second method uses the mmap system call on the DRM file handle.
770 <synopsis>void *mmap(void *addr, size_t length, int prot, int flags, int fd,
771 off_t offset);</synopsis>
772 DRM identifies the GEM object to be mapped by a fake offset passed
773 through the mmap offset argument. Prior to being mapped, a GEM object
774 must thus be associated with a fake offset. To do so, drivers must call
775 <function>drm_gem_create_mmap_offset</function> on the object. The
776 function allocates a fake offset range from a pool and stores the
777 offset divided by PAGE_SIZE in
778 <literal>obj->map_list.hash.key</literal>. Care must be taken not to
779 call <function>drm_gem_create_mmap_offset</function> if a fake offset
780 has already been allocated for the object. This can be tested by
781 <literal>obj->map_list.map</literal> being non-NULL.
784 Once allocated, the fake offset value
785 (<literal>obj->map_list.hash.key << PAGE_SHIFT</literal>)
786 must be passed to the application in a driver-specific way and can then
787 be used as the mmap offset argument.
790 The GEM core provides a helper method <function>drm_gem_mmap</function>
791 to handle object mapping. The method can be set directly as the mmap
792 file operation handler. It will look up the GEM object based on the
793 offset value and set the VMA operations to the
794 <structname>drm_driver</structname> <structfield>gem_vm_ops</structfield>
795 field. Note that <function>drm_gem_mmap</function> doesn't map memory to
796 userspace, but relies on the driver-provided fault handler to map pages
800 To use <function>drm_gem_mmap</function>, drivers must fill the struct
801 <structname>drm_driver</structname> <structfield>gem_vm_ops</structfield>
802 field with a pointer to VM operations.
805 <synopsis>struct vm_operations_struct *gem_vm_ops
807 struct vm_operations_struct {
808 void (*open)(struct vm_area_struct * area);
809 void (*close)(struct vm_area_struct * area);
810 int (*fault)(struct vm_area_struct *vma, struct vm_fault *vmf);
814 The <methodname>open</methodname> and <methodname>close</methodname>
815 operations must update the GEM object reference count. Drivers can use
816 the <function>drm_gem_vm_open</function> and
817 <function>drm_gem_vm_close</function> helper functions directly as open
821 The fault operation handler is responsible for mapping individual pages
822 to userspace when a page fault occurs. Depending on the memory
823 allocation scheme, drivers can allocate pages at fault time, or can
824 decide to allocate memory for the GEM object at the time the object is
828 Drivers that want to map the GEM object upfront instead of handling page
829 faults can implement their own mmap file operation handler.
833 <title>Memory Coherency</title>
835 When mapped to the device or used in a command buffer, backing pages
836 for an object are flushed to memory and marked write combined so as to
837 be coherent with the GPU. Likewise, if the CPU accesses an object
838 after the GPU has finished rendering to the object, then the object
839 must be made coherent with the CPU's view of memory, usually involving
840 GPU cache flushing of various kinds. This core CPU<->GPU
841 coherency management is provided by a device-specific ioctl, which
842 evaluates an object's current domain and performs any necessary
843 flushing or synchronization to put the object into the desired
844 coherency domain (note that the object may be busy, i.e. an active
845 render target; in that case, setting the domain blocks the client and
846 waits for rendering to complete before performing any necessary
847 flushing operations).
851 <title>Command Execution</title>
853 Perhaps the most important GEM function for GPU devices is providing a
854 command execution interface to clients. Client programs construct
855 command buffers containing references to previously allocated memory
856 objects, and then submit them to GEM. At that point, GEM takes care to
857 bind all the objects into the GTT, execute the buffer, and provide
858 necessary synchronization between clients accessing the same buffers.
859 This often involves evicting some objects from the GTT and re-binding
860 others (a fairly expensive operation), and providing relocation
861 support which hides fixed GTT offsets from clients. Clients must take
862 care not to submit command buffers that reference more objects than
863 can fit in the GTT; otherwise, GEM will reject them and no rendering
864 will occur. Similarly, if several objects in the buffer require fence
865 registers to be allocated for correct rendering (e.g. 2D blits on
866 pre-965 chips), care must be taken not to require more fence registers
867 than are available to the client. Such resource management should be
868 abstracted from the client in libdrm.
872 <title>GEM Function Reference</title>
873 !Edrivers/gpu/drm/drm_gem.c
877 <title>VMA Offset Manager</title>
878 !Pdrivers/gpu/drm/drm_vma_manager.c vma offset manager
879 !Edrivers/gpu/drm/drm_vma_manager.c
880 !Iinclude/drm/drm_vma_manager.h
884 <!-- Internals: mode setting -->
886 <sect1 id="drm-mode-setting">
887 <title>Mode Setting</title>
889 Drivers must initialize the mode setting core by calling
890 <function>drm_mode_config_init</function> on the DRM device. The function
891 initializes the <structname>drm_device</structname>
892 <structfield>mode_config</structfield> field and never fails. Once done,
893 mode configuration must be setup by initializing the following fields.
897 <synopsis>int min_width, min_height;
898 int max_width, max_height;</synopsis>
900 Minimum and maximum width and height of the frame buffers in pixel
905 <synopsis>struct drm_mode_config_funcs *funcs;</synopsis>
906 <para>Mode setting functions.</para>
910 <title>Frame Buffer Creation</title>
911 <synopsis>struct drm_framebuffer *(*fb_create)(struct drm_device *dev,
912 struct drm_file *file_priv,
913 struct drm_mode_fb_cmd2 *mode_cmd);</synopsis>
915 Frame buffers are abstract memory objects that provide a source of
916 pixels to scanout to a CRTC. Applications explicitly request the
917 creation of frame buffers through the DRM_IOCTL_MODE_ADDFB(2) ioctls and
918 receive an opaque handle that can be passed to the KMS CRTC control,
919 plane configuration and page flip functions.
922 Frame buffers rely on the underneath memory manager for low-level memory
923 operations. When creating a frame buffer applications pass a memory
924 handle (or a list of memory handles for multi-planar formats) through
925 the <parameter>drm_mode_fb_cmd2</parameter> argument. For drivers using
926 GEM as their userspace buffer management interface this would be a GEM
927 handle. Drivers are however free to use their own backing storage object
928 handles, e.g. vmwgfx directly exposes special TTM handles to userspace
929 and so expects TTM handles in the create ioctl and not GEM handles.
932 Drivers must first validate the requested frame buffer parameters passed
933 through the mode_cmd argument. In particular this is where invalid
934 sizes, pixel formats or pitches can be caught.
937 If the parameters are deemed valid, drivers then create, initialize and
938 return an instance of struct <structname>drm_framebuffer</structname>.
939 If desired the instance can be embedded in a larger driver-specific
940 structure. Drivers must fill its <structfield>width</structfield>,
941 <structfield>height</structfield>, <structfield>pitches</structfield>,
942 <structfield>offsets</structfield>, <structfield>depth</structfield>,
943 <structfield>bits_per_pixel</structfield> and
944 <structfield>pixel_format</structfield> fields from the values passed
945 through the <parameter>drm_mode_fb_cmd2</parameter> argument. They
946 should call the <function>drm_helper_mode_fill_fb_struct</function>
947 helper function to do so.
951 The initialization of the new framebuffer instance is finalized with a
952 call to <function>drm_framebuffer_init</function> which takes a pointer
953 to DRM frame buffer operations (struct
954 <structname>drm_framebuffer_funcs</structname>). Note that this function
955 publishes the framebuffer and so from this point on it can be accessed
956 concurrently from other threads. Hence it must be the last step in the
957 driver's framebuffer initialization sequence. Frame buffer operations
961 <synopsis>int (*create_handle)(struct drm_framebuffer *fb,
962 struct drm_file *file_priv, unsigned int *handle);</synopsis>
964 Create a handle to the frame buffer underlying memory object. If
965 the frame buffer uses a multi-plane format, the handle will
966 reference the memory object associated with the first plane.
969 Drivers call <function>drm_gem_handle_create</function> to create
974 <synopsis>void (*destroy)(struct drm_framebuffer *framebuffer);</synopsis>
976 Destroy the frame buffer object and frees all associated
977 resources. Drivers must call
978 <function>drm_framebuffer_cleanup</function> to free resources
979 allocated by the DRM core for the frame buffer object, and must
980 make sure to unreference all memory objects associated with the
981 frame buffer. Handles created by the
982 <methodname>create_handle</methodname> operation are released by
987 <synopsis>int (*dirty)(struct drm_framebuffer *framebuffer,
988 struct drm_file *file_priv, unsigned flags, unsigned color,
989 struct drm_clip_rect *clips, unsigned num_clips);</synopsis>
991 This optional operation notifies the driver that a region of the
992 frame buffer has changed in response to a DRM_IOCTL_MODE_DIRTYFB
999 The lifetime of a drm framebuffer is controlled with a reference count,
1000 drivers can grab additional references with
1001 <function>drm_framebuffer_reference</function> </para> and drop them
1002 again with <function>drm_framebuffer_unreference</function>. For
1003 driver-private framebuffers for which the last reference is never
1004 dropped (e.g. for the fbdev framebuffer when the struct
1005 <structname>drm_framebuffer</structname> is embedded into the fbdev
1006 helper struct) drivers can manually clean up a framebuffer at module
1008 <function>drm_framebuffer_unregister_private</function>.
1011 <title>Dumb Buffer Objects</title>
1013 The KMS API doesn't standardize backing storage object creation and
1014 leaves it to driver-specific ioctls. Furthermore actually creating a
1015 buffer object even for GEM-based drivers is done through a
1016 driver-specific ioctl - GEM only has a common userspace interface for
1017 sharing and destroying objects. While not an issue for full-fledged
1018 graphics stacks that include device-specific userspace components (in
1019 libdrm for instance), this limit makes DRM-based early boot graphics
1020 unnecessarily complex.
1023 Dumb objects partly alleviate the problem by providing a standard
1024 API to create dumb buffers suitable for scanout, which can then be used
1025 to create KMS frame buffers.
1028 To support dumb objects drivers must implement the
1029 <methodname>dumb_create</methodname>,
1030 <methodname>dumb_destroy</methodname> and
1031 <methodname>dumb_map_offset</methodname> operations.
1035 <synopsis>int (*dumb_create)(struct drm_file *file_priv, struct drm_device *dev,
1036 struct drm_mode_create_dumb *args);</synopsis>
1038 The <methodname>dumb_create</methodname> operation creates a driver
1039 object (GEM or TTM handle) suitable for scanout based on the
1040 width, height and depth from the struct
1041 <structname>drm_mode_create_dumb</structname> argument. It fills the
1042 argument's <structfield>handle</structfield>,
1043 <structfield>pitch</structfield> and <structfield>size</structfield>
1044 fields with a handle for the newly created object and its line
1045 pitch and size in bytes.
1049 <synopsis>int (*dumb_destroy)(struct drm_file *file_priv, struct drm_device *dev,
1050 uint32_t handle);</synopsis>
1052 The <methodname>dumb_destroy</methodname> operation destroys a dumb
1053 object created by <methodname>dumb_create</methodname>.
1057 <synopsis>int (*dumb_map_offset)(struct drm_file *file_priv, struct drm_device *dev,
1058 uint32_t handle, uint64_t *offset);</synopsis>
1060 The <methodname>dumb_map_offset</methodname> operation associates an
1061 mmap fake offset with the object given by the handle and returns
1062 it. Drivers must use the
1063 <function>drm_gem_create_mmap_offset</function> function to
1064 associate the fake offset as described in
1065 <xref linkend="drm-gem-objects-mapping"/>.
1070 Note that dumb objects may not be used for gpu acceleration, as has been
1071 attempted on some ARM embedded platforms. Such drivers really must have
1072 a hardware-specific ioctl to allocate suitable buffer objects.
1076 <title>Output Polling</title>
1077 <synopsis>void (*output_poll_changed)(struct drm_device *dev);</synopsis>
1079 This operation notifies the driver that the status of one or more
1080 connectors has changed. Drivers that use the fb helper can just call the
1081 <function>drm_fb_helper_hotplug_event</function> function to handle this
1086 <title>Locking</title>
1088 Beside some lookup structures with their own locking (which is hidden
1089 behind the interface functions) most of the modeset state is protected
1090 by the <code>dev-<mode_config.lock</code> mutex and additionally
1091 per-crtc locks to allow cursor updates, pageflips and similar operations
1092 to occur concurrently with background tasks like output detection.
1093 Operations which cross domains like a full modeset always grab all
1094 locks. Drivers there need to protect resources shared between crtcs with
1095 additional locking. They also need to be careful to always grab the
1096 relevant crtc locks if a modset functions touches crtc state, e.g. for
1097 load detection (which does only grab the <code>mode_config.lock</code>
1098 to allow concurrent screen updates on live crtcs).
1103 <!-- Internals: kms initialization and cleanup -->
1105 <sect1 id="drm-kms-init">
1106 <title>KMS Initialization and Cleanup</title>
1108 A KMS device is abstracted and exposed as a set of planes, CRTCs, encoders
1109 and connectors. KMS drivers must thus create and initialize all those
1110 objects at load time after initializing mode setting.
1113 <title>CRTCs (struct <structname>drm_crtc</structname>)</title>
1115 A CRTC is an abstraction representing a part of the chip that contains a
1116 pointer to a scanout buffer. Therefore, the number of CRTCs available
1117 determines how many independent scanout buffers can be active at any
1118 given time. The CRTC structure contains several fields to support this:
1119 a pointer to some video memory (abstracted as a frame buffer object), a
1120 display mode, and an (x, y) offset into the video memory to support
1121 panning or configurations where one piece of video memory spans multiple
1125 <title>CRTC Initialization</title>
1127 A KMS device must create and register at least one struct
1128 <structname>drm_crtc</structname> instance. The instance is allocated
1129 and zeroed by the driver, possibly as part of a larger structure, and
1130 registered with a call to <function>drm_crtc_init</function> with a
1131 pointer to CRTC functions.
1135 <title>CRTC Operations</title>
1137 <title>Set Configuration</title>
1138 <synopsis>int (*set_config)(struct drm_mode_set *set);</synopsis>
1140 Apply a new CRTC configuration to the device. The configuration
1141 specifies a CRTC, a frame buffer to scan out from, a (x,y) position in
1142 the frame buffer, a display mode and an array of connectors to drive
1143 with the CRTC if possible.
1146 If the frame buffer specified in the configuration is NULL, the driver
1147 must detach all encoders connected to the CRTC and all connectors
1148 attached to those encoders and disable them.
1151 This operation is called with the mode config lock held.
1154 Note that the drm core has no notion of restoring the mode setting
1155 state after resume, since all resume handling is in the full
1156 responsibility of the driver. The common mode setting helper library
1157 though provides a helper which can be used for this:
1158 <function>drm_helper_resume_force_mode</function>.
1162 <title>Page Flipping</title>
1163 <synopsis>int (*page_flip)(struct drm_crtc *crtc, struct drm_framebuffer *fb,
1164 struct drm_pending_vblank_event *event);</synopsis>
1166 Schedule a page flip to the given frame buffer for the CRTC. This
1167 operation is called with the mode config mutex held.
1170 Page flipping is a synchronization mechanism that replaces the frame
1171 buffer being scanned out by the CRTC with a new frame buffer during
1172 vertical blanking, avoiding tearing. When an application requests a page
1173 flip the DRM core verifies that the new frame buffer is large enough to
1174 be scanned out by the CRTC in the currently configured mode and then
1175 calls the CRTC <methodname>page_flip</methodname> operation with a
1176 pointer to the new frame buffer.
1179 The <methodname>page_flip</methodname> operation schedules a page flip.
1180 Once any pending rendering targeting the new frame buffer has
1181 completed, the CRTC will be reprogrammed to display that frame buffer
1182 after the next vertical refresh. The operation must return immediately
1183 without waiting for rendering or page flip to complete and must block
1184 any new rendering to the frame buffer until the page flip completes.
1187 If a page flip can be successfully scheduled the driver must set the
1188 <code>drm_crtc-<fb</code> field to the new framebuffer pointed to
1189 by <code>fb</code>. This is important so that the reference counting
1190 on framebuffers stays balanced.
1193 If a page flip is already pending, the
1194 <methodname>page_flip</methodname> operation must return
1195 -<errorname>EBUSY</errorname>.
1198 To synchronize page flip to vertical blanking the driver will likely
1199 need to enable vertical blanking interrupts. It should call
1200 <function>drm_vblank_get</function> for that purpose, and call
1201 <function>drm_vblank_put</function> after the page flip completes.
1204 If the application has requested to be notified when page flip completes
1205 the <methodname>page_flip</methodname> operation will be called with a
1206 non-NULL <parameter>event</parameter> argument pointing to a
1207 <structname>drm_pending_vblank_event</structname> instance. Upon page
1208 flip completion the driver must call <methodname>drm_send_vblank_event</methodname>
1209 to fill in the event and send to wake up any waiting processes.
1210 This can be performed with
1211 <programlisting><![CDATA[
1212 spin_lock_irqsave(&dev->event_lock, flags);
1214 drm_send_vblank_event(dev, pipe, event);
1215 spin_unlock_irqrestore(&dev->event_lock, flags);
1216 ]]></programlisting>
1219 FIXME: Could drivers that don't need to wait for rendering to complete
1220 just add the event to <literal>dev->vblank_event_list</literal> and
1221 let the DRM core handle everything, as for "normal" vertical blanking
1225 While waiting for the page flip to complete, the
1226 <literal>event->base.link</literal> list head can be used freely by
1227 the driver to store the pending event in a driver-specific list.
1230 If the file handle is closed before the event is signaled, drivers must
1231 take care to destroy the event in their
1232 <methodname>preclose</methodname> operation (and, if needed, call
1233 <function>drm_vblank_put</function>).
1237 <title>Miscellaneous</title>
1240 <synopsis>void (*set_property)(struct drm_crtc *crtc,
1241 struct drm_property *property, uint64_t value);</synopsis>
1243 Set the value of the given CRTC property to
1244 <parameter>value</parameter>. See <xref linkend="drm-kms-properties"/>
1245 for more information about properties.
1249 <synopsis>void (*gamma_set)(struct drm_crtc *crtc, u16 *r, u16 *g, u16 *b,
1250 uint32_t start, uint32_t size);</synopsis>
1252 Apply a gamma table to the device. The operation is optional.
1256 <synopsis>void (*destroy)(struct drm_crtc *crtc);</synopsis>
1258 Destroy the CRTC when not needed anymore. See
1259 <xref linkend="drm-kms-init"/>.
1267 <title>Planes (struct <structname>drm_plane</structname>)</title>
1269 A plane represents an image source that can be blended with or overlayed
1270 on top of a CRTC during the scanout process. Planes are associated with
1271 a frame buffer to crop a portion of the image memory (source) and
1272 optionally scale it to a destination size. The result is then blended
1273 with or overlayed on top of a CRTC.
1276 <title>Plane Initialization</title>
1278 Planes are optional. To create a plane, a KMS drivers allocates and
1279 zeroes an instances of struct <structname>drm_plane</structname>
1280 (possibly as part of a larger structure) and registers it with a call
1281 to <function>drm_plane_init</function>. The function takes a bitmask
1282 of the CRTCs that can be associated with the plane, a pointer to the
1283 plane functions and a list of format supported formats.
1287 <title>Plane Operations</title>
1290 <synopsis>int (*update_plane)(struct drm_plane *plane, struct drm_crtc *crtc,
1291 struct drm_framebuffer *fb, int crtc_x, int crtc_y,
1292 unsigned int crtc_w, unsigned int crtc_h,
1293 uint32_t src_x, uint32_t src_y,
1294 uint32_t src_w, uint32_t src_h);</synopsis>
1296 Enable and configure the plane to use the given CRTC and frame buffer.
1299 The source rectangle in frame buffer memory coordinates is given by
1300 the <parameter>src_x</parameter>, <parameter>src_y</parameter>,
1301 <parameter>src_w</parameter> and <parameter>src_h</parameter>
1302 parameters (as 16.16 fixed point values). Devices that don't support
1303 subpixel plane coordinates can ignore the fractional part.
1306 The destination rectangle in CRTC coordinates is given by the
1307 <parameter>crtc_x</parameter>, <parameter>crtc_y</parameter>,
1308 <parameter>crtc_w</parameter> and <parameter>crtc_h</parameter>
1309 parameters (as integer values). Devices scale the source rectangle to
1310 the destination rectangle. If scaling is not supported, and the source
1311 rectangle size doesn't match the destination rectangle size, the
1312 driver must return a -<errorname>EINVAL</errorname> error.
1316 <synopsis>int (*disable_plane)(struct drm_plane *plane);</synopsis>
1318 Disable the plane. The DRM core calls this method in response to a
1319 DRM_IOCTL_MODE_SETPLANE ioctl call with the frame buffer ID set to 0.
1320 Disabled planes must not be processed by the CRTC.
1324 <synopsis>void (*destroy)(struct drm_plane *plane);</synopsis>
1326 Destroy the plane when not needed anymore. See
1327 <xref linkend="drm-kms-init"/>.
1334 <title>Encoders (struct <structname>drm_encoder</structname>)</title>
1336 An encoder takes pixel data from a CRTC and converts it to a format
1337 suitable for any attached connectors. On some devices, it may be
1338 possible to have a CRTC send data to more than one encoder. In that
1339 case, both encoders would receive data from the same scanout buffer,
1340 resulting in a "cloned" display configuration across the connectors
1341 attached to each encoder.
1344 <title>Encoder Initialization</title>
1346 As for CRTCs, a KMS driver must create, initialize and register at
1347 least one struct <structname>drm_encoder</structname> instance. The
1348 instance is allocated and zeroed by the driver, possibly as part of a
1352 Drivers must initialize the struct <structname>drm_encoder</structname>
1353 <structfield>possible_crtcs</structfield> and
1354 <structfield>possible_clones</structfield> fields before registering the
1355 encoder. Both fields are bitmasks of respectively the CRTCs that the
1356 encoder can be connected to, and sibling encoders candidate for cloning.
1359 After being initialized, the encoder must be registered with a call to
1360 <function>drm_encoder_init</function>. The function takes a pointer to
1361 the encoder functions and an encoder type. Supported types are
1364 DRM_MODE_ENCODER_DAC for VGA and analog on DVI-I/DVI-A
1367 DRM_MODE_ENCODER_TMDS for DVI, HDMI and (embedded) DisplayPort
1370 DRM_MODE_ENCODER_LVDS for display panels
1373 DRM_MODE_ENCODER_TVDAC for TV output (Composite, S-Video, Component,
1377 DRM_MODE_ENCODER_VIRTUAL for virtual machine displays
1382 Encoders must be attached to a CRTC to be used. DRM drivers leave
1383 encoders unattached at initialization time. Applications (or the fbdev
1384 compatibility layer when implemented) are responsible for attaching the
1385 encoders they want to use to a CRTC.
1389 <title>Encoder Operations</title>
1392 <synopsis>void (*destroy)(struct drm_encoder *encoder);</synopsis>
1394 Called to destroy the encoder when not needed anymore. See
1395 <xref linkend="drm-kms-init"/>.
1399 <synopsis>void (*set_property)(struct drm_plane *plane,
1400 struct drm_property *property, uint64_t value);</synopsis>
1402 Set the value of the given plane property to
1403 <parameter>value</parameter>. See <xref linkend="drm-kms-properties"/>
1404 for more information about properties.
1411 <title>Connectors (struct <structname>drm_connector</structname>)</title>
1413 A connector is the final destination for pixel data on a device, and
1414 usually connects directly to an external display device like a monitor
1415 or laptop panel. A connector can only be attached to one encoder at a
1416 time. The connector is also the structure where information about the
1417 attached display is kept, so it contains fields for display data, EDID
1418 data, DPMS & connection status, and information about modes
1419 supported on the attached displays.
1422 <title>Connector Initialization</title>
1424 Finally a KMS driver must create, initialize, register and attach at
1425 least one struct <structname>drm_connector</structname> instance. The
1426 instance is created as other KMS objects and initialized by setting the
1431 <term><structfield>interlace_allowed</structfield></term>
1433 Whether the connector can handle interlaced modes.
1437 <term><structfield>doublescan_allowed</structfield></term>
1439 Whether the connector can handle doublescan.
1443 <term><structfield>display_info
1444 </structfield></term>
1446 Display information is filled from EDID information when a display
1447 is detected. For non hot-pluggable displays such as flat panels in
1448 embedded systems, the driver should initialize the
1449 <structfield>display_info</structfield>.<structfield>width_mm</structfield>
1451 <structfield>display_info</structfield>.<structfield>height_mm</structfield>
1452 fields with the physical size of the display.
1456 <term id="drm-kms-connector-polled"><structfield>polled</structfield></term>
1458 Connector polling mode, a combination of
1461 <term>DRM_CONNECTOR_POLL_HPD</term>
1463 The connector generates hotplug events and doesn't need to be
1464 periodically polled. The CONNECT and DISCONNECT flags must not
1465 be set together with the HPD flag.
1469 <term>DRM_CONNECTOR_POLL_CONNECT</term>
1471 Periodically poll the connector for connection.
1475 <term>DRM_CONNECTOR_POLL_DISCONNECT</term>
1477 Periodically poll the connector for disconnection.
1481 Set to 0 for connectors that don't support connection status
1487 The connector is then registered with a call to
1488 <function>drm_connector_init</function> with a pointer to the connector
1489 functions and a connector type, and exposed through sysfs with a call to
1490 <function>drm_sysfs_connector_add</function>.
1493 Supported connector types are
1495 <listitem>DRM_MODE_CONNECTOR_VGA</listitem>
1496 <listitem>DRM_MODE_CONNECTOR_DVII</listitem>
1497 <listitem>DRM_MODE_CONNECTOR_DVID</listitem>
1498 <listitem>DRM_MODE_CONNECTOR_DVIA</listitem>
1499 <listitem>DRM_MODE_CONNECTOR_Composite</listitem>
1500 <listitem>DRM_MODE_CONNECTOR_SVIDEO</listitem>
1501 <listitem>DRM_MODE_CONNECTOR_LVDS</listitem>
1502 <listitem>DRM_MODE_CONNECTOR_Component</listitem>
1503 <listitem>DRM_MODE_CONNECTOR_9PinDIN</listitem>
1504 <listitem>DRM_MODE_CONNECTOR_DisplayPort</listitem>
1505 <listitem>DRM_MODE_CONNECTOR_HDMIA</listitem>
1506 <listitem>DRM_MODE_CONNECTOR_HDMIB</listitem>
1507 <listitem>DRM_MODE_CONNECTOR_TV</listitem>
1508 <listitem>DRM_MODE_CONNECTOR_eDP</listitem>
1509 <listitem>DRM_MODE_CONNECTOR_VIRTUAL</listitem>
1513 Connectors must be attached to an encoder to be used. For devices that
1514 map connectors to encoders 1:1, the connector should be attached at
1515 initialization time with a call to
1516 <function>drm_mode_connector_attach_encoder</function>. The driver must
1517 also set the <structname>drm_connector</structname>
1518 <structfield>encoder</structfield> field to point to the attached
1522 Finally, drivers must initialize the connectors state change detection
1523 with a call to <function>drm_kms_helper_poll_init</function>. If at
1524 least one connector is pollable but can't generate hotplug interrupts
1525 (indicated by the DRM_CONNECTOR_POLL_CONNECT and
1526 DRM_CONNECTOR_POLL_DISCONNECT connector flags), a delayed work will
1527 automatically be queued to periodically poll for changes. Connectors
1528 that can generate hotplug interrupts must be marked with the
1529 DRM_CONNECTOR_POLL_HPD flag instead, and their interrupt handler must
1530 call <function>drm_helper_hpd_irq_event</function>. The function will
1531 queue a delayed work to check the state of all connectors, but no
1532 periodic polling will be done.
1536 <title>Connector Operations</title>
1538 Unless otherwise state, all operations are mandatory.
1542 <synopsis>void (*dpms)(struct drm_connector *connector, int mode);</synopsis>
1544 The DPMS operation sets the power state of a connector. The mode
1547 <listitem><para>DRM_MODE_DPMS_ON</para></listitem>
1548 <listitem><para>DRM_MODE_DPMS_STANDBY</para></listitem>
1549 <listitem><para>DRM_MODE_DPMS_SUSPEND</para></listitem>
1550 <listitem><para>DRM_MODE_DPMS_OFF</para></listitem>
1554 In all but DPMS_ON mode the encoder to which the connector is attached
1555 should put the display in low-power mode by driving its signals
1556 appropriately. If more than one connector is attached to the encoder
1557 care should be taken not to change the power state of other displays as
1558 a side effect. Low-power mode should be propagated to the encoders and
1559 CRTCs when all related connectors are put in low-power mode.
1563 <title>Modes</title>
1564 <synopsis>int (*fill_modes)(struct drm_connector *connector, uint32_t max_width,
1565 uint32_t max_height);</synopsis>
1567 Fill the mode list with all supported modes for the connector. If the
1568 <parameter>max_width</parameter> and <parameter>max_height</parameter>
1569 arguments are non-zero, the implementation must ignore all modes wider
1570 than <parameter>max_width</parameter> or higher than
1571 <parameter>max_height</parameter>.
1574 The connector must also fill in this operation its
1575 <structfield>display_info</structfield>
1576 <structfield>width_mm</structfield> and
1577 <structfield>height_mm</structfield> fields with the connected display
1578 physical size in millimeters. The fields should be set to 0 if the value
1579 isn't known or is not applicable (for instance for projector devices).
1583 <title>Connection Status</title>
1585 The connection status is updated through polling or hotplug events when
1586 supported (see <xref linkend="drm-kms-connector-polled"/>). The status
1587 value is reported to userspace through ioctls and must not be used
1588 inside the driver, as it only gets initialized by a call to
1589 <function>drm_mode_getconnector</function> from userspace.
1591 <synopsis>enum drm_connector_status (*detect)(struct drm_connector *connector,
1592 bool force);</synopsis>
1594 Check to see if anything is attached to the connector. The
1595 <parameter>force</parameter> parameter is set to false whilst polling or
1596 to true when checking the connector due to user request.
1597 <parameter>force</parameter> can be used by the driver to avoid
1598 expensive, destructive operations during automated probing.
1601 Return connector_status_connected if something is connected to the
1602 connector, connector_status_disconnected if nothing is connected and
1603 connector_status_unknown if the connection state isn't known.
1606 Drivers should only return connector_status_connected if the connection
1607 status has really been probed as connected. Connectors that can't detect
1608 the connection status, or failed connection status probes, should return
1609 connector_status_unknown.
1613 <title>Miscellaneous</title>
1616 <synopsis>void (*set_property)(struct drm_connector *connector,
1617 struct drm_property *property, uint64_t value);</synopsis>
1619 Set the value of the given connector property to
1620 <parameter>value</parameter>. See <xref linkend="drm-kms-properties"/>
1621 for more information about properties.
1625 <synopsis>void (*destroy)(struct drm_connector *connector);</synopsis>
1627 Destroy the connector when not needed anymore. See
1628 <xref linkend="drm-kms-init"/>.
1636 <title>Cleanup</title>
1638 The DRM core manages its objects' lifetime. When an object is not needed
1639 anymore the core calls its destroy function, which must clean up and
1640 free every resource allocated for the object. Every
1641 <function>drm_*_init</function> call must be matched with a
1642 corresponding <function>drm_*_cleanup</function> call to cleanup CRTCs
1643 (<function>drm_crtc_cleanup</function>), planes
1644 (<function>drm_plane_cleanup</function>), encoders
1645 (<function>drm_encoder_cleanup</function>) and connectors
1646 (<function>drm_connector_cleanup</function>). Furthermore, connectors
1647 that have been added to sysfs must be removed by a call to
1648 <function>drm_sysfs_connector_remove</function> before calling
1649 <function>drm_connector_cleanup</function>.
1652 Connectors state change detection must be cleanup up with a call to
1653 <function>drm_kms_helper_poll_fini</function>.
1657 <title>Output discovery and initialization example</title>
1658 <programlisting><![CDATA[
1659 void intel_crt_init(struct drm_device *dev)
1661 struct drm_connector *connector;
1662 struct intel_output *intel_output;
1664 intel_output = kzalloc(sizeof(struct intel_output), GFP_KERNEL);
1668 connector = &intel_output->base;
1669 drm_connector_init(dev, &intel_output->base,
1670 &intel_crt_connector_funcs, DRM_MODE_CONNECTOR_VGA);
1672 drm_encoder_init(dev, &intel_output->enc, &intel_crt_enc_funcs,
1673 DRM_MODE_ENCODER_DAC);
1675 drm_mode_connector_attach_encoder(&intel_output->base,
1676 &intel_output->enc);
1678 /* Set up the DDC bus. */
1679 intel_output->ddc_bus = intel_i2c_create(dev, GPIOA, "CRTDDC_A");
1680 if (!intel_output->ddc_bus) {
1681 dev_printk(KERN_ERR, &dev->pdev->dev, "DDC bus registration "
1686 intel_output->type = INTEL_OUTPUT_ANALOG;
1687 connector->interlace_allowed = 0;
1688 connector->doublescan_allowed = 0;
1690 drm_encoder_helper_add(&intel_output->enc, &intel_crt_helper_funcs);
1691 drm_connector_helper_add(connector, &intel_crt_connector_helper_funcs);
1693 drm_sysfs_connector_add(connector);
1694 }]]></programlisting>
1696 In the example above (taken from the i915 driver), a CRTC, connector and
1697 encoder combination is created. A device-specific i2c bus is also
1698 created for fetching EDID data and performing monitor detection. Once
1699 the process is complete, the new connector is registered with sysfs to
1700 make its properties available to applications.
1704 <title>KMS API Functions</title>
1705 !Edrivers/gpu/drm/drm_crtc.c
1709 <!-- Internals: kms helper functions -->
1712 <title>Mode Setting Helper Functions</title>
1714 The CRTC, encoder and connector functions provided by the drivers
1715 implement the DRM API. They're called by the DRM core and ioctl handlers
1716 to handle device state changes and configuration request. As implementing
1717 those functions often requires logic not specific to drivers, mid-layer
1718 helper functions are available to avoid duplicating boilerplate code.
1721 The DRM core contains one mid-layer implementation. The mid-layer provides
1722 implementations of several CRTC, encoder and connector functions (called
1723 from the top of the mid-layer) that pre-process requests and call
1724 lower-level functions provided by the driver (at the bottom of the
1725 mid-layer). For instance, the
1726 <function>drm_crtc_helper_set_config</function> function can be used to
1727 fill the struct <structname>drm_crtc_funcs</structname>
1728 <structfield>set_config</structfield> field. When called, it will split
1729 the <methodname>set_config</methodname> operation in smaller, simpler
1730 operations and call the driver to handle them.
1733 To use the mid-layer, drivers call <function>drm_crtc_helper_add</function>,
1734 <function>drm_encoder_helper_add</function> and
1735 <function>drm_connector_helper_add</function> functions to install their
1736 mid-layer bottom operations handlers, and fill the
1737 <structname>drm_crtc_funcs</structname>,
1738 <structname>drm_encoder_funcs</structname> and
1739 <structname>drm_connector_funcs</structname> structures with pointers to
1740 the mid-layer top API functions. Installing the mid-layer bottom operation
1741 handlers is best done right after registering the corresponding KMS object.
1744 The mid-layer is not split between CRTC, encoder and connector operations.
1745 To use it, a driver must provide bottom functions for all of the three KMS
1749 <title>Helper Functions</title>
1752 <synopsis>int drm_crtc_helper_set_config(struct drm_mode_set *set);</synopsis>
1754 The <function>drm_crtc_helper_set_config</function> helper function
1755 is a CRTC <methodname>set_config</methodname> implementation. It
1756 first tries to locate the best encoder for each connector by calling
1757 the connector <methodname>best_encoder</methodname> helper
1761 After locating the appropriate encoders, the helper function will
1762 call the <methodname>mode_fixup</methodname> encoder and CRTC helper
1763 operations to adjust the requested mode, or reject it completely in
1764 which case an error will be returned to the application. If the new
1765 configuration after mode adjustment is identical to the current
1766 configuration the helper function will return without performing any
1770 If the adjusted mode is identical to the current mode but changes to
1771 the frame buffer need to be applied, the
1772 <function>drm_crtc_helper_set_config</function> function will call
1773 the CRTC <methodname>mode_set_base</methodname> helper operation. If
1774 the adjusted mode differs from the current mode, or if the
1775 <methodname>mode_set_base</methodname> helper operation is not
1776 provided, the helper function performs a full mode set sequence by
1777 calling the <methodname>prepare</methodname>,
1778 <methodname>mode_set</methodname> and
1779 <methodname>commit</methodname> CRTC and encoder helper operations,
1784 <synopsis>void drm_helper_connector_dpms(struct drm_connector *connector, int mode);</synopsis>
1786 The <function>drm_helper_connector_dpms</function> helper function
1787 is a connector <methodname>dpms</methodname> implementation that
1788 tracks power state of connectors. To use the function, drivers must
1789 provide <methodname>dpms</methodname> helper operations for CRTCs
1790 and encoders to apply the DPMS state to the device.
1793 The mid-layer doesn't track the power state of CRTCs and encoders.
1794 The <methodname>dpms</methodname> helper operations can thus be
1795 called with a mode identical to the currently active mode.
1799 <synopsis>int drm_helper_probe_single_connector_modes(struct drm_connector *connector,
1800 uint32_t maxX, uint32_t maxY);</synopsis>
1802 The <function>drm_helper_probe_single_connector_modes</function> helper
1803 function is a connector <methodname>fill_modes</methodname>
1804 implementation that updates the connection status for the connector
1805 and then retrieves a list of modes by calling the connector
1806 <methodname>get_modes</methodname> helper operation.
1809 The function filters out modes larger than
1810 <parameter>max_width</parameter> and <parameter>max_height</parameter>
1811 if specified. It then calls the connector
1812 <methodname>mode_valid</methodname> helper operation for each mode in
1813 the probed list to check whether the mode is valid for the connector.
1819 <title>CRTC Helper Operations</title>
1821 <listitem id="drm-helper-crtc-mode-fixup">
1822 <synopsis>bool (*mode_fixup)(struct drm_crtc *crtc,
1823 const struct drm_display_mode *mode,
1824 struct drm_display_mode *adjusted_mode);</synopsis>
1826 Let CRTCs adjust the requested mode or reject it completely. This
1827 operation returns true if the mode is accepted (possibly after being
1828 adjusted) or false if it is rejected.
1831 The <methodname>mode_fixup</methodname> operation should reject the
1832 mode if it can't reasonably use it. The definition of "reasonable"
1833 is currently fuzzy in this context. One possible behaviour would be
1834 to set the adjusted mode to the panel timings when a fixed-mode
1835 panel is used with hardware capable of scaling. Another behaviour
1836 would be to accept any input mode and adjust it to the closest mode
1837 supported by the hardware (FIXME: This needs to be clarified).
1841 <synopsis>int (*mode_set_base)(struct drm_crtc *crtc, int x, int y,
1842 struct drm_framebuffer *old_fb)</synopsis>
1844 Move the CRTC on the current frame buffer (stored in
1845 <literal>crtc->fb</literal>) to position (x,y). Any of the frame
1846 buffer, x position or y position may have been modified.
1849 This helper operation is optional. If not provided, the
1850 <function>drm_crtc_helper_set_config</function> function will fall
1851 back to the <methodname>mode_set</methodname> helper operation.
1854 FIXME: Why are x and y passed as arguments, as they can be accessed
1855 through <literal>crtc->x</literal> and
1856 <literal>crtc->y</literal>?
1860 <synopsis>void (*prepare)(struct drm_crtc *crtc);</synopsis>
1862 Prepare the CRTC for mode setting. This operation is called after
1863 validating the requested mode. Drivers use it to perform
1864 device-specific operations required before setting the new mode.
1868 <synopsis>int (*mode_set)(struct drm_crtc *crtc, struct drm_display_mode *mode,
1869 struct drm_display_mode *adjusted_mode, int x, int y,
1870 struct drm_framebuffer *old_fb);</synopsis>
1872 Set a new mode, position and frame buffer. Depending on the device
1873 requirements, the mode can be stored internally by the driver and
1874 applied in the <methodname>commit</methodname> operation, or
1875 programmed to the hardware immediately.
1878 The <methodname>mode_set</methodname> operation returns 0 on success
1879 or a negative error code if an error occurs.
1883 <synopsis>void (*commit)(struct drm_crtc *crtc);</synopsis>
1885 Commit a mode. This operation is called after setting the new mode.
1886 Upon return the device must use the new mode and be fully
1893 <title>Encoder Helper Operations</title>
1896 <synopsis>bool (*mode_fixup)(struct drm_encoder *encoder,
1897 const struct drm_display_mode *mode,
1898 struct drm_display_mode *adjusted_mode);</synopsis>
1900 Let encoders adjust the requested mode or reject it completely. This
1901 operation returns true if the mode is accepted (possibly after being
1902 adjusted) or false if it is rejected. See the
1903 <link linkend="drm-helper-crtc-mode-fixup">mode_fixup CRTC helper
1904 operation</link> for an explanation of the allowed adjustments.
1908 <synopsis>void (*prepare)(struct drm_encoder *encoder);</synopsis>
1910 Prepare the encoder for mode setting. This operation is called after
1911 validating the requested mode. Drivers use it to perform
1912 device-specific operations required before setting the new mode.
1916 <synopsis>void (*mode_set)(struct drm_encoder *encoder,
1917 struct drm_display_mode *mode,
1918 struct drm_display_mode *adjusted_mode);</synopsis>
1920 Set a new mode. Depending on the device requirements, the mode can
1921 be stored internally by the driver and applied in the
1922 <methodname>commit</methodname> operation, or programmed to the
1923 hardware immediately.
1927 <synopsis>void (*commit)(struct drm_encoder *encoder);</synopsis>
1929 Commit a mode. This operation is called after setting the new mode.
1930 Upon return the device must use the new mode and be fully
1937 <title>Connector Helper Operations</title>
1940 <synopsis>struct drm_encoder *(*best_encoder)(struct drm_connector *connector);</synopsis>
1942 Return a pointer to the best encoder for the connecter. Device that
1943 map connectors to encoders 1:1 simply return the pointer to the
1944 associated encoder. This operation is mandatory.
1948 <synopsis>int (*get_modes)(struct drm_connector *connector);</synopsis>
1950 Fill the connector's <structfield>probed_modes</structfield> list
1951 by parsing EDID data with <function>drm_add_edid_modes</function> or
1952 calling <function>drm_mode_probed_add</function> directly for every
1953 supported mode and return the number of modes it has detected. This
1954 operation is mandatory.
1957 When adding modes manually the driver creates each mode with a call to
1958 <function>drm_mode_create</function> and must fill the following fields.
1961 <synopsis>__u32 type;</synopsis>
1963 Mode type bitmask, a combination of
1966 <term>DRM_MODE_TYPE_BUILTIN</term>
1967 <listitem><para>not used?</para></listitem>
1970 <term>DRM_MODE_TYPE_CLOCK_C</term>
1971 <listitem><para>not used?</para></listitem>
1974 <term>DRM_MODE_TYPE_CRTC_C</term>
1975 <listitem><para>not used?</para></listitem>
1979 DRM_MODE_TYPE_PREFERRED - The preferred mode for the connector
1982 <para>not used?</para>
1986 <term>DRM_MODE_TYPE_DEFAULT</term>
1987 <listitem><para>not used?</para></listitem>
1990 <term>DRM_MODE_TYPE_USERDEF</term>
1991 <listitem><para>not used?</para></listitem>
1994 <term>DRM_MODE_TYPE_DRIVER</term>
1997 The mode has been created by the driver (as opposed to
1998 to user-created modes).
2003 Drivers must set the DRM_MODE_TYPE_DRIVER bit for all modes they
2004 create, and set the DRM_MODE_TYPE_PREFERRED bit for the preferred
2009 <synopsis>__u32 clock;</synopsis>
2010 <para>Pixel clock frequency in kHz unit</para>
2013 <synopsis>__u16 hdisplay, hsync_start, hsync_end, htotal;
2014 __u16 vdisplay, vsync_start, vsync_end, vtotal;</synopsis>
2015 <para>Horizontal and vertical timing information</para>
2017 Active Front Sync Back
2019 <-----------------------><----------------><-------------><-------------->
2021 //////////////////////|
2022 ////////////////////// |
2023 ////////////////////// |.................. ................
2026 <----- [hv]display ----->
2027 <------------- [hv]sync_start ------------>
2028 <--------------------- [hv]sync_end --------------------->
2029 <-------------------------------- [hv]total ----------------------------->
2033 <synopsis>__u16 hskew;
2034 __u16 vscan;</synopsis>
2035 <para>Unknown</para>
2038 <synopsis>__u32 flags;</synopsis>
2040 Mode flags, a combination of
2043 <term>DRM_MODE_FLAG_PHSYNC</term>
2045 Horizontal sync is active high
2049 <term>DRM_MODE_FLAG_NHSYNC</term>
2051 Horizontal sync is active low
2055 <term>DRM_MODE_FLAG_PVSYNC</term>
2057 Vertical sync is active high
2061 <term>DRM_MODE_FLAG_NVSYNC</term>
2063 Vertical sync is active low
2067 <term>DRM_MODE_FLAG_INTERLACE</term>
2073 <term>DRM_MODE_FLAG_DBLSCAN</term>
2075 Mode uses doublescan
2079 <term>DRM_MODE_FLAG_CSYNC</term>
2081 Mode uses composite sync
2085 <term>DRM_MODE_FLAG_PCSYNC</term>
2087 Composite sync is active high
2091 <term>DRM_MODE_FLAG_NCSYNC</term>
2093 Composite sync is active low
2097 <term>DRM_MODE_FLAG_HSKEW</term>
2099 hskew provided (not used?)
2103 <term>DRM_MODE_FLAG_BCAST</term>
2109 <term>DRM_MODE_FLAG_PIXMUX</term>
2115 <term>DRM_MODE_FLAG_DBLCLK</term>
2121 <term>DRM_MODE_FLAG_CLKDIV2</term>
2129 Note that modes marked with the INTERLACE or DBLSCAN flags will be
2131 <function>drm_helper_probe_single_connector_modes</function> if
2132 the connector's <structfield>interlace_allowed</structfield> or
2133 <structfield>doublescan_allowed</structfield> field is set to 0.
2137 <synopsis>char name[DRM_DISPLAY_MODE_LEN];</synopsis>
2139 Mode name. The driver must call
2140 <function>drm_mode_set_name</function> to fill the mode name from
2141 <structfield>hdisplay</structfield>,
2142 <structfield>vdisplay</structfield> and interlace flag after
2143 filling the corresponding fields.
2149 The <structfield>vrefresh</structfield> value is computed by
2150 <function>drm_helper_probe_single_connector_modes</function>.
2153 When parsing EDID data, <function>drm_add_edid_modes</function> fill the
2154 connector <structfield>display_info</structfield>
2155 <structfield>width_mm</structfield> and
2156 <structfield>height_mm</structfield> fields. When creating modes
2157 manually the <methodname>get_modes</methodname> helper operation must
2158 set the <structfield>display_info</structfield>
2159 <structfield>width_mm</structfield> and
2160 <structfield>height_mm</structfield> fields if they haven't been set
2161 already (for instance at initialization time when a fixed-size panel is
2162 attached to the connector). The mode <structfield>width_mm</structfield>
2163 and <structfield>height_mm</structfield> fields are only used internally
2164 during EDID parsing and should not be set when creating modes manually.
2168 <synopsis>int (*mode_valid)(struct drm_connector *connector,
2169 struct drm_display_mode *mode);</synopsis>
2171 Verify whether a mode is valid for the connector. Return MODE_OK for
2172 supported modes and one of the enum drm_mode_status values (MODE_*)
2173 for unsupported modes. This operation is mandatory.
2176 As the mode rejection reason is currently not used beside for
2177 immediately removing the unsupported mode, an implementation can
2178 return MODE_BAD regardless of the exact reason why the mode is not
2182 Note that the <methodname>mode_valid</methodname> helper operation is
2183 only called for modes detected by the device, and
2184 <emphasis>not</emphasis> for modes set by the user through the CRTC
2185 <methodname>set_config</methodname> operation.
2191 <title>Modeset Helper Functions Reference</title>
2192 !Edrivers/gpu/drm/drm_crtc_helper.c
2195 <title>fbdev Helper Functions Reference</title>
2196 !Pdrivers/gpu/drm/drm_fb_helper.c fbdev helpers
2197 !Edrivers/gpu/drm/drm_fb_helper.c
2198 !Iinclude/drm/drm_fb_helper.h
2201 <title>Display Port Helper Functions Reference</title>
2202 !Pdrivers/gpu/drm/drm_dp_helper.c dp helpers
2203 !Iinclude/drm/drm_dp_helper.h
2204 !Edrivers/gpu/drm/drm_dp_helper.c
2207 <title>EDID Helper Functions Reference</title>
2208 !Edrivers/gpu/drm/drm_edid.c
2211 <title>Rectangle Utilities Reference</title>
2212 !Pinclude/drm/drm_rect.h rect utils
2213 !Iinclude/drm/drm_rect.h
2214 !Edrivers/gpu/drm/drm_rect.c
2217 <title>Flip-work Helper Reference</title>
2218 !Pinclude/drm/drm_flip_work.h flip utils
2219 !Iinclude/drm/drm_flip_work.h
2220 !Edrivers/gpu/drm/drm_flip_work.c
2223 <title>HDMI Infoframes Helper Reference</title>
2225 Strictly speaking this is not a DRM helper library but generally useable
2226 by any driver interfacing with HDMI outputs like v4l or alsa drivers.
2227 But it nicely fits into the overall topic of mode setting helper
2228 libraries and hence is also included here.
2230 !Iinclude/linux/hdmi.h
2231 !Edrivers/video/hdmi.c
2235 <!-- Internals: kms properties -->
2237 <sect1 id="drm-kms-properties">
2238 <title>KMS Properties</title>
2240 Drivers may need to expose additional parameters to applications than
2241 those described in the previous sections. KMS supports attaching
2242 properties to CRTCs, connectors and planes and offers a userspace API to
2243 list, get and set the property values.
2246 Properties are identified by a name that uniquely defines the property
2247 purpose, and store an associated value. For all property types except blob
2248 properties the value is a 64-bit unsigned integer.
2251 KMS differentiates between properties and property instances. Drivers
2252 first create properties and then create and associate individual instances
2253 of those properties to objects. A property can be instantiated multiple
2254 times and associated with different objects. Values are stored in property
2255 instances, and all other property information are stored in the propery
2256 and shared between all instances of the property.
2259 Every property is created with a type that influences how the KMS core
2260 handles the property. Supported property types are
2263 <term>DRM_MODE_PROP_RANGE</term>
2264 <listitem><para>Range properties report their minimum and maximum
2265 admissible values. The KMS core verifies that values set by
2266 application fit in that range.</para></listitem>
2269 <term>DRM_MODE_PROP_ENUM</term>
2270 <listitem><para>Enumerated properties take a numerical value that
2271 ranges from 0 to the number of enumerated values defined by the
2272 property minus one, and associate a free-formed string name to each
2273 value. Applications can retrieve the list of defined value-name pairs
2274 and use the numerical value to get and set property instance values.
2278 <term>DRM_MODE_PROP_BITMASK</term>
2279 <listitem><para>Bitmask properties are enumeration properties that
2280 additionally restrict all enumerated values to the 0..63 range.
2281 Bitmask property instance values combine one or more of the
2282 enumerated bits defined by the property.</para></listitem>
2285 <term>DRM_MODE_PROP_BLOB</term>
2286 <listitem><para>Blob properties store a binary blob without any format
2287 restriction. The binary blobs are created as KMS standalone objects,
2288 and blob property instance values store the ID of their associated
2290 <para>Blob properties are only used for the connector EDID property
2291 and cannot be created by drivers.</para></listitem>
2296 To create a property drivers call one of the following functions depending
2297 on the property type. All property creation functions take property flags
2298 and name, as well as type-specific arguments.
2301 <synopsis>struct drm_property *drm_property_create_range(struct drm_device *dev, int flags,
2303 uint64_t min, uint64_t max);</synopsis>
2304 <para>Create a range property with the given minimum and maximum
2308 <synopsis>struct drm_property *drm_property_create_enum(struct drm_device *dev, int flags,
2310 const struct drm_prop_enum_list *props,
2311 int num_values);</synopsis>
2312 <para>Create an enumerated property. The <parameter>props</parameter>
2313 argument points to an array of <parameter>num_values</parameter>
2314 value-name pairs.</para>
2317 <synopsis>struct drm_property *drm_property_create_bitmask(struct drm_device *dev,
2318 int flags, const char *name,
2319 const struct drm_prop_enum_list *props,
2320 int num_values);</synopsis>
2321 <para>Create a bitmask property. The <parameter>props</parameter>
2322 argument points to an array of <parameter>num_values</parameter>
2323 value-name pairs.</para>
2328 Properties can additionally be created as immutable, in which case they
2329 will be read-only for applications but can be modified by the driver. To
2330 create an immutable property drivers must set the DRM_MODE_PROP_IMMUTABLE
2331 flag at property creation time.
2334 When no array of value-name pairs is readily available at property
2335 creation time for enumerated or range properties, drivers can create
2336 the property using the <function>drm_property_create</function> function
2337 and manually add enumeration value-name pairs by calling the
2338 <function>drm_property_add_enum</function> function. Care must be taken to
2339 properly specify the property type through the <parameter>flags</parameter>
2343 After creating properties drivers can attach property instances to CRTC,
2344 connector and plane objects by calling the
2345 <function>drm_object_attach_property</function>. The function takes a
2346 pointer to the target object, a pointer to the previously created property
2347 and an initial instance value.
2351 <!-- Internals: vertical blanking -->
2353 <sect1 id="drm-vertical-blank">
2354 <title>Vertical Blanking</title>
2356 Vertical blanking plays a major role in graphics rendering. To achieve
2357 tear-free display, users must synchronize page flips and/or rendering to
2358 vertical blanking. The DRM API offers ioctls to perform page flips
2359 synchronized to vertical blanking and wait for vertical blanking.
2362 The DRM core handles most of the vertical blanking management logic, which
2363 involves filtering out spurious interrupts, keeping race-free blanking
2364 counters, coping with counter wrap-around and resets and keeping use
2365 counts. It relies on the driver to generate vertical blanking interrupts
2366 and optionally provide a hardware vertical blanking counter. Drivers must
2367 implement the following operations.
2371 <synopsis>int (*enable_vblank) (struct drm_device *dev, int crtc);
2372 void (*disable_vblank) (struct drm_device *dev, int crtc);</synopsis>
2374 Enable or disable vertical blanking interrupts for the given CRTC.
2378 <synopsis>u32 (*get_vblank_counter) (struct drm_device *dev, int crtc);</synopsis>
2380 Retrieve the value of the vertical blanking counter for the given
2381 CRTC. If the hardware maintains a vertical blanking counter its value
2382 should be returned. Otherwise drivers can use the
2383 <function>drm_vblank_count</function> helper function to handle this
2389 Drivers must initialize the vertical blanking handling core with a call to
2390 <function>drm_vblank_init</function> in their
2391 <methodname>load</methodname> operation. The function will set the struct
2392 <structname>drm_device</structname>
2393 <structfield>vblank_disable_allowed</structfield> field to 0. This will
2394 keep vertical blanking interrupts enabled permanently until the first mode
2395 set operation, where <structfield>vblank_disable_allowed</structfield> is
2396 set to 1. The reason behind this is not clear. Drivers can set the field
2397 to 1 after <function>calling drm_vblank_init</function> to make vertical
2398 blanking interrupts dynamically managed from the beginning.
2401 Vertical blanking interrupts can be enabled by the DRM core or by drivers
2402 themselves (for instance to handle page flipping operations). The DRM core
2403 maintains a vertical blanking use count to ensure that the interrupts are
2404 not disabled while a user still needs them. To increment the use count,
2405 drivers call <function>drm_vblank_get</function>. Upon return vertical
2406 blanking interrupts are guaranteed to be enabled.
2409 To decrement the use count drivers call
2410 <function>drm_vblank_put</function>. Only when the use count drops to zero
2411 will the DRM core disable the vertical blanking interrupts after a delay
2412 by scheduling a timer. The delay is accessible through the vblankoffdelay
2413 module parameter or the <varname>drm_vblank_offdelay</varname> global
2414 variable and expressed in milliseconds. Its default value is 5000 ms.
2417 When a vertical blanking interrupt occurs drivers only need to call the
2418 <function>drm_handle_vblank</function> function to account for the
2422 Resources allocated by <function>drm_vblank_init</function> must be freed
2423 with a call to <function>drm_vblank_cleanup</function> in the driver
2424 <methodname>unload</methodname> operation handler.
2428 <!-- Internals: open/close, file operations and ioctls -->
2431 <title>Open/Close, File Operations and IOCTLs</title>
2433 <title>Open and Close</title>
2434 <synopsis>int (*firstopen) (struct drm_device *);
2435 void (*lastclose) (struct drm_device *);
2436 int (*open) (struct drm_device *, struct drm_file *);
2437 void (*preclose) (struct drm_device *, struct drm_file *);
2438 void (*postclose) (struct drm_device *, struct drm_file *);</synopsis>
2439 <abstract>Open and close handlers. None of those methods are mandatory.
2442 The <methodname>firstopen</methodname> method is called by the DRM core
2443 for legacy UMS (User Mode Setting) drivers only when an application
2444 opens a device that has no other opened file handle. UMS drivers can
2445 implement it to acquire device resources. KMS drivers can't use the
2446 method and must acquire resources in the <methodname>load</methodname>
2450 Similarly the <methodname>lastclose</methodname> method is called when
2451 the last application holding a file handle opened on the device closes
2452 it, for both UMS and KMS drivers. Additionally, the method is also
2453 called at module unload time or, for hot-pluggable devices, when the
2454 device is unplugged. The <methodname>firstopen</methodname> and
2455 <methodname>lastclose</methodname> calls can thus be unbalanced.
2458 The <methodname>open</methodname> method is called every time the device
2459 is opened by an application. Drivers can allocate per-file private data
2460 in this method and store them in the struct
2461 <structname>drm_file</structname> <structfield>driver_priv</structfield>
2462 field. Note that the <methodname>open</methodname> method is called
2463 before <methodname>firstopen</methodname>.
2466 The close operation is split into <methodname>preclose</methodname> and
2467 <methodname>postclose</methodname> methods. Drivers must stop and
2468 cleanup all per-file operations in the <methodname>preclose</methodname>
2469 method. For instance pending vertical blanking and page flip events must
2470 be cancelled. No per-file operation is allowed on the file handle after
2471 returning from the <methodname>preclose</methodname> method.
2474 Finally the <methodname>postclose</methodname> method is called as the
2475 last step of the close operation, right before calling the
2476 <methodname>lastclose</methodname> method if no other open file handle
2477 exists for the device. Drivers that have allocated per-file private data
2478 in the <methodname>open</methodname> method should free it here.
2481 The <methodname>lastclose</methodname> method should restore CRTC and
2482 plane properties to default value, so that a subsequent open of the
2483 device will not inherit state from the previous user. It can also be
2484 used to execute delayed power switching state changes, e.g. in
2485 conjunction with the vga-switcheroo infrastructure. Beyond that KMS
2486 drivers should not do any further cleanup. Only legacy UMS drivers might
2487 need to clean up device state so that the vga console or an independent
2488 fbdev driver could take over.
2492 <title>File Operations</title>
2493 <synopsis>const struct file_operations *fops</synopsis>
2494 <abstract>File operations for the DRM device node.</abstract>
2496 Drivers must define the file operations structure that forms the DRM
2497 userspace API entry point, even though most of those operations are
2498 implemented in the DRM core. The <methodname>open</methodname>,
2499 <methodname>release</methodname> and <methodname>ioctl</methodname>
2500 operations are handled by
2502 .owner = THIS_MODULE,
2504 .release = drm_release,
2505 .unlocked_ioctl = drm_ioctl,
2506 #ifdef CONFIG_COMPAT
2507 .compat_ioctl = drm_compat_ioctl,
2512 Drivers that implement private ioctls that requires 32/64bit
2513 compatibility support must provide their own
2514 <methodname>compat_ioctl</methodname> handler that processes private
2515 ioctls and calls <function>drm_compat_ioctl</function> for core ioctls.
2518 The <methodname>read</methodname> and <methodname>poll</methodname>
2519 operations provide support for reading DRM events and polling them. They
2524 .llseek = no_llseek,
2528 The memory mapping implementation varies depending on how the driver
2529 manages memory. Pre-GEM drivers will use <function>drm_mmap</function>,
2530 while GEM-aware drivers will use <function>drm_gem_mmap</function>. See
2531 <xref linkend="drm-gem"/>.
2533 .mmap = drm_gem_mmap,
2537 No other file operation is supported by the DRM API.
2541 <title>IOCTLs</title>
2542 <synopsis>struct drm_ioctl_desc *ioctls;
2543 int num_ioctls;</synopsis>
2544 <abstract>Driver-specific ioctls descriptors table.</abstract>
2546 Driver-specific ioctls numbers start at DRM_COMMAND_BASE. The ioctls
2547 descriptors table is indexed by the ioctl number offset from the base
2548 value. Drivers can use the DRM_IOCTL_DEF_DRV() macro to initialize the
2552 <programlisting>DRM_IOCTL_DEF_DRV(ioctl, func, flags)</programlisting>
2554 <parameter>ioctl</parameter> is the ioctl name. Drivers must define
2555 the DRM_##ioctl and DRM_IOCTL_##ioctl macros to the ioctl number
2556 offset from DRM_COMMAND_BASE and the ioctl number respectively. The
2557 first macro is private to the device while the second must be exposed
2558 to userspace in a public header.
2561 <parameter>func</parameter> is a pointer to the ioctl handler function
2562 compatible with the <type>drm_ioctl_t</type> type.
2563 <programlisting>typedef int drm_ioctl_t(struct drm_device *dev, void *data,
2564 struct drm_file *file_priv);</programlisting>
2567 <parameter>flags</parameter> is a bitmask combination of the following
2568 values. It restricts how the ioctl is allowed to be called.
2571 DRM_AUTH - Only authenticated callers allowed
2574 DRM_MASTER - The ioctl can only be called on the master file
2578 DRM_ROOT_ONLY - Only callers with the SYSADMIN capability allowed
2581 DRM_CONTROL_ALLOW - The ioctl can only be called on a control
2585 DRM_UNLOCKED - The ioctl handler will be called without locking
2586 the DRM global mutex
2594 <title>Legacy Support Code</title>
2596 The section very brievely covers some of the old legacy support code which
2597 is only used by old DRM drivers which have done a so-called shadow-attach
2598 to the underlying device instead of registering as a real driver. This
2599 also includes some of the old generic buffer mangement and command
2600 submission code. Do not use any of this in new and modern drivers.
2604 <title>Legacy Suspend/Resume</title>
2606 The DRM core provides some suspend/resume code, but drivers wanting full
2607 suspend/resume support should provide save() and restore() functions.
2608 These are called at suspend, hibernate, or resume time, and should perform
2609 any state save or restore required by your device across suspend or
2612 <synopsis>int (*suspend) (struct drm_device *, pm_message_t state);
2613 int (*resume) (struct drm_device *);</synopsis>
2615 Those are legacy suspend and resume methods which
2616 <emphasis>only</emphasis> work with the legacy shadow-attach driver
2617 registration functions. New driver should use the power management
2618 interface provided by their bus type (usually through
2619 the struct <structname>device_driver</structname> dev_pm_ops) and set
2620 these methods to NULL.
2625 <title>Legacy DMA Services</title>
2627 This should cover how DMA mapping etc. is supported by the core.
2628 These functions are deprecated and should not be used.
2637 - Document the struct_mutex catch-all lock
2638 - Document connector properties
2640 - Why is the load method optional?
2641 - What are drivers supposed to set the initial display state to, and how?
2642 Connector's DPMS states are not initialized and are thus equal to
2643 DRM_MODE_DPMS_ON. The fbcon compatibility layer calls
2644 drm_helper_disable_unused_functions(), which disables unused encoders and
2645 CRTCs, but doesn't touch the connectors' DPMS state, and
2646 drm_helper_connector_dpms() in reaction to fbdev blanking events. Do drivers
2647 that don't implement (or just don't use) fbcon compatibility need to call
2648 those functions themselves?
2649 - KMS drivers must call drm_vblank_pre_modeset() and drm_vblank_post_modeset()
2650 around mode setting. Should this be done in the DRM core?
2651 - vblank_disable_allowed is set to 1 in the first drm_vblank_post_modeset()
2652 call and never set back to 0. It seems to be safe to permanently set it to 1
2653 in drm_vblank_init() for KMS driver, and it might be safe for UMS drivers as
2654 well. This should be investigated.
2655 - crtc and connector .save and .restore operations are only used internally in
2656 drivers, should they be removed from the core?
2657 - encoder mid-layer .save and .restore operations are only used internally in
2658 drivers, should they be removed from the core?
2659 - encoder mid-layer .detect operation is only used internally in drivers,
2660 should it be removed from the core?
2663 <!-- External interfaces -->
2665 <chapter id="drmExternals">
2666 <title>Userland interfaces</title>
2668 The DRM core exports several interfaces to applications,
2669 generally intended to be used through corresponding libdrm
2670 wrapper functions. In addition, drivers export device-specific
2671 interfaces for use by userspace drivers & device-aware
2672 applications through ioctls and sysfs files.
2675 External interfaces include: memory mapping, context management,
2676 DMA operations, AGP management, vblank control, fence
2677 management, memory management, and output management.
2680 Cover generic ioctls and sysfs layout here. We only need high-level
2681 info, since man pages should cover the rest.
2684 <!-- External: render nodes -->
2687 <title>Render nodes</title>
2689 DRM core provides multiple character-devices for user-space to use.
2690 Depending on which device is opened, user-space can perform a different
2691 set of operations (mainly ioctls). The primary node is always created
2692 and called card<num>. Additionally, a currently
2693 unused control node, called controlD<num> is also
2694 created. The primary node provides all legacy operations and
2695 historically was the only interface used by userspace. With KMS, the
2696 control node was introduced. However, the planned KMS control interface
2697 has never been written and so the control node stays unused to date.
2700 With the increased use of offscreen renderers and GPGPU applications,
2701 clients no longer require running compositors or graphics servers to
2702 make use of a GPU. But the DRM API required unprivileged clients to
2703 authenticate to a DRM-Master prior to getting GPU access. To avoid this
2704 step and to grant clients GPU access without authenticating, render
2705 nodes were introduced. Render nodes solely serve render clients, that
2706 is, no modesetting or privileged ioctls can be issued on render nodes.
2707 Only non-global rendering commands are allowed. If a driver supports
2708 render nodes, it must advertise it via the DRIVER_RENDER
2709 DRM driver capability. If not supported, the primary node must be used
2710 for render clients together with the legacy drmAuth authentication
2714 If a driver advertises render node support, DRM core will create a
2715 separate render node called renderD<num>. There will
2716 be one render node per device. No ioctls except PRIME-related ioctls
2717 will be allowed on this node. Especially GEM_OPEN will be
2718 explicitly prohibited. Render nodes are designed to avoid the
2719 buffer-leaks, which occur if clients guess the flink names or mmap
2720 offsets on the legacy interface. Additionally to this basic interface,
2721 drivers must mark their driver-dependent render-only ioctls as
2722 DRM_RENDER_ALLOW so render clients can use them. Driver
2723 authors must be careful not to allow any privileged ioctls on render
2727 With render nodes, user-space can now control access to the render node
2728 via basic file-system access-modes. A running graphics server which
2729 authenticates clients on the privileged primary/legacy node is no longer
2730 required. Instead, a client can open the render node and is immediately
2731 granted GPU access. Communication between clients (or servers) is done
2732 via PRIME. FLINK from render node to legacy node is not supported. New
2733 clients must not use the insecure FLINK interface.
2736 Besides dropping all modeset/global ioctls, render nodes also drop the
2737 DRM-Master concept. There is no reason to associate render clients with
2738 a DRM-Master as they are independent of any graphics server. Besides,
2739 they must work without any running master, anyway.
2740 Drivers must be able to run without a master object if they support
2741 render nodes. If, on the other hand, a driver requires shared state
2742 between clients which is visible to user-space and accessible beyond
2743 open-file boundaries, they cannot support render nodes.
2747 <!-- External: vblank handling -->
2750 <title>VBlank event handling</title>
2752 The DRM core exposes two vertical blank related ioctls:
2755 <term>DRM_IOCTL_WAIT_VBLANK</term>
2758 This takes a struct drm_wait_vblank structure as its argument,
2759 and it is used to block or request a signal when a specified
2760 vblank event occurs.
2765 <term>DRM_IOCTL_MODESET_CTL</term>
2768 This should be called by application level drivers before and
2769 after mode setting, since on many devices the vertical blank
2770 counter is reset at that time. Internally, the DRM snapshots
2771 the last vblank count when the ioctl is called with the
2772 _DRM_PRE_MODESET command, so that the counter won't go backwards
2773 (which is dealt with when _DRM_POST_MODESET is used).
2778 <!--!Edrivers/char/drm/drm_irq.c-->
2784 <part id="drmDrivers">
2785 <title>DRM Drivers</title>
2789 This second part of the DRM Developer's Guide documents driver code,
2790 implementation details and also all the driver-specific userspace
2791 interfaces. Especially since all hardware-acceleration interfaces to
2792 userspace are driver specific for efficiency and other reasons these
2793 interfaces can be rather substantial. Hence every driver has its own
2798 <chapter id="drmI915">
2799 <title>drm/i915 Intel GFX Driver</title>
2801 The drm/i915 driver supports all (with the exception of some very early
2802 models) integrated GFX chipsets with both Intel display and rendering
2803 blocks. This excludes a set of SoC platforms with an SGX rendering unit,
2804 those have basic support through the gma500 drm driver.
2807 <title>Display Hardware Handling</title>
2809 This section covers everything related to the display hardware including
2810 the mode setting infrastructure, plane, sprite and cursor handling and
2811 display, output probing and related topics.
2814 <title>Mode Setting Infrastructure</title>
2816 The i915 driver is thus far the only DRM driver which doesn't use the
2817 common DRM helper code to implement mode setting sequences. Thus it
2818 has its own tailor-made infrastructure for executing a display
2819 configuration change.
2823 <title>Plane Configuration</title>
2825 This section covers plane configuration and composition with the
2826 primary plane, sprites, cursors and overlays. This includes the
2827 infrastructure to do atomic vsync'ed updates of all this state and
2828 also tightly coupled topics like watermark setup and computation,
2829 framebuffer compression and panel self refresh.
2833 <title>Output Probing</title>
2835 This section covers output probing and related infrastructure like the
2836 hotplug interrupt storm detection and mitigation code. Note that the
2837 i915 driver still uses most of the common DRM helper code for output
2838 probing, so those sections fully apply.
2844 <title>Memory Management and Command Submission</title>
2846 This sections covers all things related to the GEM implementation in the
projects / firefly-linux-kernel-4.4.55.git / blob