From patchwork Thu Aug 20 13:47:49 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: Kieran Bingham X-Patchwork-Id: 9338 Return-Path: X-Original-To: parsemail@patchwork.libcamera.org Delivered-To: parsemail@patchwork.libcamera.org Received: from lancelot.ideasonboard.com (lancelot.ideasonboard.com [92.243.16.209]) by patchwork.libcamera.org (Postfix) with ESMTPS id C04C4BE173 for ; Thu, 20 Aug 2020 13:47:59 +0000 (UTC) Received: from lancelot.ideasonboard.com (localhost [IPv6:::1]) by lancelot.ideasonboard.com (Postfix) with ESMTP id 791A961FE7; Thu, 20 Aug 2020 15:47:59 +0200 (CEST) Authentication-Results: lancelot.ideasonboard.com; dkim=fail reason="signature verification failed" (1024-bit key; unprotected) header.d=ideasonboard.com header.i=@ideasonboard.com header.b="AXm5BLm+"; dkim-atps=neutral Received: from perceval.ideasonboard.com (perceval.ideasonboard.com [IPv6:2001:4b98:dc2:55:216:3eff:fef7:d647]) by lancelot.ideasonboard.com (Postfix) with ESMTPS id 4F49760381 for ; Thu, 20 Aug 2020 15:47:57 +0200 (CEST) Received: from Q.local (cpc89244-aztw30-2-0-cust3082.18-1.cable.virginm.net [86.31.172.11]) by perceval.ideasonboard.com (Postfix) with ESMTPSA id B17C9329; Thu, 20 Aug 2020 15:47:56 +0200 (CEST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=ideasonboard.com; s=mail; t=1597931277; bh=chTvLguGK5lwpgNpV5CZqy8olyZPydGled1PQlfSzPY=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=AXm5BLm++Tqf1bOtFxO2UqFqc09hF5ThYwHdK1pdTWak9W58bGzY5Y51eECHvZbbb PA9rv7srNG93INg831NN3vhyRFdBlzJEfg3yRQ/llUXi21kqivoU4Y+OCdfk9BQj8R 2QnkY5nmq44lB145hLE/ibBgBvQ2tI7ZE+JFb/bM= From: Kieran Bingham To: libcamera devel , Chris Ward , Jacopo Mondi Date: Thu, 20 Aug 2020 14:47:49 +0100 Message-Id: <20200820134751.278033-2-kieran.bingham@ideasonboard.com> X-Mailer: git-send-email 2.25.1 In-Reply-To: <20200820134751.278033-1-kieran.bingham@ideasonboard.com> References: <20200820134751.278033-1-kieran.bingham@ideasonboard.com> MIME-Version: 1.0 Subject: [libcamera-devel] [PATCH v4 1/3] Documentation: Guides: Developers Guide to Libcamera X-BeenThere: libcamera-devel@lists.libcamera.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: libcamera-devel-bounces@lists.libcamera.org Sender: "libcamera-devel" From: Chris Chinchilla Create an introduction and overview for new developers to libcamera. Provide an overview of the Camera Stack, and Architecture of libcamera and introduce the main concepts of libcamera. Signed-off-by: Chris Chinchilla [Kieran: Rework/Reflow, add diagrams, licensing] Signed-off-by: Kieran Bingham Acked-by: Laurent Pinchart --- Documentation/guides/introduction.rst | 319 ++++++++++++++++++++++++++ Documentation/index.rst | 2 + Documentation/meson.build | 1 + 3 files changed, 322 insertions(+) create mode 100644 Documentation/guides/introduction.rst diff --git a/Documentation/guides/introduction.rst b/Documentation/guides/introduction.rst new file mode 100644 index 000000000000..f34d2cf7cbdc --- /dev/null +++ b/Documentation/guides/introduction.rst @@ -0,0 +1,319 @@ +.. SPDX-License-Identifier: CC-BY-SA-4.0 + +Developers guide to libcamera +============================= + +The Linux kernel handles multimedia devices through the 'Linux media' subsystem +and provides a set of APIs (application programming interfaces) known +collectively as V4L2 (`Video for Linux 2`_) and the `Media Controller`_ API +which provide an interface to interact and control media devices. + +Included in this subsystem are drivers for camera sensors, CSI2 (Camera +Serial Interface) recievers, and ISPs (Image Signal Processors) + +The usage of these drivers to provide a functioning camera stack is a +responsibility that lies in userspace which is commonly implemented separately +by vendors without a common architecture or API for application developers. + +libcamera provides a complete camera stack for Linux based systems to abstract +functionality desired by camera application developers and process the +configuration of hardware and image control algorithms requried to obtain +desireable results from the camera. + +.. _Video for Linux 2: https://www.linuxtv.org/downloads/v4l-dvb-apis-new/userspace-api/v4l/v4l2.html +.. _Media Controller: https://www.linuxtv.org/downloads/v4l-dvb-apis-new/userspace-api/mediactl/media-controller.html + + +In this developers guide, we will explore the `Camera Stack`_ and how it is +can be visualised at a high level, and explore the internal `Architecture`_ of +the libcamera library with its components. The current `Platform Support`_ is +detailed, as well as an overview of the `Licensing`_ requirements of the +project. + +This introduction is followed by a walkthrough tutorial to newcomers wishing to +support a new platform with the `Pipeline Handler Writers Guide`_ and for those +looking to make use of the libcamera native API an `Application Writers Guide`_ +provides a tutorial of the key APIs exposed by libcamera. + +.. _Pipeline Handler Writers Guide: pipeline-handler.html +.. _Application Writers Guide: application-developer.html + +.. TODO: Correctly link to the other articles of the guide + +Camera Stack +------------ + +The libcamera library is implemented in userspace, and makes use of underlying +kernel drivers that directly interact with hardware. + +Applications can make use of libcamera through the native `libcamera API`_'s or +through an adaptation layer integrating libcamera into a larger framework. + +.. _libcamera API: https://www.libcamera.org/api-html/index.html + +:: + + Application Layer + / +--------------+ +--------------+ +--------------+ +--------------+ + | | Native | | Framework | | Native | | Android | + | | V4L2 | | Application | | libcamera | | Camera | + | | Application | | (gstreamer) | | Application | | Framework | + \ +--------------+ +--------------+ +--------------+ +--------------+ + + ^ ^ ^ ^ + | | | | + | | | | + v v | v + Adaptation Layer | + / +--------------+ +--------------+ | +--------------+ + | | V4L2 | | gstreamer | | | Android | + | | Compatability| | element | | | Camera | + | | (preload) | |(libcamerasrc)| | | HAL | + \ +--------------+ +--------------+ | +--------------+ + | + ^ ^ | ^ + | | | | + | | | | + v v v v + libcamera Framework + / +--------------------------------------------------------------------+ + | | | + | | libcamera | + | | | + \ +--------------------------------------------------------------------+ + + ^ ^ ^ + Userspace | | | + --------------------- | ---------------- | ---------------- | --------------- + Kernel | | | + v v v + + +-----------+ +-----------+ +-----------+ + | Media | <--> | Video | <--> | V4L2 | + | Device | | Device | | Subdev | + +-----------+ +-----------+ +-----------+ + +The camera stack comprises of four software layers. From bottom to top: + +* The kernel drivers control the camera hardware and expose a low-level + interface to userspace through the Linux kernel V4L2 family of APIs + (Media Controller API, V4L2 Video Device API and V4L2 Subdev API). + +* The libcamera framework is the core part of the stack. It handles all control + of the camera devices in its core component, libcamera, and exposes a native + C++ API to upper layers. + +* The libcamera adaptation layer is an umbrella term designating the components + that interface to libcamera in other frameworks. Notable examples are the V4L2 + compatibility layer, the gstreamer libcamera element, and the Android camera + HAL implementation based on libcamera which are provided as a part of the + libcamera project. + +* The applications and upper level frameworks are based on the libcamera + framework or libcamera adaptation, and are outside of the scope of the + libcamera project, however example native applications (cam, qcam) are + provided for testing. + + +V4L2 Compatibility Layer + V4L2 compatibility is achieved through a shared library that traps all + accesses to camera devices and routes them to libcamera to emulate high-level + V4L2 camera devices. It is injected in a process address space through + `LD_PRELOAD` and is completely transparent for applications. + + The compatibility layer exposes camera device features on a best-effort basis, + and aims for the level of features traditionally available from a UVC camera + designed for video conferencing. + +Android Camera HAL + Camera support for Android is achieved through a generic Android camera HAL + implementation on top of libcamera. The HAL implements features required by + Android and out of scope from libcamera, such as JPEG encoding support. + + This component is used to provide support for ChromeOS platforms + +GStreamer element (gstlibcamerasrc) + A `GStreamer element`_ is provided to allow capture from libcamera supported + devices through GStreamer pipelines, and connect to other elements for further + processing. + + Development of this element is ongoing and is limited to a single stream. + +Native libcamera API + Applications can make use of the libcamera API directly using the C++ + API. An example application and walkthrough using the libcamera API can be + followed in the `Application Writers Guide`_ + +.. _GStreamer element: https://gstreamer.freedesktop.org/documentation/application-development/basics/elements.html + +Architecture +------------ + +While offering a unified API towards upper layers, and presenting itself as a +single library, libcamera isn’t monolithic. It exposes multiple components +through its public API and is built around a set of separate helpers internally. +Hardware abstractractions are handled through the use of device-specific +components where required and dynamically loadable plugins are used to separate +image processing algorithms from the core libcamera codebase. + +:: + + --------------------------< libcamera Public API >--------------------------- + ^ ^ + | | + v v + +-------------+ +---------------------------------------------------+ + | Camera | | Camera Device | + | Manager | | +-----------------------------------------------+ | + +-------------+ | | Device-Agnostic | | + ^ | | | | + | | | +--------------------------+ | + | | | | ~~~~~~~~~~~~~~~~~~~~~~~ | + | | | | { +-----------------+ } | + | | | | } | //// Image //// | { | + | | | | <-> | / Processing // | } | + | | | | } | / Algorithms // | { | + | | | | { +-----------------+ } | + | | | | ~~~~~~~~~~~~~~~~~~~~~~~ | + | | | | ========================== | + | | | | +-----------------+ | + | | | | | // Pipeline /// | | + | | | | <-> | /// Handler /// | | + | | | | | /////////////// | | + | | +--------------------+ +-----------------+ | + | | Device-Specific | + | +---------------------------------------------------+ + | ^ ^ + | | | + v v v + +--------------------------------------------------------------------+ + | Helpers and Support Classes | + | +-------------+ +-------------+ +-------------+ +-------------+ | + | | MC & V4L2 | | Buffers | | Sandboxing | | Plugins | | + | | Support | | Allocator | | IPC | | Manager | | + | +-------------+ +-------------+ +-------------+ +-------------+ | + | +-------------+ +-------------+ | + | | Pipeline | | ... | | + | | Runner | | | | + | +-------------+ +-------------+ | + +--------------------------------------------------------------------+ + + /// Device-Specific Components + ~~~ Sandboxing + + +Camera Manager + The Camera Manager enumerates cameras and instantiates Pipeline Handlers to + manage each Camera that libcamera supports. The Camera Manager supports + hotplug detection and notification events when supported by the underlying + kernel devices. + + There is only ever one instance of the Camera Manager running per application. + Each application's instance of the Camera Manager ensures that only a single + application can take control of a camera device at once. + + Read the `Camera Manager API`_ documentation for more details. + +.. _Camera Manager API: http://libcamera.org/api-html/classlibcamera_1_1CameraManager.html + +Camera Device + The Camera class represents a single item of camera hardware that is capable + of producing one or more image streams, and provides the API to interact with + the underlying device. + + If a system has multiple instances of the same hardware attached, each has it's + own instance of the camera class. + + The API exposes full control of the device to upper layers of libcamera through + the public API, making it the highest level object libcamera exposes, and the + object that all other API operations interact with from configuration to + capture. + + Read the `Camera API`_ documentation for more details. + +.. _Camera API: http://libcamera.org/api-html/classlibcamera_1_1Camera.html + +Pipeline Handler + The Pipeline Handler manages the complex pipelines exposed by the kernel + drivers through the Media Controller and V4L2 APIs. It abstracts pipeline + handling to hide device-specific details from the rest of the library, and + implements both pipeline configuration based on stream configuration, and + pipeline runtime execution and scheduling when needed by the device. + + The Pipeline Handler lives in the same process as the rest of the library, and + has access to all helpers and kernel camera-related devices. + + Hardware abstraction is handled by device specific Pipeline Handlers which are + derived from the Pipeline Handler base class allowing commonality to be shared + among the implementations. + + Derived pipeline handlers create Camera device instances based on the devices + they detect and support on the running system, and are responsible for + managing the interactions with a camera device. + + More details can be found in the `PipelineHandler API`_ documentation, and the + `Pipeline Handler Writers Guide`_. + +.. _PipelineHandler API: http://libcamera.org/api-html/classlibcamera_1_1PipelineHandler.html + +Image Processing Algorithms + An image processing algorithm (IPA) component is a loadable plugin that + implements 3A (Auto-Exposure, Auto-White Balance, and Auto-Focus) and other + algorithms. + + The algorithms run on the CPU and interact with the camera devices through the + Pipeline Handler to control hardware image processing based on the parameters + supplied by upper layers, maintaining state and closing the control loop + of the ISP. + + The component is sandboxed and can only interact with libcamera through the + API provided by the Pipeline Handler and an IPA has no direct access to kernel + camera devices. + + Open source IPA modules built with libcamera can be run in the same process + space as libcamera, however external IPA modules are run in a separate process + from the main libcamera process. IPA modules have a restricted view of the + system, including no access to networking APIs and limited access to file + systems. + + IPA modules are only required for platforms and devices with an ISP controlled + by the host CPU. Camera sensors which have an integrated ISP are not + controlled through the IPA module. + +Platform Support +---------------- + +The library currently supports the following hardware platforms specifically +with dedicated pipeline handlers: + + - Intel IPU3 (ipu3) + - Rockchip RK3399 (rkisp1) + - RaspberryPi 3 and 4 (raspberrypi) + +Furthermore, generic platform support is provided for the following: + + - USB video device class cameras (uvcvideo) + - iMX7, Allwinner Sun6i (simple) + - Virtual media controller driver for test use cases (vimc) + +Licensing +--------- + +The libcamera core, is covered by the `LGPL-2.1-or-later`_ license. Pipeline +Handlers are a part of the libcamera code base and need to be contributed +upstream by device vendors. IPA modules included in libcamera are covered by a +free software license, however third-parties may develop IPA modules outside of +libcamera and distribute them under a closed-source license, provided they do +not include source code from the libcamera project. + +The libcamera project itself contains multiple libraries, applications and +utilities. Licenses are expressed through SPDX tags in text-based files that +support comments, and through the .reuse/dep5 file otherwise. A copy of all +licenses are stored in the LICENSES directory, and a full summary of the +licensing used throughout the project can be found in the COPYING.rst document. + +Applications which link dynamically against libcamera and use only the public +API are an independent work of the authors and have no license restrictions +imposed upon them from libcamera. + +.. _LGPL-2.1-or-later: https://spdx.org/licenses/LGPL-2.1-or-later.html \ No newline at end of file diff --git a/Documentation/index.rst b/Documentation/index.rst index 4e746bb17c4a..cfcfd388699b 100644 --- a/Documentation/index.rst +++ b/Documentation/index.rst @@ -12,3 +12,5 @@ Home Docs Contribute + + Developer Guide diff --git a/Documentation/meson.build b/Documentation/meson.build index 6d9a397cf1a3..dd7ae700af11 100644 --- a/Documentation/meson.build +++ b/Documentation/meson.build @@ -52,6 +52,7 @@ if sphinx.found() 'contributing.rst', 'docs.rst', 'index.rst', + 'guides/introduction.rst', ] release = 'release=v' + libcamera_git_version From patchwork Thu Aug 20 13:47:50 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: Kieran Bingham X-Patchwork-Id: 9339 Return-Path: X-Original-To: parsemail@patchwork.libcamera.org Delivered-To: parsemail@patchwork.libcamera.org Received: from lancelot.ideasonboard.com (lancelot.ideasonboard.com [92.243.16.209]) by patchwork.libcamera.org (Postfix) with ESMTPS id 5D204BE173 for ; Thu, 20 Aug 2020 13:48:01 +0000 (UTC) Received: from lancelot.ideasonboard.com (localhost [IPv6:::1]) by lancelot.ideasonboard.com (Postfix) with ESMTP id 2954162082; Thu, 20 Aug 2020 15:48:01 +0200 (CEST) Authentication-Results: lancelot.ideasonboard.com; dkim=fail reason="signature verification failed" (1024-bit key; unprotected) header.d=ideasonboard.com header.i=@ideasonboard.com header.b="id96eGC7"; dkim-atps=neutral Received: from perceval.ideasonboard.com (perceval.ideasonboard.com [213.167.242.64]) by lancelot.ideasonboard.com (Postfix) with ESMTPS id AA04260381 for ; Thu, 20 Aug 2020 15:47:57 +0200 (CEST) Received: from Q.local (cpc89244-aztw30-2-0-cust3082.18-1.cable.virginm.net [86.31.172.11]) by perceval.ideasonboard.com (Postfix) with ESMTPSA id 291C0A17; Thu, 20 Aug 2020 15:47:57 +0200 (CEST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=ideasonboard.com; s=mail; t=1597931277; bh=GWvDPfPUk+SUcze+ekmyZ9u+5COtFbSDxJz9Hn7HBmg=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=id96eGC7/uDeTkcGaAeKa3jkp+AfrKR5nDbtcikgIE2Zabv1CgiJ8e4nIBdHSdFCq 5N6fgDVO2S+QzzkMGszEra5QPx/1Ii2PFJGLiv7lB2ce9fd2ef/n9w0LFqlMKS9Rww FyM1xes6LBcw3g0dnLsKRmplnFFYdRYDkOzUddIY= From: Kieran Bingham To: libcamera devel , Chris Ward , Jacopo Mondi Date: Thu, 20 Aug 2020 14:47:50 +0100 Message-Id: <20200820134751.278033-3-kieran.bingham@ideasonboard.com> X-Mailer: git-send-email 2.25.1 In-Reply-To: <20200820134751.278033-1-kieran.bingham@ideasonboard.com> References: <20200820134751.278033-1-kieran.bingham@ideasonboard.com> MIME-Version: 1.0 Subject: [libcamera-devel] [PATCH v4 2/3] Documentation: Guides: Pipeline Handler Writers Guide X-BeenThere: libcamera-devel@lists.libcamera.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: libcamera-devel-bounces@lists.libcamera.org Sender: "libcamera-devel" From: Chris Chinchilla Introduce a pipeline-handler writers guide to provide a walk through of the steps and concepts required to implement a new Pipeline Handler. Signed-off-by: Chris Chinchilla [Reflow/Rework, update to mainline API] Signed-off-by: Jacopo Mondi [Further reworks and review] Signed-off-by: Kieran Bingham Acked-by: Laurent Pinchart --- Documentation/guides/pipeline-handler.rst | 1473 +++++++++++++++++++++ Documentation/index.rst | 1 + Documentation/meson.build | 1 + 3 files changed, 1475 insertions(+) create mode 100644 Documentation/guides/pipeline-handler.rst diff --git a/Documentation/guides/pipeline-handler.rst b/Documentation/guides/pipeline-handler.rst new file mode 100644 index 000000000000..eb795dcc35b6 --- /dev/null +++ b/Documentation/guides/pipeline-handler.rst @@ -0,0 +1,1473 @@ +.. SPDX-License-Identifier: CC-BY-SA-4.0 + +Pipeline Handler Writers Guide +============================== + +Pipeline handlers are the abstraction layer for device-specific hardware +configuration. They access and control hardware through the V4L2 and Media +Controller kernel interfaces, and implement an internal API to control the ISP +and capture components of a pipeline directly. + +Prerequisite knowledge: system architecture +------------------------------------------- + +A pipeline handler configures and manages the image acquisition and +transformation pipeline realized by specialized system peripherals combined with +an image source connected to the system through a data and control bus. The +presence, number and characteristics of them vary depending on the system design +and the product integration of the target platform. + +System components can be classified in three macro-categories: + +.. TODO: Insert references to the open CSI-2 (and other) specification. + +- Input ports: Interfaces to external devices, usually image sensors, + which transfer data from the physical bus to locations accessible by other + system peripherals. An input port needs to be configured according to the + input image format and size and could optionally apply basic transformations + on the received images, most typically cropping/scaling and some formats + conversion. The industry standard for the system typically targeted by + libcamera is to have receivers compliant with the MIPI CSI-2 specifications, + implemented on a compatible physical layer such as MIPI D-PHY or MIPI C-PHY. + Other design are possible but less common, such as LVDS or the legacy BT.601 + and BT.656 parallel protocols. + +- Image Signal Processor (ISP): A specialized media processor which applies + digital transformations on image streams. ISPs can be integrated as part of + the SoC as a memory interfaced system peripheral or packaged as stand-alone + chips connected to the application processor through a bus. Most hardware used + by libcamera makes use of in-system ISP designs but pipelines can equally + support external ISP chips or be instrumented to use other system resources + such as a GPU or an FPGA IP block. ISPs expose a software programming + interface that allows the configuration of multiple processing blocks which + form an "Image Transformation Pipeline". An ISP usually produces 'processed' + image streams along with the metadata describing the processing steps which + have been applied to generate the output frames. + +- Camera Sensor: Digital components that integrate an image sensor with control + electronics and usually a lens. It interfaces to the SoC image receiver ports + and is programmed to produce images in a format and size suitable for the + current system configuration. Complex camera modules can integrate on-board + ISP or DSP chips and process images before delivering them to the system. Most + systems with a dedicated ISP processor will usually integrate camera sensors + which produce images in Raw Bayer format and defer processing to it. + +It is the responsibility of the pipeline handler to interface with these (and +possibly other) components of the system and implement the following +functionalities: + +- Detect and register camera devices available in the system with an associated + set of image streams. + +- Configure the image acquisition and processing pipeline by assigning the + system resources (memory, shared components, etc.) to satisfy the + configuration requested by the application. + +- Start and the stop the image acquisition and processing sessions. + +- Apply configuration settings requested by applications and computed by image + processing algorithms integrated in libcamera to the hardware devices. + +- Notify applications of the availability of new images and deliver them to the + correct locations. + +Prerequisite knowledge: libcamera architecture +---------------------------------------------- + +A pipeline handler makes use of the following libcamera classes to realize the +functionalities descibed above. Below is a brief overview of each of those: + +.. TODO: (All) Convert to sphinx refs +.. TODO: (MediaDevice) Reference to the Media Device API (possibly with versioning requirements) +.. TODO: (IPAInterface) refer to the IPA guide + +- `MediaDevice `_: + Instances of this class are associated with a kernel media controller + device and its connected objects. + +- `DeviceEnumerator `_: + Enumerates all media devices attached to the system and the media entities + registered with it, by creating instances of the ``MediaDevice`` class and + storing them. + +- `DeviceMatch `_: + Describes a media device search pattern using entity names, or other + properties. + +- `V4L2VideoDevice `_: + Models an instance of a V4L2 video device constructed with the path to a V4L2 + video device node. + +- `V4L2SubDevice `_: + Provides an API to the sub-devices that model the hardware components of a + V4L2 device. + +- `CameraSensor `_: + Abstracts camera sensor handling by hiding the details of the V4L2 subdevice + kernel API and caching sensor information. + +- `CameraData `_: + Represents device-specific data a pipeline handler associates to each Camera + instance. + +- `StreamConfiguration `_: + Models the current configuration of an image stream produced by the camera by + reporting its format and sizes. + +- `CameraConfiguration `_: + Represents the current configuration of a camera, which includes a list of + stream configurations for each active stream in a capture session. When + validated, it is applied to the camera. + +- `IPAInterface `_: + The interface to the Image Processing Algorithm (IPA) module which performs + the computation of the image processing pipeline tuning parameters. + +- `ControlList `_: + A list of control items, indexed by Control<> instances or by numerical index + which contains values used by application and IPA to change parameters of + image streams, used to return to applications and share with IPA the metadata + associated with the captured images, and to advertise the immutable camera + characteristics enumerated at system initialization time. + +Creating a PipelineHandler +-------------------------- + +This guide walks through the steps to create a simple pipeline handler +called “Vivid” that supports the `V4L2 Virtual Video Test Driver`_ (vivid). + +To use the vivid test driver, you first need to check that the vivid kernel +module is loaded, for example with the ``modprobe vivid`` command. + +.. _V4L2 Virtual Video Test Driver: https://www.kernel.org/doc/html/latest/admin-guide/media/vivid.html + +Create the skeleton file structure +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +To add a new pipeline handler, create a directory to hold the pipeline code in +the *src/libcamera/pipeline/* directory that matches the name of the pipeline +(in this case *vivid*). Inside the new directory add a *meson.build* file that +integrates with the libcamera build system, and a *vivid.cpp* file that matches +the name of the pipeline. + +In the *meson.build* file, add the *vivid.cpp* file as a build source for +libcamera by adding it to the global meson ``libcamera_sources`` variable: + +.. code-block:: none + + # SPDX-License-Identifier: CC0-1.0 + + libcamera_sources += files([ + 'vivid.cpp', + ]) + +Users of libcamera can selectively enable pipelines while building libcamera +using the ``pipelines`` option. + +For example, to enable only the IPU3, UVC, and VIVID pipelines, specify them as +a comma separated list with ``-Dpipelines`` when generating a build directory: + +.. code-block:: shell + + meson build -Dpipelines=ipu3,uvcvideo,vivid + +Read the `Meson build configuration`_ documentation for more information on +configuring a build directory. + +.. _Meson build configuration: https://mesonbuild.com/Configuring-a-build-directory.html + +To add the new pipeline handler to this list of options, add its directory name +to the libcamera build options in the top level _meson_options.txt_. + +.. code-block:: none + + option('pipelines', + type : 'array', + choices : ['ipu3', 'raspberrypi', 'rkisp1', 'simple', 'uvcvideo', 'vimc', 'vivid'], + description : 'Select which pipeline handlers to include') + + +In *vivid.cpp* add the pipeline handler to the ``libcamera`` namespace, defining +a `PipelineHandler`_ derived class named PipelineHandlerVivid, and add stub +methods for the overridden class members. + +.. _PipelineHandler: http://libcamera.org/api-html/classlibcamera_1_1PipelineHandler.html + +.. code-block:: cpp + + namespace libcamera { + + class PipelineHandlerVivid : public PipelineHandler + { + public: + PipelineHandlerVivid(CameraManager *manager); + + CameraConfiguration *generateConfiguration(Camera *camera, + const StreamRoles &roles) override; + int configure(Camera *camera, CameraConfiguration *config) override; + + int exportFrameBuffers(Camera *camera, Stream *stream, + std::vector> *buffers) override; + + int start(Camera *camera) override; + void stop(Camera *camera) override; + + int queueRequestDevice(Camera *camera, Request *request) override; + + bool match(DeviceEnumerator *enumerator) override; + }; + + PipelineHandlerVivid::PipelineHandlerVivid(CameraManager *manager) + : PipelineHandler(manager) + { + } + + CameraConfiguration *PipelineHandlerVivid::generateConfiguration(Camera *camera, + const StreamRoles &roles) + { + return nullptr; + } + + int PipelineHandlerVivid::configure(Camera *camera, CameraConfiguration *config) + { + return -1; + } + + int PipelineHandlerVivid::exportFrameBuffers(Camera *camera, Stream *stream, + std::vector> *buffers) + { + return -1; + } + + int PipelineHandlerVivid::start(Camera *camera) + { + return -1; + } + + void PipelineHandlerVivid::stop(Camera *camera) + { + } + + int PipelineHandlerVivid::queueRequestDevice(Camera *camera, Request *request) + { + return -1; + } + + bool PipelineHandlerVivid::match(DeviceEnumerator *enumerator) + { + return false; + } + + REGISTER_PIPELINE_HANDLER(PipelineHandlerVivid); + + } /* namespace libcamera */ + +Note that you must register the ``PipelineHandler`` subclass with the pipeline +handler factory using the `REGISTER_PIPELINE_HANDLER`_ macro which +registers it and creates a global symbol to reference the class and make it +available to try and match devices. + +.. _REGISTER_PIPELINE_HANDLER: http://libcamera.org/api-html/pipeline__handler_8h.html + +For debugging and testing a pipeline handler during development, you can define +a log message category for the pipeline handler. The ``LOG_DEFINE_CATEGORY`` +macro and ``LIBCAMERA_LOG_LEVELS`` environment variable help you use the inbuilt +libcamera `logging infrastructure`_ that allow for the inspection of internal +operations in a user-configurable way. + +.. _logging infrastructure: http://libcamera.org/api-html/log_8h.html + +Add the following before the ``PipelineHandlerVivid`` class declaration: + +.. code-block:: cpp + + LOG_DEFINE_CATEGORY(VIVID) + +At this point you need the following includes for logging and pipeline handler +features: + +.. code-block:: cpp + + #include "libcamera/internal/log.h" + #include "libcamera/internal/pipeline_handler.h" + +Run the following commands: + +.. code-block:: shell + + meson build + ninja -C build + +To build the libcamera code base, and confirm that the build system found the +new pipeline handler by running: + +.. code-block:: shell + + LIBCAMERA_LOG_LEVELS=Pipeline:0 ./build/src/cam/cam -l + +And you should see output like the below: + +.. code-block:: shell + + DEBUG Pipeline pipeline_handler.cpp:680 Registered pipeline handler "PipelineHandlerVivid" + +Matching devices +~~~~~~~~~~~~~~~~ + +Each pipeline handler registered in libcamera gets tested against the current +system configuration, by matching a ``DeviceMatch`` with the system +``DeviceEnumerator``. A successful match makes sure all the requested components +have been registered in the system and allows the pipeline handler to be +initialized. + +The main entry point of a pipeline handler is the `match()`_ class member +function. When the ``CameraManager`` is started (using the `start()`_ method), +all the registered pipeline handlers are iterated and their ``match`` function +called with an enumerator of all devices it found on a system. + +The match method should identify if there are suitable devices available in the +``DeviceEnumerator`` which the pipeline supports, returning ``true`` if it +matches a device, and ``false`` if it does not. To do this, construct a +`DeviceMatch`_ class with the name of the ``MediaController`` device to match. +You can specify the search further by adding specific media entities to the +search using the ``.add()`` method on the DeviceMatch. + +.. _match(): https://www.libcamera.org/api-html/classlibcamera_1_1PipelineHandler.html#a7cd5b652a2414b543ec20ba9dabf61b6 +.. _start(): http://libcamera.org/api-html/classlibcamera_1_1CameraManager.html#a49e322880a2a26013bb0076788b298c5 +.. _DeviceMatch: http://libcamera.org/api-html/classlibcamera_1_1DeviceMatch.html + +This example uses search patterns that match vivid, but when developing a new +pipeline handler, you should change this value to suit your device identifier. + +Replace the contents of the ``PipelineHandlerVivid::match`` method with the +following: + +.. code-block:: cpp + + DeviceMatch dm("vivid"); + dm.add("vivid-000-vid-cap"); + return false; // Prevent infinite loops for now + +With the device matching criteria defined, attempt to acquire exclusive access +to the matching media controller device with the `acquireMediaDevice`_ method. +If the method attempts to acquire a device it has already matched, it returns +``false``. + +.. _acquireMediaDevice: http://libcamera.org/api-html/classlibcamera_1_1PipelineHandler.html#a77e424fe704e7b26094164b9189e0f84 + +Add the following below ``dm.add("vivid-000-vid-cap");``: + +.. code-block:: cpp + + MediaDevice *media = acquireMediaDevice(enumerator, dm); + if (!media) + return false; + +The pipeline handler now needs an additional include. Add the following to the +existing include block for device enumeration functionality: + +.. code-block:: cpp + + #include "libcamera/internal/device_enumerator.h" + +At this stage, you should test that the pipeline handler can successfully match +the devices, but have not yet added any code to create a Camera which libcamera +reports to applications. + +As a temporary validation step, add a debug print with + +.. code-block:: cpp + + LOG(VIVID, Debug) << "Vivid Device Identified"; + +before the final closing return statement in the ``PipelineHandlerVivid::match`` +method for when when the pipeline handler successfully matches the +``MediaDevice`` and ``MediaEntity`` names. + +Test that the pipeline handler matches and finds a device by rebuilding, and +running + +.. code-block:: shell + + ninja -C build + LIBCAMERA_LOG_LEVELS=Pipeline,VIVID:0 ./build/src/cam/cam -l + +And you should see output like the below: + +.. code-block:: shell + + DEBUG VIVID vivid.cpp:74 Vivid Device Identified + +Creating camera devices +~~~~~~~~~~~~~~~~~~~~~~~ + +If the pipeline handler successfully matches with the system it is running on, +it can proceed to initialization, by creating all the required instances of the +``V4L2VideoDevice``, ``V4L2Subdevice`` and ``CameraSensor`` hardware abstraction +classes. If the Pipeline handler supports an ISP, it can then also Initialise +the IPA module before proceeding to the creation of the Camera devices. + +An image ``Stream`` represents a sequence of images and data of known size and +format, stored in application-accessible memory locations. Typical examples of +streams are the ISP processed outputs and the raw images captured at the +receivers port output. + +The Pipeline Handler is responsible for defining the set of Streams associated +with the Camera. + +Each Camera has instance-specific data represented using the `CameraData`_ +class, which can be extended for the specific needs of the pipeline handler. + +.. _CameraData: http://libcamera.org/api-html/classlibcamera_1_1CameraData.html + + +To support the Camera we will later register, we need to create a CameraData +class that we can implement for our specific Pipeline Handler. + +Define a new ``VividCameraData()`` class derived from ``CameraData`` by adding +the following code before the PipelineHandlerVivid class definition where it +will be used: + +.. code-block:: cpp + + class VividCameraData : public CameraData + { + public: + VividCameraData(PipelineHandler *pipe, MediaDevice *media) + : CameraData(pipe), media_(media), video_(nullptr) + { + } + + ~VividCameraData() + { + delete video_; + } + + int init(); + void bufferReady(FrameBuffer *buffer); + + MediaDevice *media_; + V4L2VideoDevice *video_; + Stream stream_; + }; + +This example pipeline handler handles a single video device and supports a +single stream, represented by the ``VividCameraData`` class members. More +complex pipeline handlers might register cameras composed of several video +devices and sub-devices, or multiple streams per camera that represent the +several components of the image capture pipeline. You should represent all these +components in the ``CameraData`` derived class when developing a custom +PipelineHandler. + +In our example VividCameraData we implement an ``init()`` function to prepare +the object from our PipelineHandler, however the CameraData class does not +specify the interface for initialisation and PipelineHandlers can manage this +based on their own needs. Derived CameraData classes are used only by their +respective pipeline handlers. + +The CameraData class stores the context required for each camera instance and +is usually responsible for opening all Devices used in the capture pipeline. + +We can now implement the ``init`` method for our example Pipeline Handler to +create a new V4L2 video device from the media entity, which we can specify using +the `MediaDevice::getEntityByName`_ method from the MediaDevice. As our example +is based upon the simplistic Vivid test device, we only need to open a single +capture device named 'vivid-000-vid-cap' by the device. + +.. _MediaDevice::getEntityByName: http://libcamera.org/api-html/classlibcamera_1_1MediaDevice.html#ad5d9279329ef4987ceece2694b33e230 + +.. code-block:: cpp + + int VividCameraData::init() + { + video_ = new V4L2VideoDevice(media_->getEntityByName("vivid-000-vid-cap")); + if (video_->open()) + return -ENODEV; + + return 0; + } + +The CameraData should be created and initialised before we move on to register a +new Camera device so we need to construct and initialise our +VividCameraData after we have identified our device within +PipelineHandlerVivid::match(). The VividCameraData is wrapped by a +std::unique_ptr to help manage the lifetime of our CameraData instance. + +If the camera data initialization fails, return ``false`` to indicate the +failure to the ``match()`` method and prevent retrying of the pipeline handler. + +.. code-block:: cpp + + std::unique_ptr data = std::make_unique(this, media); + + if (data->init()) + return false; + + +Once the camera data has been initialized, the Camera device instances and the +associated streams have to be registered. Create a set of streams for the +camera, which for this device is only one. You create a camera using the static +`Camera::create`_ method, passing the pipeline handler, the id of the camera, +and the streams available. Then register the camera and its data with the +pipeline handler and camera manager using `registerCamera`_. + +Finally with a successful construction, we return 'true' indicating that the +PipelineHandler successfully matched and constructed a device. + +.. _Camera::create: http://libcamera.org/api-html/classlibcamera_1_1Camera.html#a453740e0d2a2f495048ae307a85a2574 +.. _registerCamera: http://libcamera.org/api-html/classlibcamera_1_1PipelineHandler.html#adf02a7f1bbd87aca73c0e8d8e0e6c98b + +.. code-block:: cpp + + std::set streams{ &data->stream_ }; + std::shared_ptr camera = Camera::create(this, data->video_->deviceName(), streams); + registerCamera(std::move(camera), std::move(data)); + + return true; + + +Our match function should now look like the following: + +.. code-block:: cpp + + bool PipelineHandlerVivid::match(DeviceEnumerator *enumerator) + { + DeviceMatch dm("vivid"); + dm.add("vivid-000-vid-cap"); + + MediaDevice *media = acquireMediaDevice(enumerator, dm); + if (!media) + return false; + + std::unique_ptr data = std::make_unique(this, media); + + /* Locate and open the capture video node. */ + if (data->init()) + return false; + + /* Create and register the camera. */ + std::set streams{ &data->stream_ }; + std::shared_ptr camera = Camera::create(this, data->video_->deviceName(), streams); + registerCamera(std::move(camera), std::move(data)); + + return true; + } + +We will need to use our custom CameraData class frequently throughout the +pipeline handler, so we add a private convenience helper to our Pipeline handler +to obtain and cast the custom CameraData instance from a Camera instance. + +.. code-block:: cpp + + private: + VividCameraData *cameraData(const Camera *camera) + { + return static_cast( + PipelineHandler::cameraData(camera)); + } + +At this point, you need to add the following new includes to provide the Camera +interface, and device interaction interfaces. + +.. code-block:: cpp + + #include + #include "libcamera/internal/media_device.h" + #include "libcamera/internal/v4l2_videodevice.h" + +Registering controls and properties +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +The libcamera `controls framework`_ allows an application to configure the +streams capture parameters on a per-frame basis and is also used to advertise +immutable properties of the ``Camera`` device. + +The libcamera controls and properties are defined in YAML form which is +processed to automatically generate documentation and interfaces. Controls are +defined by the src/libcamera/`control_ids.yaml`_ file and camera properties +are defined by src/libcamera/`properties_ids.yaml`_. + +.. _controls framework: http://libcamera.org/api-html/controls_8h.html +.. _control_ids.yaml: http://libcamera.org/api-html/control__ids_8h.html +.. _properties_ids.yaml: http://libcamera.org/api-html/property__ids_8h.html + +Pipeline handlers can optionally register the list of controls an application +can set as well as a list of immutable camera properties. Being both +Camera-specific values, they are represented in the ``CameraData`` base class, +which provides two members for this purpose: the `CameraData::controlInfo_`_ and +the `CameraData::properties_`_ fields. + +.. _CameraData::controlInfo_: http://libcamera.org/api-html/classlibcamera_1_1CameraData.html#ab9fecd05c655df6084a2233872144a52 +.. _CameraData::properties_: http://libcamera.org/api-html/classlibcamera_1_1CameraData.html#a84002c29f45bd35566c172bb65e7ec0b + +The ``controlInfo_`` field represents a map of ``ControlId`` instances +associated with the limits of valid values supported for the control. More +information can be found in the `ControlInfoMap`_ class documentation. + +.. _ControlInfoMap: http://libcamera.org/api-html/classlibcamera_1_1ControlInfoMap.html + +Pipeline handlers register controls to expose the tunable device and IPA +parameters to applications. Our example pipeline handler only exposes trivial +controls of the video device, by registering a ``ControlId`` instance with +associated values for each supported V4L2 control but demonstrates the mapping +of V4L2 Controls to libcamera ControlIDs. + +Complete the initialization of the ``VividCameraData`` class by adding the +following code to the ``VividCameraData::init()`` method to initialise the +controls. For more complex control configurations, this could of course be +broken out to a separate function, but for now we just initialise the small set +inline in our CameraData init: + +.. code-block:: cpp + + /* Initialise the supported controls. */ + const ControlInfoMap &controls = video_->controls(); + ControlInfoMap::Map ctrls; + + for (const auto &ctrl : controls) { + const ControlId *id; + ControlInfo info; + + switch (ctrl.first->id()) { + case V4L2_CID_BRIGHTNESS: + id = &controls::Brightness; + info = ControlInfo{ { -1.0f }, { 1.0f }, { 0.0f } }; + break; + case V4L2_CID_CONTRAST: + id = &controls::Contrast; + info = ControlInfo{ { 0.0f }, { 2.0f }, { 1.0f } }; + break; + case V4L2_CID_SATURATION: + id = &controls::Saturation; + info = ControlInfo{ { 0.0f }, { 2.0f }, { 1.0f } }; + break; + default: + continue; + } + + ctrls.emplace(id, info); + } + + controlInfo_ = std::move(ctrls); + +The ``properties_`` field is a list of ``ControlId`` instances +associated with immutable values, which represent static characteristics that can +be used by applications to identify camera devices in the system. Properties can be +registered by inspecting the values of V4L2 controls from the video devices and +camera sensor (for example to retrieve the position and orientation of a camera) +or to express other immutable characteristics. The example pipeline handler does +not register any property, but examples are available in the libcamera code +base. + +.. TODO: Add a property example to the pipeline handler. At least the model. + +At this point you need to add the following includes to the top of the file for +handling controls: + +.. code-block:: cpp + + #include + #include + +Generating a default configuration +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Once ``Camera`` devices and the associated ``Streams`` have been registered, an +application can proceed to acquire and configure the camera to prepare it for a +frame capture session. + +Applications specify the requested configuration by assigning a +``StreamConfiguration`` instance to each stream they want to enable which +expresses the desired image size and pixel format. The stream configurations are +grouped in a ``CameraConfiguration`` which is inspected by the pipeline handler +and validated to adjust it to a supported configuration. This may involve +adjusting the formats or image sizes or alignments for example to match the +capabilities of the device. + +Applications may choose to repeat validation stages, adjusting paramters until a +set of validated StreamConfigurations are returned that is acceptable for the +applications needs. When the pipeline handler receives a valid camera +configuration it can use the image stream configurations to apply settings to +the hardware devices. + +This configuration and validation process is managed with another Pipeline +specific class derived from a common base implementation and interface. + +To support validation in our example pipeline handler, Create a new class called +``VividCameraConfiguration`` derived from the base `CameraConfiguration`_ class +which we can implement and use within our ``PipelineHandlerVivid`` class. + +.. _CameraConfiguration: http://libcamera.org/api-html/classlibcamera_1_1CameraConfiguration.html + +The derived ``CameraConfiguration`` class must override the base class +``validate()`` function, where the stream configuration inspection and +adjustment happens. + +.. code-block:: cpp + + class VividCameraConfiguration : public CameraConfiguration + { + public: + VividCameraConfiguration(); + + Status validate() override; + }; + + VividCameraConfiguration::VividCameraConfiguration() + : CameraConfiguration() + { + } + +Applications generate a ``CameraConfiguration`` instance by calling the +`Camera::generateConfiguration()`_ function, which calls into the pipeline +implementation of the overridden `PipelineHandler::generateConfiguration()`_ +method. + +.. _Camera::generateConfiguration(): http://libcamera.org/api-html/classlibcamera_1_1Camera.html#a25c80eb7fc9b1cf32692ce0c7f09991d +.. _PipelineHandler::generateConfiguration(): http://libcamera.org/api-html/classlibcamera_1_1PipelineHandler.html#a7932e87735695500ce1f8c7ae449b65b + +Configurations are generated by receiving a list of ``StreamRoles`` instances, +which libcamera uses as predefined ways an application intends to use a camera +(You can read the full list in the `StreamRole API`_ documentation). These are +optional hints on how an application intends to use a stream, and a pipeline +handler should return an ideal configuration for each role that is requested. + +.. _StreamRole API: http://libcamera.org/api-html/stream_8h.html#file_a295d1f5e7828d95c0b0aabc0a8baac03 + +In the pipeline handler ``generateConfiguration`` implementation, remove the +``return nullptr;``, create a new instance of the ``CameraConfiguration`` +derived class, and assign it to a base class pointer. + +.. code-block:: cpp + + VividCameraData *data = cameraData(camera); + CameraConfiguration *config = new VividCameraConfiguration(); + +A ``CameraConfiguration`` is specific to each pipeline, so you can only create +it from the pipeline handler code path. Applications can also generate an empty +configuration and add desired stream configurations manually. Pipelines must +allow for this by returning an empty configuration if no roles are requested. + +To support this in our PipelineHandlerVivid, next add the following check in +``generateConfiguration`` after the Cameraconfiguration has been constructed: + +.. code-block:: cpp + + if (roles.empty()) + return config; + +A production pipeline handler should generate the ``StreamConfiguration`` for +all the appropriate stream roles a camera device supports. For this simpler +example (with only one stream), the pipeline handler always returns the same +configuration, inferred from the underlying V4L2VideoDevice. + +How it does this is shown below, but examination of the more full-featured +pipelines for IPU3, RKISP1 and RaspberryPi are recommend to explore more +complex examples. + +To generate a ``StreamConfiguration``, you need a list of pixel formats and +frame sizes which supported outputs of the stream. You can fetch a map of the +``V4LPixelFormat`` and ``SizeRange`` supported by the underlying output device, +but the pipeline handler needs to convert this to a ``libcamera::PixelFormat`` +type to pass to applications. We do this here using ``std::transform`` to +convert the formats and populate a new ``PixelFormat`` map as shown below. + +Continue adding the following code example to our ``generateConfiguration`` +implementation. + +.. code-block:: cpp + + std::map> v4l2Formats = + data->video_->formats(); + std::map> deviceFormats; + std::transform(v4l2Formats.begin(), v4l2Formats.end(), + std::inserter(deviceFormats, deviceFormats.begin()), + [&](const decltype(v4l2Formats)::value_type &format) { + return decltype(deviceFormats)::value_type{ + format.first.toPixelFormat(), + format.second + }; + }); + +The `StreamFormats`_ class holds information about the pixel formats and frame +sizes that a stream can support. The class groups size information by the pixel +format, which can produce it. + +.. _StreamFormats: http://libcamera.org/api-html/classlibcamera_1_1StreamFormats.html + +The code below uses the ``StreamFormats`` class to represent all of the +supported pixel formats, associated with a list of frame sizes. It then +generates a supported StreamConfiguration to model the information an +application can use to configure a single stream. + +Continue adding the following code to support this: + +.. code-block:: cpp + + StreamFormats formats(deviceFormats); + StreamConfiguration cfg(formats); + +As well as a list of supported StreamFormats, the StreamConfiguration is also +expected to provide an initialsed default configuration. This may be arbitrary, +but depending on use case you may which to select an output that matches the +Sensor output, or prefer a pixelformat which might provide higher performance on +the hardware. The bufferCount represents the number of buffers required to +support functional continuous processing on this stream. + +.. code-block:: cpp + + cfg.pixelFormat = formats::BGR888; + cfg.size = { 1280, 720 }; + cfg.bufferCount = 4; + +Finally add each ``StreamConfiguration`` generated to the +``CameraConfiguration``, and ensure that it has been validated before returning +it to the application. With only a single supported stream, this code adds only +a single StreamConfiguration however a StreamConfiguration should be added for +each supported role in a device that can handle more streams. + +Add the following code to complete the implementation of +``generateConfiguration``: + +.. code-block:: cpp + + config->addConfiguration(cfg); + + config->validate(); + + return config; + +To validate a camera configuration, a pipeline handler must implement the +`CameraConfiguration::validate()`_ method in it's derived class to inspect all +the stream configuration associated to it, make any adjustments required to make +the configuration valid, and return the validation status. + +If changes are made, it marks the configuration as ``Adjusted``, however if the +requested configuration is not supported and cannot be adjusted it shall be +refused and marked as ``Invalid``. + +.. _CameraConfiguration::validate(): http://libcamera.org/api-html/classlibcamera_1_1CameraConfiguration.html#a29f8f263384c6149775b6011c7397093 + +The validation phase makes sure all the platform-specific constraints are +respected by the requested configuration. The most trivial examples being making +sure the requested image formats are supported and the image alignment +restrictions adhered to. The pipeline handler specific implementation of +``validate()`` shall inspect all the configuration parameters received and never +assume they are correct, as applications are free to change the requested stream +parameters after the configuration has been generated. + +Again, this example pipeline handler is simpler, look at the more complex +implementations for a realistic example. + +Add the following function implementation to your file: + +.. code-block:: cpp + + CameraConfiguration::Status VividCameraConfiguration::validate() + { + Status status = Valid; + + if (config_.empty()) + return Invalid; + + if (config_.size() > 1) { + config_.resize(1); + status = Adjusted; + } + + StreamConfiguration &cfg = config_[0]; + + const std::vector formats = cfg.formats().pixelformats(); + if (std::find(formats.begin(), formats.end(), cfg.pixelFormat) == formats.end()) { + cfg.pixelFormat = cfg.formats().pixelformats()[0]; + LOG(VIVID, Debug) << "Adjusting format to " << cfg.pixelFormat.toString(); + status = Adjusted; + } + + cfg.bufferCount = 4; + + return status; + } + +Now that we are handling the ``PixelFormat`` type, we also need to add +``#include `` to the include section before we rebuild the +codebase, and test: + +.. code-block:: shell + + ninja -C build + LIBCAMERA_LOG_LEVELS=Pipeline,VIVID:0 ./build/src/cam/cam -c vivid -I + +You should see the following output showing the capabilites of our new pipeline +handler, and showing that our configurations have been generated: + +.. code-block:: shell + + Using camera vivid + 0: 1280x720-BGR888 + * Pixelformat: NV21 (320x180)-(3840x2160)/(+0,+0) + - 320x180 + - 640x360 + - 640x480 + - 1280x720 + - 1920x1080 + - 3840x2160 + * Pixelformat: NV12 (320x180)-(3840x2160)/(+0,+0) + - 320x180 + - 640x360 + - 640x480 + - 1280x720 + - 1920x1080 + - 3840x2160 + * Pixelformat: BGRA8888 (320x180)-(3840x2160)/(+0,+0) + - 320x180 + - 640x360 + - 640x480 + - 1280x720 + - 1920x1080 + - 3840x2160 + * Pixelformat: RGBA8888 (320x180)-(3840x2160)/(+0,+0) + - 320x180 + - 640x360 + - 640x480 + - 1280x720 + - 1920x1080 + - 3840x2160 + +Configuring a device +~~~~~~~~~~~~~~~~~~~~ + +With the configuration generated, and optionally modified and re-validated, a +pipeline handler needs a method that allows an application to apply a +configuration to the hardware devices. + +The `PipelineHandler::configure()`_ method receives a valid +`CameraConfiguration`_ and applies the settings to hardware devices, using its +parameters to prepare a device for a streaming session with the desired +properties. + +.. _PipelineHandler::configure(): http://libcamera.org/api-html/classlibcamera_1_1PipelineHandler.html#a930f2a9cdfb51dfb4b9ca3824e84fc29 +.. _CameraConfiguration: http://libcamera.org/api-html/classlibcamera_1_1CameraConfiguration.html + +Replace the contents of the stubbed ``PipelineHandlerVivid::configure`` method +with the following to obtain the camera data and stream configuration. This +pipeline handler supports only a single stream, so it directly obtains the first +``StreamConfiguration`` from the camera configuration. A pipeline handler with +multiple streams should inspect each StreamConfiguration and configure the +system accordingly. + +.. code-block:: cpp + + VividCameraData *data = cameraData(camera); + StreamConfiguration &cfg = config->at(0); + int ret; + +The Vivid capture device is a V4L2 video device, so we use a `V4L2DeviceFormat`_ +with the fourcc and size attributes to apply directly to the capture device +node. The fourcc attribute is a `V4L2PixelFormat`_ and differs from the +``libcamera::PixelFormat``. Converting the format requires knowledge of the +plane configuration for multiplanar formats, so you must explicitly convert it +using the helper ``V4L2VideoDevice::toV4L2PixelFormat()`` provided by the +V4L2VideoDevice instance of which the format will be applied on. + +.. _V4L2DeviceFormat: http://libcamera.org/api-html/classlibcamera_1_1V4L2DeviceFormat.html +.. _V4L2PixelFormat: http://libcamera.org/api-html/classlibcamera_1_1V4L2PixelFormat.html + +Add the following code beneath the code from above: + +.. code-block:: cpp + + V4L2DeviceFormat format = {}; + format.fourcc = data->video_->toV4L2PixelFormat(cfg.pixelFormat); + format.size = cfg.size; + +Set the video device format defined above using the +`V4L2VideoDevice::setFormat()`_ method. You should check if the kernel +driver has adjusted the format, as this shows the pipeline handler has failed to +handle the validation stages correctly, and the configure operation shall also +fail. + +.. _V4L2VideoDevice::setFormat(): http://libcamera.org/api-html/classlibcamera_1_1V4L2VideoDevice.html#ad67b47dd9327ce5df43350b80c083cca + +Continue the implementation with the following code: + +.. code-block:: cpp + + ret = data->video_->setFormat(&format); + if (ret) + return ret; + + if (format.size != cfg.size || + format.fourcc != data->video_->toV4L2PixelFormat(cfg.pixelFormat)) + return -EINVAL; + +Finally, store and set stream-specific data reflecting the state of the stream. +Associate the configuration with the stream by using the +`StreamConfiguration::setStream`_ method, and set the values of individual +stream configuration members as required. + +.. _StreamConfiguration::setStream: http://libcamera.org/api-html/structlibcamera_1_1StreamConfiguration.html#a74a0eb44dad1b00112c7c0443ae54a12 + +.. NOTE: the cfg.setStream() call here associates the stream to the + StreamConfiguration however that should quite likely be done as part of + the validation process. TBD + +Complete the configure implementation with the following code: + +.. code-block:: cpp + + cfg.setStream(&data->stream_); + cfg.stride = format.planes[0].bpl; + + return 0; + +.. TODO: stride SHALL be assigned in validate + +Initializing device controls +~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Pipeline handlers can optionally initialize the video devices and camera sensor +controls at system configuration time, to make sure to make sure they are +defaulted to sane values. Handling of device controls is again performed using +the libcamera `controls framework`_. + +.. _Controls Framework: http://libcamera.org/api-html/controls_8h.html + +This section is particularly specific to Vivid as it sets the initial values of +controls to match `Vivid Controls`_ defined by the kernel driver. You won’t need +any of the code below for your pipeline handler, but it’s included as an example +of how to implement functionality your pipeline handler might need. + +.. _Vivid Controls: https://www.kernel.org/doc/html/latest/admin-guide/media/vivid.html#controls + +We need to add some definitions at the top of the file for convenience. These +come directly from the kernel sources: + +.. code-block:: cpp + + #define VIVID_CID_VIVID_BASE (0x00f00000 | 0xf000) + #define VIVID_CID_VIVID_CLASS (0x00f00000 | 1) + #define VIVID_CID_TEST_PATTERN (VIVID_CID_VIVID_BASE + 0) + #define VIVID_CID_OSD_TEXT_MODE (VIVID_CID_VIVID_BASE + 1) + #define VIVID_CID_HOR_MOVEMENT (VIVID_CID_VIVID_BASE + 2) + +We can now use the V4L2 control IDs to prepare a list of controls with the +`ControlList`_ class, and set them using the `ControlList::set()`_ method. + +.. _ControlList: http://libcamera.org/api-html/classlibcamera_1_1ControlList.html +.. _ControlList::set(): http://libcamera.org/api-html/classlibcamera_1_1ControlList.html#a74a1a29abff5243e6e37ace8e24eb4ba + +In our pipeline ``configure`` method, add the following code after the format +has been set and checked to initialise the ControlList and apply it to the +device: + +.. code-block:: cpp + + ControlList controls(data->video_->controls()); + controls.set(VIVID_CID_TEST_PATTERN, 0); + controls.set(VIVID_CID_OSD_TEXT_MODE, 0); + + controls.set(V4L2_CID_BRIGHTNESS, 128); + controls.set(V4L2_CID_CONTRAST, 128); + controls.set(V4L2_CID_SATURATION, 128); + + controls.set(VIVID_CID_HOR_MOVEMENT, 5); + + ret = data->video_->setControls(&controls); + if (ret) { + LOG(VIVID, Error) << "Failed to set controls: " << ret; + return ret < 0 ? ret : -EINVAL; + } + +These controls configure VIVID to use a default test pattern, and enable all +on-screen display text, while configuring sensible brightness, contrast and +saturation values. Use the ``controls.set`` method to set individual controls. + +Buffer handling and stream control +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Once the system has been configured with the requested parameters, it is now +possible for applications to start capturing frames from the ``Camera`` device. + +Libcamera implements a per-frame request capture model, realized by queueing +``Request`` instances to a ``Camera`` object. Before applications can start +submitting capture requests the capture pipeline needs to be prepared to deliver +frames as soon as they are requested. Memory should be initialized and made +available to the devices which have to be started and ready to produce +images. At the end of a capture session the ``Camera`` device needs to be +stopped, to gracefully clean up any allocated memory and stop the hardware +devices. Pipeline handlers implement two methods for these purposes, the +``start()`` and ``stop()`` methods. + +The memory initialization phase that happens at ``start()`` time serves to +configure video devices to be able to use memory buffers exported as dma-buf +file descriptors. From the pipeline handlers perspective the video devices that +provide application facing streams always act as memory importers which use, +in V4L2 terminology, buffers of V4L2_MEMORY_DMABUF memory type. + +Libcamera also provides an API to allocate and export memory to applications +realized through the `exportFrameBuffers`_ function and the +`FrameBufferAllocator`_ class which will be presented later. + +.. _exportFrameBuffers: http://libcamera.org/api-html/classlibcamera_1_1PipelineHandler.html#a6312a69da7129c2ed41f9d9f790adf7c +.. _FrameBufferAllocator: http://libcamera.org/api-html/classlibcamera_1_1FrameBufferAllocator.html + +Please refer to the V4L2VideoDevice API documentation, specifically the +`allocateBuffers`_, `importBuffers`_ and `exportBuffers`_ functions for a +detailed description of the video device memory management. + +.. _allocateBuffers: http://libcamera.org/api-html/classlibcamera_1_1V4L2VideoDevice.html#a3a1a77e5e6c220ea7878e89485864a1c +.. _importBuffers: http://libcamera.org/api-html/classlibcamera_1_1V4L2VideoDevice.html#a154f5283d16ebd5e15d63e212745cb64 +.. _exportBuffers: http://libcamera.org/api-html/classlibcamera_1_1V4L2VideoDevice.html#ae9c0b0a68f350725b63b73a6da5a2ecd + +Video memory buffers are represented in libcamera by the `FrameBuffer`_ class. +A ``FrameBuffer`` instance has to be associated to each ``Stream`` which is part +of a capture ``Request``. Pipeline handlers should prepare the capture devices +by importing the dma-buf file descriptors it needs to operate on. This operation +is performed by using the ``V4L2VideoDevice`` API, which provides an +``importBuffers()`` function that prepares the video device accordingly. + +.. _FrameBuffer: http://libcamera.org/api-html/classlibcamera_1_1FrameBuffer.html + +Implement the pipeline handler ``start()`` function by replacing the stub +version with the following code: + +.. code-block:: c++ + + VividCameraData *data = cameraData(camera); + unsigned int count = data->stream_.configuration().bufferCount; + + int ret = data->video_->importBuffers(count); + if (ret < 0) + return ret; + + return 0; + +During the startup phase pipeline handlers allocate any internal buffer pool +required to transfer data between different components of the image capture +pipeline, for example, between the CSI-2 receiver and the ISP input. The example +pipeline does not require any internal pool, but examples are available in more +complex pipeline handlers in the libcamera code base. + +Applications might want to use memory allocated in the video devices instead of +allocating it from other parts of the system. libcamera provides an abstraction +to assist with this task in the `FrameBufferAllocator`_ class. The +``FrameBufferAllocator`` reserves memory for a ``Stream`` in the video device +and exports it as dma-buf file descriptors. From this point on, the allocated +``FrameBuffer`` are associated to ``Stream`` instances in a ``Request`` and then +imported by the pipeline hander in exactly the same fashion as if they were +allocated elsewhere. + +.. _FrameBufferAllocator: http://libcamera.org/api-html/classlibcamera_1_1FrameBufferAllocator.html + +Pipeline handlers support the ``FrameBufferAllocator`` operations by +implementing the `exportFrameBuffers`_ function, which will allocate memory in +the video device associated with a stream and export it. + +.. _exportFrameBuffers: http://libcamera.org/api-html/classlibcamera_1_1PipelineHandler.html#a6312a69da7129c2ed41f9d9f790adf7c + +Implement the ``exportFrameBuffers`` stub method with the following code to +handle this: + +.. code-block:: cpp + + unsigned int count = stream->configuration().bufferCount; + VividCameraData *data = cameraData(camera); + + return data->video_->exportBuffers(count, buffers); + +Once memory has been properly setup, the video devices can be started, to +prepare for capture operations. Complete the ``start`` method implementation +with the following code: + +.. code-block:: cpp + + ret = data->video_->streamOn(); + if (ret < 0) { + data->video_->releaseBuffers(); + return ret; + } + + return 0; + +The method starts the video device associated with the stream with the +`streamOn`_ method. If the call fails, the error value is propagated to the +caller and the `releaseBuffers`_ method releases any buffers to leave the device +in a consistent state. If your pipeline handler uses any image processing +algorithms, or other devices you should also stop them. + +.. _streamOn: http://libcamera.org/api-html/classlibcamera_1_1V4L2VideoDevice.html#a588a5dc9d6f4c54c61136ac43ff9a8cc +.. _releaseBuffers: http://libcamera.org/api-html/classlibcamera_1_1V4L2VideoDevice.html#a191619c152f764e03bc461611f3fcd35 + +Of course we also need to handle the corresponding actions to stop streaming on +a device, Add the following to the ``stop`` method, to stop the stream with the +`streamOff`_ method and release all buffers. + +.. _streamOff: http://libcamera.org/api-html/classlibcamera_1_1V4L2VideoDevice.html#a61998710615bdf7aa25a046c8565ed66 + +.. code-block:: cpp + + VividCameraData *data = cameraData(camera); + data->video_->streamOff(); + data->video_->releaseBuffers(); + +Queuing requests between applications and hardware +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +libcamera implements a streaming model based on capture requests queued by an +application to the ``Camera`` device. Each request contains at least one +``Stream`` instance with an associated ``FrameBuffer`` object. + +When an application sends a capture request, the pipeline handler identifies +which video devices have to be provided with buffers to generate a frame from +the enabled streams. + +This example pipeline handler identifies the buffer using the `findBuffer`_ +helper from the only supported stream and queues it to the capture device +directly with the `queueBuffer`_ method provided by the V4L2VideoDevice. + +.. _findBuffer: http://libcamera.org/api-html/classlibcamera_1_1Request.html#ac66050aeb9b92c64218945158559c4d4 +.. _queueBuffer: http://libcamera.org/api-html/classlibcamera_1_1V4L2VideoDevice.html#a594cd594686a8c1cf9ae8dba0b2a8a75 + +Replace the stubbed contents of ``queueRequestDevice`` with the following: + +.. code-block:: cpp + + VividCameraData *data = cameraData(camera); + FrameBuffer *buffer = request->findBuffer(&data->stream_); + if (!buffer) { + LOG(VIVID, Error) + << "Attempt to queue request with invalid stream"; + + return -ENOENT; + } + + int ret = data->video_->queueBuffer(buffer); + if (ret < 0) + return ret; + + return 0; + +Processing controls +~~~~~~~~~~~~~~~~~~~ + +Capture requests not only contain streams and memory buffers, but can +optionally contain a list of controls the application has set to modify the +streaming parameters. + +Applications can set controls registered by the pipeline handler in the +initialization phase, as explained in the `Registering controls and properties`_ +section. + +Implement a ``processControls`` method above the ``queueRequestDevice`` method +to loop through the control list received with each request, and inspect the +control values. Controls may need to be converted between the libcamera control +range definitions and their corresponding values on the device before being set. + +.. code-block:: cpp + + int PipelineHandlerVivid::processControls(VividCameraData *data, Request *request) + { + ControlList controls(data->video_->controls()); + + for (auto it : request->controls()) { + unsigned int id = it.first; + unsigned int offset; + uint32_t cid; + + if (id == controls::Brightness) { + cid = V4L2_CID_BRIGHTNESS; + offset = 128; + } else if (id == controls::Contrast) { + cid = V4L2_CID_CONTRAST; + offset = 0; + } else if (id == controls::Saturation) { + cid = V4L2_CID_SATURATION; + offset = 0; + } else { + continue; + } + + int32_t value = lroundf(it.second.get() * 128 + offset); + controls.set(cid, utils::clamp(value, 0, 255)); + } + + for (const auto &ctrl : controls) + LOG(VIVID, Debug) + << "Setting control " << utils::hex(ctrl.first) + << " to " << ctrl.second.toString(); + + int ret = data->video_->setControls(&controls); + if (ret) { + LOG(VIVID, Error) << "Failed to set controls: " << ret; + return ret < 0 ? ret : -EINVAL; + } + + return ret; + } + +Declare the function prototype for the ``processControls`` method within the +private ``PipelineHandlerVivid`` class members, as it is only used internally as +a helper when processing Requests. + +.. code-block:: cpp + + private: + int processControls(VividCameraData *data, Request *request); + +A pipeline handler is responsible for applying controls provided in a Request to +the relevant hardware devices. This could be directly on the capture device, or +where appropriate by setting controls on V4L2Subdevices directly. Each pipeline +handler is responsible for understanding the correct procedure for applying +controls to the device they support. + +This example pipeline handler applies controls during the `queueRequestDevice`_ +method for each request, and applies them to the capture device through the +capture node. + +.. _queueRequestDevice: http://libcamera.org/api-html/classlibcamera_1_1PipelineHandler.html#a106914cca210640c9da9ee1f0419e83c + +In the ``queueRequestDevice`` method, replace the following: + +.. code-block:: cpp + + int ret = data->video_->queueBuffer(buffer); + if (ret < 0) + return ret; + +With the following code: + +.. code-block:: cpp + + int ret = processControls(data, request); + if (ret < 0) + return ret; + + ret = data->video_->queueBuffer(buffer); + if (ret < 0) + return ret; + +We also need to add the following include directive to support the control +value translation operations: + +.. code-block:: cpp + + #include + +Frame completion and event handling +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Libcamera implements a signals and slots mechanism (similar to `Qt Signals and +Slots`_) to connect event sources with callbacks to handle them. + +As a general summary, a ``Slot`` can be connected to a ``Signal``, which when +emitted triggers the execution of the connected slots. A detailed description +of the libcamera implementation is available in the `libcamera Signal and Slot`_ +classes documentation. + +.. _Qt Signals and Slots: https://doc.qt.io/qt-5/signalsandslots.html +.. _libcamera Signal and Slot: http://libcamera.org/api-html/classlibcamera_1_1Signal.html#details + +In order to notify applications about the availability of new frames and data, +the ``Camera`` device exposes two ``Signals`` which applications can connect to +be notified of frame completion events. The ``bufferComplete`` signal serves to +report to applications the completion event of a single ``Stream`` part of a +``Request``, while the ``requestComplete`` signal notifies the completion of all +the ``Streams`` and data submitted as part of a request. This mechanism allows +implementation of partial request completion, which allows an application to +inspect completed buffers associated with the single streams without waiting for +all of them to be ready. + +The ``bufferComplete`` and ``requestComplete`` signals are emitted by the +``Camera`` device upon notifications received from the pipeline handler, which +tracks the buffers and request completion status. + +The single buffer completion notification is implemented by pipeline handlers by +`connecting`_ the ``bufferReady`` signal of the capture devices they have queued +buffers to, to a member function slot that handles processing of the completed +frames. When a buffer is ready, the pipeline handler must propagate the +completion of that buffer to the Camera by using the PipelineHandler base class +``completeBuffer`` function. When all of the buffers referenced by a ``Request`` +have been completed, the pipeline handler must again notify the ``Camera`` using +the PipelineHandler base class ``completeRequest`` function. The PipelineHandler +class implementation makes sure the request completion notifications are +delivered to applications in the same order as they have been submitted. + +.. _connecting: http://libcamera.org/api-html/classlibcamera_1_1Signal.html#aa04db72d5b3091ffbb4920565aeed382 + +Returning to the ``int VividCameraData::init()`` method, add the following above +the closing ``return 0;`` to connects the pipeline handler ``bufferReady`` +method to the V4L2 device buffer signal. + +.. code-block:: cpp + + video_->bufferReady.connect(this, &VividCameraData::bufferReady); + +Create the matching ``VividCameraData::bufferReady`` method after your +VividCameradata::init() impelementation. + +The ``bufferReady`` method obtains the request from the buffer using the +``request`` method, and notifies the ``Camera`` that the buffer and +request are completed. In this simpler pipeline handler, there is only one +stream, so it completes the request immediately. You can find a more complex +example of event handling with supporting multiple streams in the libcamera +code-base. + +.. TODO: Add link + +.. code-block:: cpp + + void VividCameraData::bufferReady(FrameBuffer *buffer) + { + Request *request = buffer->request(); + + pipe_->completeBuffer(camera_, request, buffer); + pipe_->completeRequest(camera_, request); + } + +Testing a pipeline handler +~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Once you've built the pipeline handler, we ca rebuild the code base, and test +capture through the pipeline through both of the cam and qcam utilities. + +.. code-block:: shell + + ninja -C build + ./build/src/cam/cam -c vivid -C 5 + +To test that the pipeline handler can detect a device, and capture input. + +Running the command above outputs (a lot of) information about pixel formats, +and then starts capturing frame data, and should provide an output such as the +following: + +.. code-block:: none + + user@dev:/home/libcamera$ ./build/src/cam/cam -c vivid -C5 + [42:34:08.573066847] [186470] INFO IPAManager ipa_manager.cpp:136 libcamera is not installed. Adding '/home/libcamera/build/src/ipa' to the IPA search path + [42:34:08.575908115] [186470] INFO Camera camera_manager.cpp:287 libcamera v0.0.11+876-7b27d262 + [42:34:08.610334268] [186471] INFO IPAProxy ipa_proxy.cpp:122 libcamera is not installed. Loading IPA configuration from '/home/libcamera/src/ipa/vimc/data' + Using camera vivid + [42:34:08.618462130] [186470] WARN V4L2 v4l2_pixelformat.cpp:176 Unsupported V4L2 pixel format Y10 + ... + [42:34:08.619901297] [186470] INFO Camera camera.cpp:793 configuring streams: (0) 1280x720-BGR888 + Capture 5 frames + fps: 0.00 stream0 seq: 000000 bytesused: 2764800 + fps: 4.98 stream0 seq: 000001 bytesused: 2764800 + fps: 5.00 stream0 seq: 000002 bytesused: 2764800 + fps: 5.03 stream0 seq: 000003 bytesused: 2764800 + fps: 5.03 stream0 seq: 000004 bytesused: 2764800 + +This demonstrates that the pipeline handler is successfully capturing frames, +but it is helpful to see the visual output and validate the images are being +processed correctly. The libcamera project also implements a Qt based +application which will render the frames in a window for visual inspection: + +.. code-block:: shell + + ./build/src/qcam/qcam -c vivid + +.. TODO: Running qcam with the vivid pipeline handler appears to have a bug and + no visual frames are seen. However disabling zero-copy on qcam renders + them successfully. \ No newline at end of file diff --git a/Documentation/index.rst b/Documentation/index.rst index cfcfd388699b..fb391d2b6ebf 100644 --- a/Documentation/index.rst +++ b/Documentation/index.rst @@ -14,3 +14,4 @@ Contribute Developer Guide + Pipeline Handler Writers Guide diff --git a/Documentation/meson.build b/Documentation/meson.build index dd7ae700af11..9f6c67071da9 100644 --- a/Documentation/meson.build +++ b/Documentation/meson.build @@ -53,6 +53,7 @@ if sphinx.found() 'docs.rst', 'index.rst', 'guides/introduction.rst', + 'guides/pipeline-handler.rst', ] release = 'release=v' + libcamera_git_version From patchwork Thu Aug 20 13:47:51 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: Kieran Bingham X-Patchwork-Id: 9340 Return-Path: X-Original-To: parsemail@patchwork.libcamera.org Delivered-To: parsemail@patchwork.libcamera.org Received: from lancelot.ideasonboard.com (lancelot.ideasonboard.com [92.243.16.209]) by patchwork.libcamera.org (Postfix) with ESMTPS id D6CFBBE173 for ; Thu, 20 Aug 2020 13:48:01 +0000 (UTC) Received: from lancelot.ideasonboard.com (localhost [IPv6:::1]) by lancelot.ideasonboard.com (Postfix) with ESMTP id 7FDA6620B6; Thu, 20 Aug 2020 15:48:01 +0200 (CEST) Authentication-Results: lancelot.ideasonboard.com; dkim=fail reason="signature verification failed" (1024-bit key; unprotected) header.d=ideasonboard.com header.i=@ideasonboard.com header.b="vg9cTgi5"; dkim-atps=neutral Received: from perceval.ideasonboard.com (perceval.ideasonboard.com [213.167.242.64]) by lancelot.ideasonboard.com (Postfix) with ESMTPS id 28E1661FA6 for ; Thu, 20 Aug 2020 15:47:58 +0200 (CEST) Received: from Q.local (cpc89244-aztw30-2-0-cust3082.18-1.cable.virginm.net [86.31.172.11]) by perceval.ideasonboard.com (Postfix) with ESMTPSA id A1BADA62; Thu, 20 Aug 2020 15:47:57 +0200 (CEST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=ideasonboard.com; s=mail; t=1597931277; bh=sd2oHITHJO3VHB5TPNUb4i3aOoLqljXqApNRN35sCUQ=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=vg9cTgi5ZY7Ykkbd74Usa8zFMBaxTAbjAt/PIRQIB0DsjoevT36t7jl8oAe0y5Yho 0T+nHDfpiAiBR3QNTgN9DkDAxRJgRCRPtiFTOENqiByMSq0MVHVYKoH6QDXCmobRw9 1Dp1wvb8nsJu2DEYKUVWU/KGqWcZotl6gDLQ2ttk= From: Kieran Bingham To: libcamera devel , Chris Ward , Jacopo Mondi Date: Thu, 20 Aug 2020 14:47:51 +0100 Message-Id: <20200820134751.278033-4-kieran.bingham@ideasonboard.com> X-Mailer: git-send-email 2.25.1 In-Reply-To: <20200820134751.278033-1-kieran.bingham@ideasonboard.com> References: <20200820134751.278033-1-kieran.bingham@ideasonboard.com> MIME-Version: 1.0 Subject: [libcamera-devel] [PATCH v4 3/3] Documentation: Guides: Application Writers Guide X-BeenThere: libcamera-devel@lists.libcamera.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: libcamera-devel-bounces@lists.libcamera.org Sender: "libcamera-devel" From: Chris Chinchilla Provide a tutorial and walk through guide for writing an applications with libcamera. Signed-off-by: Chris Chinchilla [Reflow/Rework, update to mainline API] Signed-off-by: Jacopo Mondi [Further reworks and review] Signed-off-by: Kieran Bingham Acked-by: Laurent Pinchart --- .../guides/application-developer.rst | 644 ++++++++++++++++++ Documentation/index.rst | 1 + Documentation/meson.build | 1 + 3 files changed, 646 insertions(+) create mode 100644 Documentation/guides/application-developer.rst diff --git a/Documentation/guides/application-developer.rst b/Documentation/guides/application-developer.rst new file mode 100644 index 000000000000..14ab21550cbe --- /dev/null +++ b/Documentation/guides/application-developer.rst @@ -0,0 +1,644 @@ +.. SPDX-License-Identifier: CC-BY-SA-4.0 + +Using libcamera in a C++ application +==================================== + +This tutorial shows how to create a C++ application that uses libcamera to +interface with a camera on a system, capture frames from it for 3 seconds, and +write metadata about the frames to standard out. + +.. TODO: Check how much of the example code runs before camera start etc? + +Application skeleton +-------------------- + +Most of the code in this tutorial runs in the ``int main()`` function +with a separate global function to handle events. The two functions need +to share data, which are stored in global variables for simplicity. A +production-ready application would organize the various objects created +in classes, and the event handler would be a class member function to +provide context data without requiring global variables. + +Use the following code snippets as the initial application skeleton. +It already lists all the necessary includes directives and instructs the +compiler to use the libcamera namespace, which gives access to the libcamera +defined names and types without the need of prefixing them. + +.. code:: cpp + + #include + #include + #include + + #include + + using namespace libcamera; + + int main() + { + // Code to follow + + return 0; + } + +Camera Manager +-------------- + +Every libcamera-based application needs an instance of a `CameraManager`_ that +runs for the life of the application. When the Camera Manager starts, it +enumerates all the cameras detected in the system. Behind the scenes, libcamera +abstracts and manages the complex pipelines that kernel drivers expose through +the `Linux Media Controller`_ and `Video for Linux`_ (V4L2) APIs, meaning that +an application doesn’t need to handle device or driver specific details. + +.. _CameraManager: http://libcamera.org/api-html/classlibcamera_1_1CameraManager.html +.. _Linux Media Controller: https://www.kernel.org/doc/html/latest/media/uapi/mediactl/media-controller-intro.html +.. _Video for Linux: https://www.linuxtv.org/docs.php + +Before the ``int main()`` function, create a global shared pointer +variable for the camera to support the event call back later: + +.. code:: cpp + + std::shared_ptr camera; + +Create a Camera Manager instance at the beginning of the main function, and then +start it. An application should only create a single Camera Manager instance. + +.. code:: cpp + + CameraManager *cm = new CameraManager(); + cm->start(); + +During the application initialization, the Camera Manager is started to +enumerate all the supported devices and create cameras that the application can +interact with. + +Once the camera manager is started, we can use it to iterate the available +cameras in the system: + +.. code:: cpp + + for (auto const &camera : cm->cameras()) + std::cout << camera->id() << std::endl; + +Printing the camera id lists the machine-readable unique identifiers, so for +example, the output on a Linux machine with a connected USB webcam is +``\_SB_.PCI0.XHC_.RHUB.HS08-8:1.0-5986:2115``. + +What libcamera considers a camera +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +The libcamera library considers any unique source of video frames, which usually +correspond to a camera sensor, as a single camera device. Camera devices expose +streams, which are obtained by processing data from the single image source and +all share some basic properties such as the frame duration and the image +exposure time, as they only depend by the image source configuration. + +Applications select one or multiple Camera devices they wish to operate on, and +require frames from at least one of their Streams. + +Create and acquire a camera +--------------------------- + +This example application uses a single camera (the first enumerated one) that +the Camera Manager reports as available to applications. + +Camera devices are stored by the CameraManager in a list accessible by index, or +can be retrieved by name through the ``CameraManager::get()`` function. The +code below retrieves the name of the first available camera and gets the camera +by name from the Camera Manager. + +.. code:: cpp + + std::string cameraId = cm->cameras()[0]->id(); + camera = cm->get(cameraId); + + /* + * Note that is equivalent to: + * camera = cm->cameras()[0]; + */ + +Once a camera has been selected an application needs to acquire an exclusive +lock to it so no other application can use it. + +.. code:: cpp + + camera->acquire(); + +Configure the camera +-------------------- + +Before the application can do anything with the camera, it needs to configure +the image format and sizes of the streams it wants to capture frames from. + +Stream configurations are represented by instances of the +``StreamConfiguration`` class, which are grouped together in a +``CameraConfiguration`` object. Before an application can start setting its +desired configuration, a ``CameraConfiguration`` instance needs to be generated +from the ``Camera`` device using the ``Camera::generateConfiguration()`` +function. + +The libcamera library uses the ``StreamRole`` enumeration to define predefined +ways an application intends to use a camera. The +``Camera::generateConfiguration()`` function accepts a list of desired roles and +generates a ``CameraConfiguration`` with the best stream parameters +configuration for each of the requested roles. If the camera can handle the +requested roles, it returns an initialized ``CameraConfiguration`` and a null +pointer if it can't. + +It is possible for applications to generate an empty ``CameraConfiguration`` +instance by not providing any role. The desired configuration will have to be +filled-in manually and manually validated. + +In the example application, create a new configuration variable and use the +``Camera::generateConfiguration`` function to produce a ``CameraConfiguration`` +for the single ``StreamRole::Viewfinder`` role. + +.. code:: cpp + + std::unique_ptr config = camera->generateConfiguration( { StreamRole::Viewfinder } ); + +The generated ``CameraConfiguration`` has a ``StreamConfiguration`` instance for +each ``StreamRole`` the application requested. Each of these has a default size +and format that the camera assigned, and a list of supported pixel formats and +sizes. + +The code below accesses the first and only ``StreamConfiguration`` item in the +``CameraConfiguration`` and outputs its parameters to standard output. + +.. code:: cpp + + StreamConfiguration &streamConfig = config->at(0); + std::cout << "Default viewfinder configuration is: " << streamConfig.toString() << std::endl; + +This is expected to output something like: + + ``Default viewfinder configuration is: 1280x720-MJPEG`` + +Change and validate the configuration +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +With an initialized ``CameraConfiguration``, an application can make changes to +the parameters it contains, for example, to change the width and height, use the +following code: + +.. code:: cpp + + streamConfig.size.width = 640; + streamConfig.size.height = 480; + +If an application changes any parameters, it must validate the configuration +before applying it to the camera using the ``CameraConfiguration::validate()`` +function. If the new values are not supported by the ``Camera`` device, the +validation process adjusts the parameters to what it considers to be the closest +supported values. + +The ``validate`` method returns a `Status`_ which applications shall check to +see if the Pipeline Handler adjusted the configuration. + +.. _Status: http://libcamera.org/api-html/classlibcamera_1_1CameraConfiguration.html#a64163f21db2fe1ce0a6af5a6f6847744 + +For example, the code above set the width and height to 640x480, but if the +camera cannot produce an image that large, it might adjust the configuration to +the supported size of 320x240 and return ``Adjusted`` as validation status +result. + +If the configuration to validate cannot be adjusted to a set of supported +values, the validation procedure fails and returns the ``Invalid`` status. + +For this example application, the code below prints the adjusted values to +standard out. + +.. code:: cpp + + config->validate(); + std::cout << "Validated viewfinder configuration is: " << streamConfig.toString() << std::endl; + +For example, the output might be something like + + ``Validated viewfinder configuration is: 320x240-MJPEG`` + +A validated ``CameraConfiguration`` can bet given to the ``Camera`` device to be +applied to the system. + +.. code:: cpp + + camera->configure(config.get()); + +If an application doesn’t first validate the configuration before calling +``Camera::configure()``, there’s a chance that calling the function can fail, if +the given configuration would have to be adjusted. + +Allocate FrameBuffers +--------------------- + +An application needs to reserve the memory that libcamera can write incoming +frames and data to, and that the application can then read. The libcamera +library uses ``FrameBuffer`` instances to represent memory buffers allocated in +memory. An application should reserve enough memory for the frame size the +streams need based on the configured image sizes and formats. + +The libcamera library consumes buffers provided by applications as +``FrameBuffer`` instances, which makes libcamera a consumer of buffers exported +by other devices (such as displays or video encoders), or allocated from an +external allocator (such as ION on Android). + +In some situations, applications do not have any means to allocate or get hold +of suitable buffers, for instance, when no other device is involved, or on Linux +platforms that lack a centralized allocator. The ``FrameBufferAllocator`` class +provides a buffer allocator an application can use in these situations. + +An application doesn’t have to use the default ``FrameBufferAllocator`` that +libcamera provides. It can instead allocate memory manually and pass the buffers +in ``Request``\s (read more about ``Request`` in `the frame capture section +<#frame-capture>`_ of this guide). The example in this guide covers using the +``FrameBufferAllocator`` that libcamera provides. + +Using the libcamera ``FrameBufferAllocator`` +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Applications create a ``FrameBufferAllocator`` for a Camera and use it +to allocate buffers for streams of a ``CameraConfiguration`` with the +``allocate()`` function. + +The list of allocated buffers can be retrieved using the ``Stream`` instance +as the parameter of the ``FrameBufferAllocator::buffers()`` function. + +.. code:: cpp + + FrameBufferAllocator *allocator = new FrameBufferAllocator(camera); + + for (StreamConfiguration &cfg : *config) { + int ret = allocator->allocate(cfg.stream()); + if (ret < 0) { + std::cerr << "Can't allocate buffers" << std::endl; + return -ENOMEM; + } + + unsigned int allocated = allocator->buffers(cfg.stream()).size(); + std::cout << "Allocated " << allocated << " buffers for stream" << std::endl; + } + +Frame Capture +~~~~~~~~~~~~~ + +The libcamera library implements a streaming model based on per-frame requests. +For each frame an application wants to capture it must queue a request for it to +the camera. With libcamera, a ``Request`` is at least one ``Stream`` associated +with a ``FrameBuffer`` representing the memory location where frames have to be +stored. + +First, by using the ``Stream`` instance associated to each +``StreamConfiguration``, retrieve the list of ``FrameBuffer``\s created for it +using the frame allocator. Then create a vector of requests to be submitted to +the camera. + +.. code:: cpp + + Stream *stream = streamConfig.stream(); + const std::vector> &buffers = allocator->buffers(stream); + std::vector requests; + +Proceed to fill the request vector by creating ``Request`` instances from the +camera device, and associate a buffer for each of them for the ``Stream``. + +.. code:: cpp + + for (unsigned int i = 0; i < buffers.size(); ++i) { + Request *request = camera->createRequest(); + if (!request) + { + std::cerr << "Can't create request" << std::endl; + return -ENOMEM; + } + + const std::unique_ptr &buffer = buffers[i]; + int ret = request->addBuffer(stream, buffer.get()); + if (ret < 0) + { + std::cerr << "Can't set buffer for request" + << std::endl; + return ret; + } + + requests.push_back(request); + } + +.. TODO: Controls + +.. TODO: A request can also have controls or parameters that you can apply to the image. + +Event handling and callbacks +---------------------------- + +The libcamera library uses the concept of `signals and slots` (similar to `Qt +Signals and Slots`_) to connect events with callbacks to handle them. + +.. _signals and slots: http://libcamera.org/api-html/classlibcamera_1_1Signal.html#details +.. _Qt Signals and Slots: https://doc.qt.io/qt-5/signalsandslots.html + +The ``Camera`` device emits two signals that applications can connect to in +order to execute callbacks on frame completion events. + +The ``Camera::bufferCompleted`` signal notifies applications that a buffer with +image data is available. Receiving notifications about the single buffer +completion event allows applications to implement partial request completion +support, and to inspect the buffer content before the request it is part of has +fully completed. + +The ``Camera::requestCompleted`` signal notifies applications that a request +has completed, which means all the buffers the request contains have now +completed. Request completion notifications are always emitted in the same order +as the requests have been queued to the camera. + +To receive the signals emission notifications, connect a slot function to the +signal to handle it in the application code. + +.. code:: cpp + + camera->requestCompleted.connect(requestComplete); + +For this example application, only the ``Camera::requestCompleted`` signal gets +handled and the matching ``requestComplete`` slot method outputs information +about the FrameBuffer to standard output. This callback is typically where an +application accesses the image data from the camera and does something with it. + +Signals operate in the libcamera ``CameraManager`` thread context, so it is +important not to block the thread for a long time, as this blocks internal +processing of the camera pipelines, and can affect realtime performance. + +Handle request completion events +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Create the ``requestComplete`` function by matching the slot signature: + +.. code:: cpp + + static void requestComplete(Request *request) + { + // Code to follow + } + +Request completion events can be emitted for requests which have been canceled, +for example, by unexpected application shutdown. To avoid an application +processing invalid image data, it’s worth checking that the request has +completed successfully. The list of request completion statuses is available in +the `Request::Status`_ class enum documentation. + +.. _Request::Status: https://www.libcamera.org/api-html/classlibcamera_1_1Request.html#a2209ba8d51af8167b25f6e3e94d5c45b + +.. code:: cpp + + if (request->status() == Request::RequestCancelled) + return; + +If the ``Request`` has completed successfully, applications can access the +completed buffers using the ``Request::buffers()`` function, which returns a map +of ``FrameBuffer`` instances associated with the ``Stream`` that produced the +images. + +.. code:: cpp + + const std::map &buffers = request->buffers(); + +Iterating through the map allows applications to inspect each completed buffer +in this request, and access the metadata associated to each frame. + +The metadata buffer contains information such the capture status, a timestamp, +and the bytes used, as described in the `FrameMetadata`_ documentation. + +.. _FrameMetaData: http://libcamera.org/api-html/structlibcamera_1_1FrameMetadata.html + +.. code:: cpp + + for (auto bufferPair : buffers) { + FrameBuffer *buffer = bufferPair.second; + const FrameMetadata &metadata = buffer->metadata(); + } + +For this example application, inside the ``for`` loop from above, we ca print +the Frame sequence number and details of the planes. + +.. code:: cpp + + std::cout << " seq: " << std::setw(6) << std::setfill('0') << metadata.sequence << " bytesused: "; + + unsigned int nplane = 0; + for (const FrameMetadata::Plane &plane : metadata.planes) + { + std::cout << plane.bytesused; + if (++nplane < metadata.planes.size()) std::cout << "/"; + } + + std::cout << std::endl; + +The expected output shows each monotonically increasing frame sequence number +and the bytes used by planes. + +.. code:: text + + seq: 000000 bytesused: 1843200 + seq: 000002 bytesused: 1843200 + seq: 000004 bytesused: 1843200 + seq: 000006 bytesused: 1843200 + seq: 000008 bytesused: 1843200 + seq: 000010 bytesused: 1843200 + seq: 000012 bytesused: 1843200 + seq: 000014 bytesused: 1843200 + seq: 000016 bytesused: 1843200 + seq: 000018 bytesused: 1843200 + seq: 000020 bytesused: 1843200 + seq: 000022 bytesused: 1843200 + seq: 000024 bytesused: 1843200 + seq: 000026 bytesused: 1843200 + seq: 000028 bytesused: 1843200 + seq: 000030 bytesused: 1843200 + seq: 000032 bytesused: 1843200 + seq: 000034 bytesused: 1843200 + seq: 000036 bytesused: 1843200 + seq: 000038 bytesused: 1843200 + seq: 000040 bytesused: 1843200 + seq: 000042 bytesused: 1843200 + +A completed buffer contains of course image data which can be accessed through +the per-plane dma-buf file descriptor transported by the ``FrameBuffer`` +instance. An example of how to write image data to disk is available in the +`BufferWriter class`_ which is a part of the ``cam`` utility application in the +libcamera repository. + +.. _BufferWriter class: https://git.linuxtv.org/libcamera.git/tree/src/cam/buffer_writer.cpp + +With the handling of this request completed, it is possible to re-use the +buffers by adding them to a new ``Request`` instance with their matching +streams, and finally, queue the new capture request to the camera device: + +.. code:: cpp + + request = camera->createRequest(); + if (!request) + { + std::cerr << "Can't create request" << std::endl; + return; + } + + for (auto it = buffers.begin(); it != buffers.end(); ++it) + { + Stream *stream = it->first; + FrameBuffer *buffer = it->second; + + request->addBuffer(stream, buffer); + } + + camera->queueRequest(request); + +Request queueing +---------------- + +The ``Camera`` device is now ready to receive frame capture requests and +actually start delivering frames. In order to prepare for that, an application +needs to first start the camera, and queue requests to it for them to be +processed. + +In the main() function, just after having connected the +``Camera::requestCompleted`` signal to the callback handler, start the camera +and queue all the previously created requests. + +.. code:: cpp + + camera->start(); + for (Request *request : requests) + camera->queueRequest(request); + +Start an event loop +~~~~~~~~~~~~~~~~~~~ + +The libcamera library needs an event loop to monitor and dispatch events +generated by the video devices part of the capture pipeline. Libcamera provides +its own ``EventDispatcher`` class (inspired by the `Qt event system`_) to +process and deliver events generated by ``EventNotifiers``. + +.. _Qt event system: https://doc.qt.io/qt-5/eventsandfilters.html + +The libcamera library implements this by creating instances of the +``EventNotifier`` class, which models a file descriptor event source registered +to an ``EventDispatcher``. Whenever the ``EventDispatcher`` detects an event on +a notifier it is monitoring, it emits the notifier's +``EventNotifier::activated`` signal. The libcamera components connect to the +notifiers' signals and emit application visible events, such as the +``Camera::bufferReady`` and ``Camera::requestCompleted`` signals. + +The code below retrieves a reference to the system-wide event dispatcher and for +the a fixed duration of 3 seconds, processes all the events detected in the +system. + +.. code:: cpp + + EventDispatcher *dispatcher = cm->eventDispatcher(); + Timer timer; + timer.start(3000); + while (timer.isRunning()) + dispatcher->processEvents(); + +Clean up and stop the application +--------------------------------- + +The application is now finished with the camera and the resources the camera +uses, so needs to do the following: + +- stop the camera +- free the buffers in the FrameBufferAllocator and delete it +- release the lock on the camera and reset the pointer to it +- stop the camera manager + +.. code:: cpp + + camera->stop(); + allocator->free(stream); + delete allocator; + camera->release(); + camera.reset(); + cm->stop(); + + return 0; + +Build and run instructions +-------------------------- + +To build the application, use the `Meson build system`_ which is also the +official build system of the libcamera library. + +Make sure both ``meson`` and ``libcamera`` are installed in your system. Please +refer to your distribution documentation to install meson and install the most +recent version of libcamera from the git repository at `Linux TV`_. You would +also need to install the ``pkg-config`` tool to correctly identify the +libcamera.so object install location in the system. + +.. _Meson build system: https://mesonbuild.com/ +.. _Linux TV: https://git.linuxtv.org/libcamera.git/ + +Dependencies +~~~~~~~~~~~~ + +The test application presented here depends on the libcamera library to be +available in a path that meson can identify. The libcamera install procedure +performed using the ``ninja install`` command may by default deploy the +libcamera components in the ``/usr/local/lib`` path, or a package manager may +install it to ``/usr/lib`` depending on your distribution. If meson is unable to +find the location of the libcamera installation, you may need to instruct meson +to look into a specific path when searching for ``libcamera.so`` by setting the +``PKG_CONFIG_PATH`` environment variable to the right location. + +Adjust the following command to use the ``pkgconfig`` directory where libcamera +has been installed in your system. + +.. code:: shell + + export PKG_CONFIG_PATH=/usr/local/lib/pkgconfig/ + +Verify that ``pkg-config`` can identify the ``camera`` library with + +.. code:: shell + + $ pkg-config --libs --cflags camera + -I/usr/local/include/libcamera -L/usr/local/lib -lcamera + +``meson`` can alternatively use ``cmake`` to locate packages, please refer to +the ``meson`` documentation if you prefer to use it in place of ``pkgconfig`` + +Build file +~~~~~~~~~~ + +With the dependencies correctly identified, prepare a ``meson.build`` build file +to be placed in the same directory where the application lives. You can +name your application as you like, but be sure to update the following snippet +accordingly. In this example, the application file has been named +``simple-cam.cpp``. + +.. code:: + + project('simple-cam', 'cpp') + + simpler_cam = executable('simple-cam', + 'simple-cam.cpp', + dependencies: dependency('camera', required : true)) + +The ``dependencies`` line instructs meson to ask ``pkgconfig`` (or ``cmake``) to +locate the ``camera`` library, (libcamera without the lib prefix) which the test +application will be dynamically linked against. + +With the build file in place, compile and run the application with: + +.. code:: shell + + $ meson build + $ cd build + $ ninja + $ ./simple-cam + +It is possible to increase the library debug output by using environment +variables which control the library log filtering system: + +.. code:: shell + + $ LIBCAMERA_LOG_LEVELS=0 ./simple-cam diff --git a/Documentation/index.rst b/Documentation/index.rst index fb391d2b6ebf..68b7ac06c506 100644 --- a/Documentation/index.rst +++ b/Documentation/index.rst @@ -15,3 +15,4 @@ Developer Guide Pipeline Handler Writers Guide + Application Writers Guide diff --git a/Documentation/meson.build b/Documentation/meson.build index 9f6c67071da9..6bff2e4a4912 100644 --- a/Documentation/meson.build +++ b/Documentation/meson.build @@ -54,6 +54,7 @@ if sphinx.found() 'index.rst', 'guides/introduction.rst', 'guides/pipeline-handler.rst', + 'guides/application-developer.rst', ] release = 'release=v' + libcamera_git_version