[v2,2/2] media: docs-rst: Document memory-to-memory video encoder interface
diff mbox series

Message ID 20181022144901.113852-3-tfiga@chromium.org
State New
Headers show
Series
  • Document memory-to-memory video codec interfaces
Related show

Commit Message

Tomasz Figa Oct. 22, 2018, 2:49 p.m. UTC
Due to complexity of the video encoding process, the V4L2 drivers of
stateful encoder hardware require specific sequences of V4L2 API calls
to be followed. These include capability enumeration, initialization,
encoding, encode parameters change, drain and reset.

Specifics of the above have been discussed during Media Workshops at
LinuxCon Europe 2012 in Barcelona and then later Embedded Linux
Conference Europe 2014 in Düsseldorf. The de facto Codec API that
originated at those events was later implemented by the drivers we already
have merged in mainline, such as s5p-mfc or coda.

The only thing missing was the real specification included as a part of
Linux Media documentation. Fix it now and document the encoder part of
the Codec API.

Signed-off-by: Tomasz Figa <tfiga@chromium.org>
---
 Documentation/media/uapi/v4l/dev-encoder.rst  | 579 ++++++++++++++++++
 Documentation/media/uapi/v4l/devices.rst      |   1 +
 Documentation/media/uapi/v4l/pixfmt-v4l2.rst  |   5 +
 Documentation/media/uapi/v4l/v4l2.rst         |   2 +
 .../media/uapi/v4l/vidioc-encoder-cmd.rst     |  38 +-
 5 files changed, 610 insertions(+), 15 deletions(-)
 create mode 100644 Documentation/media/uapi/v4l/dev-encoder.rst

Comments

Hans Verkuil Nov. 12, 2018, 1:23 p.m. UTC | #1
On 10/22/2018 04:49 PM, Tomasz Figa wrote:
> Due to complexity of the video encoding process, the V4L2 drivers of
> stateful encoder hardware require specific sequences of V4L2 API calls
> to be followed. These include capability enumeration, initialization,
> encoding, encode parameters change, drain and reset.
> 
> Specifics of the above have been discussed during Media Workshops at
> LinuxCon Europe 2012 in Barcelona and then later Embedded Linux
> Conference Europe 2014 in Düsseldorf. The de facto Codec API that
> originated at those events was later implemented by the drivers we already
> have merged in mainline, such as s5p-mfc or coda.
> 
> The only thing missing was the real specification included as a part of
> Linux Media documentation. Fix it now and document the encoder part of
> the Codec API.
> 
> Signed-off-by: Tomasz Figa <tfiga@chromium.org>
> ---
>  Documentation/media/uapi/v4l/dev-encoder.rst  | 579 ++++++++++++++++++
>  Documentation/media/uapi/v4l/devices.rst      |   1 +
>  Documentation/media/uapi/v4l/pixfmt-v4l2.rst  |   5 +
>  Documentation/media/uapi/v4l/v4l2.rst         |   2 +
>  .../media/uapi/v4l/vidioc-encoder-cmd.rst     |  38 +-
>  5 files changed, 610 insertions(+), 15 deletions(-)
>  create mode 100644 Documentation/media/uapi/v4l/dev-encoder.rst
> 
> diff --git a/Documentation/media/uapi/v4l/dev-encoder.rst b/Documentation/media/uapi/v4l/dev-encoder.rst
> new file mode 100644
> index 000000000000..41139e5e48eb
> --- /dev/null
> +++ b/Documentation/media/uapi/v4l/dev-encoder.rst
> @@ -0,0 +1,579 @@
> +.. -*- coding: utf-8; mode: rst -*-
> +
> +.. _encoder:
> +
> +*************************************************
> +Memory-to-memory Stateful Video Encoder Interface
> +*************************************************
> +
> +A stateful video encoder takes raw video frames in display order and encodes
> +them into a bitstream. It generates complete chunks of the bitstream, including
> +all metadata, headers, etc. The resulting bitstream does not require any
> +further post-processing by the client.
> +
> +Performing software stream processing, header generation etc. in the driver
> +in order to support this interface is strongly discouraged. In case such
> +operations are needed, use of the Stateless Video Encoder Interface (in
> +development) is strongly advised.
> +
> +Conventions and notation used in this document
> +==============================================
> +
> +1. The general V4L2 API rules apply if not specified in this document
> +   otherwise.
> +
> +2. The meaning of words "must", "may", "should", etc. is as per RFC
> +   2119.
> +
> +3. All steps not marked "optional" are required.
> +
> +4. :c:func:`VIDIOC_G_EXT_CTRLS`, :c:func:`VIDIOC_S_EXT_CTRLS` may be used
> +   interchangeably with :c:func:`VIDIOC_G_CTRL`, :c:func:`VIDIOC_S_CTRL`,
> +   unless specified otherwise.
> +
> +5. Single-plane API (see spec) and applicable structures may be used
> +   interchangeably with Multi-plane API, unless specified otherwise,
> +   depending on encoder capabilities and following the general V4L2
> +   guidelines.
> +
> +6. i = [a..b]: sequence of integers from a to b, inclusive, i.e. i =
> +   [0..2]: i = 0, 1, 2.
> +
> +7. Given an ``OUTPUT`` buffer A, A' represents a buffer on the ``CAPTURE``
> +   queue containing data (encoded frame/stream) that resulted from processing
> +   buffer A.

The same comments as I mentioned for the previous patch apply to this section.

> +
> +Glossary
> +========
> +
> +Refer to :ref:`decoder-glossary`.

Ah, you refer to the same glossary. Then my comment about the source resolution
terms is obviously wrong.

I wonder if it wouldn't be better to split off the sections above into a separate
HW codec intro section where you explain the differences between stateful/stateless
encoders and decoders, and add the conventions and glossary.

After that you have the three documents for each variant (later four when we get
stateless encoders).

Up to you, and it can be done later in a follow-up patch.

> +
> +State machine
> +=============
> +
> +.. kernel-render:: DOT
> +   :alt: DOT digraph of encoder state machine
> +   :caption: Encoder state machine
> +
> +   digraph encoder_state_machine {
> +       node [shape = doublecircle, label="Encoding"] Encoding;
> +
> +       node [shape = circle, label="Initialization"] Initialization;
> +       node [shape = circle, label="Stopped"] Stopped;
> +       node [shape = circle, label="Drain"] Drain;
> +       node [shape = circle, label="Reset"] Reset;
> +
> +       node [shape = point]; qi
> +       qi -> Initialization [ label = "open()" ];
> +
> +       Initialization -> Encoding [ label = "Both queues streaming" ];
> +
> +       Encoding -> Drain [ label = "V4L2_DEC_CMD_STOP" ];
> +       Encoding -> Reset [ label = "VIDIOC_STREAMOFF(CAPTURE)" ];
> +       Encoding -> Stopped [ label = "VIDIOC_STREAMOFF(OUTPUT)" ];
> +       Encoding -> Encoding;
> +
> +       Drain -> Stopped [ label = "All CAPTURE\nbuffers dequeued\nor\nVIDIOC_STREAMOFF(CAPTURE)" ];
> +       Drain -> Reset [ label = "VIDIOC_STREAMOFF(CAPTURE)" ];
> +
> +       Reset -> Encoding [ label = "VIDIOC_STREAMON(CAPTURE)" ];
> +       Reset -> Initialization [ label = "VIDIOC_REQBUFS(OUTPUT, 0)" ];
> +
> +       Stopped -> Encoding [ label = "V4L2_DEC_CMD_START\nor\nVIDIOC_STREAMON(OUTPUT)" ];
> +       Stopped -> Reset [ label = "VIDIOC_STREAMOFF(CAPTURE)" ];
> +   }
> +
> +Querying capabilities
> +=====================
> +
> +1. To enumerate the set of coded formats supported by the encoder, the
> +   client may call :c:func:`VIDIOC_ENUM_FMT` on ``CAPTURE``.
> +
> +   * The full set of supported formats will be returned, regardless of the
> +     format set on ``OUTPUT``.
> +
> +2. To enumerate the set of supported raw formats, the client may call
> +   :c:func:`VIDIOC_ENUM_FMT` on ``OUTPUT``.
> +
> +   * Only the formats supported for the format currently active on ``CAPTURE``
> +     will be returned.
> +
> +   * In order to enumerate raw formats supported by a given coded format,
> +     the client must first set that coded format on ``CAPTURE`` and then
> +     enumerate the formats on ``OUTPUT``.
> +
> +3. The client may use :c:func:`VIDIOC_ENUM_FRAMESIZES` to detect supported
> +   resolutions for a given format, passing desired pixel format in
> +   :c:type:`v4l2_frmsizeenum` ``pixel_format``.
> +
> +   * Values returned by :c:func:`VIDIOC_ENUM_FRAMESIZES` for a coded pixel
> +     format will include all possible coded resolutions supported by the
> +     encoder for given coded pixel format.
> +
> +   * Values returned by :c:func:`VIDIOC_ENUM_FRAMESIZES` for a raw pixel format
> +     will include all possible frame buffer resolutions supported by the
> +     encoder for given raw pixel format and coded format currently set on
> +     ``CAPTURE``.
> +
> +4. Supported profiles and levels for given format, if applicable, may be

format -> the coded format currently set on ``CAPTURE``

> +   queried using their respective controls via :c:func:`VIDIOC_QUERYCTRL`.
> +
> +5. Any additional encoder capabilities may be discovered by querying
> +   their respective controls.
> +
> +Initialization
> +==============
> +
> +1. **Optional.** Enumerate supported formats and resolutions. See
> +   `Querying capabilities` above.

Can be dropped IMHO.

> +
> +2. Set a coded format on the ``CAPTURE`` queue via :c:func:`VIDIOC_S_FMT`
> +
> +   * **Required fields:**
> +
> +     ``type``
> +         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``CAPTURE``
> +
> +     ``pixelformat``
> +         the coded format to be produced
> +
> +     ``sizeimage``
> +         desired size of ``CAPTURE`` buffers; the encoder may adjust it to
> +         match hardware requirements
> +
> +     other fields
> +         follow standard semantics
> +
> +   * **Return fields:**
> +
> +     ``sizeimage``
> +         adjusted size of ``CAPTURE`` buffers
> +
> +   .. warning::
> +
> +      Changing the ``CAPTURE`` format may change the currently set ``OUTPUT``
> +      format. The encoder will derive a new ``OUTPUT`` format from the
> +      ``CAPTURE`` format being set, including resolution, colorimetry
> +      parameters, etc. If the client needs a specific ``OUTPUT`` format, it
> +      must adjust it afterwards.
> +
> +3. **Optional.** Enumerate supported ``OUTPUT`` formats (raw formats for
> +   source) for the selected coded format via :c:func:`VIDIOC_ENUM_FMT`.

Does this return the same set of formats as in the 'Querying Capabilities' phase?

> +
> +   * **Required fields:**
> +
> +     ``type``
> +         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``OUTPUT``
> +
> +     other fields
> +         follow standard semantics
> +
> +   * **Return fields:**
> +
> +     ``pixelformat``
> +         raw format supported for the coded format currently selected on
> +         the ``OUTPUT`` queue.

OUTPUT -> CAPTURE

> +
> +     other fields
> +         follow standard semantics
> +
> +4. Set the raw source format on the ``OUTPUT`` queue via
> +   :c:func:`VIDIOC_S_FMT`.
> +
> +   * **Required fields:**
> +
> +     ``type``
> +         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``OUTPUT``
> +
> +     ``pixelformat``
> +         raw format of the source
> +
> +     ``width``, ``height``
> +         source resolution
> +
> +     other fields
> +         follow standard semantics
> +
> +   * **Return fields:**
> +
> +     ``width``, ``height``
> +         may be adjusted by encoder to match alignment requirements, as
> +         required by the currently selected formats
> +
> +     other fields
> +         follow standard semantics
> +
> +   * Setting the source resolution will reset the selection rectangles to their
> +     default values, based on the new resolution, as described in the step 5
> +     below.
> +
> +5. **Optional.** Set the visible resolution for the stream metadata via
> +   :c:func:`VIDIOC_S_SELECTION` on the ``OUTPUT`` queue.
> +
> +   * **Required fields:**
> +
> +     ``type``
> +         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``OUTPUT``
> +
> +     ``target``
> +         set to ``V4L2_SEL_TGT_CROP``
> +
> +     ``r.left``, ``r.top``, ``r.width``, ``r.height``
> +         visible rectangle; this must fit within the `V4L2_SEL_TGT_CROP_BOUNDS`
> +         rectangle and may be subject to adjustment to match codec and
> +         hardware constraints
> +
> +   * **Return fields:**
> +
> +     ``r.left``, ``r.top``, ``r.width``, ``r.height``
> +         visible rectangle adjusted by the encoder
> +
> +   * The following selection targets are supported on ``OUTPUT``:
> +
> +     ``V4L2_SEL_TGT_CROP_BOUNDS``
> +         equal to the full source frame, matching the active ``OUTPUT``
> +         format
> +
> +     ``V4L2_SEL_TGT_CROP_DEFAULT``
> +         equal to ``V4L2_SEL_TGT_CROP_BOUNDS``
> +
> +     ``V4L2_SEL_TGT_CROP``
> +         rectangle within the source buffer to be encoded into the
> +         ``CAPTURE`` stream; defaults to ``V4L2_SEL_TGT_CROP_DEFAULT``

Since this defaults to the CROP_DEFAULT rectangle this means that if you have
a 16x16 macroblock size and you want to encode 1080p, you will always have to
explicitly set the CROP rectangle to 1920x1080, right? Since the default will
be 1088 instead of 1080.

It is probably wise to explicitly mention this.

> +
> +     ``V4L2_SEL_TGT_COMPOSE_BOUNDS``
> +         maximum rectangle within the coded resolution, which the cropped
> +         source frame can be output into; if the hardware does not support

output -> composed

> +         composition or scaling, then this is always equal to the rectangle of
> +         width and height matching ``V4L2_SEL_TGT_CROP`` and located at (0, 0)
> +
> +     ``V4L2_SEL_TGT_COMPOSE_DEFAULT``
> +         equal to a rectangle of width and height matching
> +         ``V4L2_SEL_TGT_CROP`` and located at (0, 0)
> +
> +     ``V4L2_SEL_TGT_COMPOSE``
> +         rectangle within the coded frame, which the cropped source frame
> +         is to be output into; defaults to

output -> composed

> +         ``V4L2_SEL_TGT_COMPOSE_DEFAULT``; read-only on hardware without
> +         additional compose/scaling capabilities; resulting stream will
> +         have this rectangle encoded as the visible rectangle in its
> +         metadata
> +
> +   .. warning::
> +
> +      The encoder may adjust the crop/compose rectangles to the nearest
> +      supported ones to meet codec and hardware requirements. The client needs
> +      to check the adjusted rectangle returned by :c:func:`VIDIOC_S_SELECTION`.
> +
> +6. Allocate buffers for both ``OUTPUT`` and ``CAPTURE`` via
> +   :c:func:`VIDIOC_REQBUFS`. This may be performed in any order.
> +
> +   * **Required fields:**
> +
> +     ``count``
> +         requested number of buffers to allocate; greater than zero
> +
> +     ``type``
> +         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``OUTPUT`` or
> +         ``CAPTURE``
> +
> +     other fields
> +         follow standard semantics
> +
> +   * **Return fields:**
> +
> +     ``count``
> +          actual number of buffers allocated
> +
> +   .. warning::
> +
> +      The actual number of allocated buffers may differ from the ``count``
> +      given. The client must check the updated value of ``count`` after the
> +      call returns.
> +
> +   .. note::
> +
> +      To allocate more than the minimum number of buffers (for pipeline depth),
> +      the client may query the ``V4L2_CID_MIN_BUFFERS_FOR_OUTPUT`` or
> +      ``V4L2_CID_MIN_BUFFERS_FOR_CAPTURE`` control respectively, to get the
> +      minimum number of buffers required by the encoder/format, and pass the
> +      obtained value plus the number of additional buffers needed in the
> +      ``count`` field to :c:func:`VIDIOC_REQBUFS`.

Does V4L2_CID_MIN_BUFFERS_FOR_CAPTURE make any sense for encoders?

V4L2_CID_MIN_BUFFERS_FOR_OUTPUT can make sense depending on GOP size etc.

> +
> +   Alternatively, :c:func:`VIDIOC_CREATE_BUFS` can be used to have more
> +   control over buffer allocation.
> +
> +   * **Required fields:**
> +
> +     ``count``
> +         requested number of buffers to allocate; greater than zero
> +
> +     ``type``
> +         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``OUTPUT``
> +
> +     other fields
> +         follow standard semantics
> +
> +   * **Return fields:**
> +
> +     ``count``
> +         adjusted to the number of allocated buffers
> +
> +7. Begin streaming on both ``OUTPUT`` and ``CAPTURE`` queues via
> +   :c:func:`VIDIOC_STREAMON`. This may be performed in any order. The actual
> +   encoding process starts when both queues start streaming.
> +
> +.. note::
> +
> +   If the client stops the ``CAPTURE`` queue during the encode process and then
> +   restarts it again, the encoder will begin generating a stream independent
> +   from the stream generated before the stop. The exact constraints depend
> +   on the coded format, but may include the following implications:
> +
> +   * encoded frames produced after the restart must not reference any
> +     frames produced before the stop, e.g. no long term references for
> +     H.264,
> +
> +   * any headers that must be included in a standalone stream must be
> +     produced again, e.g. SPS and PPS for H.264.
> +
> +Encoding
> +========
> +
> +This state is reached after the `Initialization` sequence finishes succesfully.

successfully

> +In this state, client queues and dequeues buffers to both queues via

client -> the client

> +:c:func:`VIDIOC_QBUF` and :c:func:`VIDIOC_DQBUF`, following the standard
> +semantics.
> +
> +The contents of encoded ``CAPTURE`` buffers depend on the active coded pixel

contents ... depend -> content ... depends

> +format and may be affected by codec-specific extended controls, as stated
> +in the documentation of each format.
> +
> +Both queues operate independently, following standard behavior of V4L2 buffer
> +queues and memory-to-memory devices. In addition, the order of encoded frames
> +dequeued from the ``CAPTURE`` queue may differ from the order of queuing raw
> +frames to the ``OUTPUT`` queue, due to properties of the selected coded format,
> +e.g. frame reordering.
> +
> +The client must not assume any direct relationship between ``CAPTURE`` and
> +``OUTPUT`` buffers and any specific timing of buffers becoming
> +available to dequeue. Specifically,
> +
> +* a buffer queued to ``OUTPUT`` may result in more than 1 buffer produced on
> +  ``CAPTURE`` (if returning an encoded frame allowed the encoder to return a
> +  frame that preceded it in display, but succeeded it in the decode order),
> +
> +* a buffer queued to ``OUTPUT`` may result in a buffer being produced on
> +  ``CAPTURE`` later into encode process, and/or after processing further
> +  ``OUTPUT`` buffers, or be returned out of order, e.g. if display
> +  reordering is used,
> +
> +* buffers may become available on the ``CAPTURE`` queue without additional
> +  buffers queued to ``OUTPUT`` (e.g. during drain or ``EOS``), because of the
> +  ``OUTPUT`` buffers queued in the past whose decoding results are only
> +  available at later time, due to specifics of the decoding process,
> +
> +* buffers queued to ``OUTPUT`` may not become available to dequeue instantly
> +  after being encoded into a corresponding ``CATPURE`` buffer, e.g. if the
> +  encoder needs to use the frame as a reference for encoding further frames.
> +
> +.. note::
> +
> +   To allow matching encoded ``CAPTURE`` buffers with ``OUTPUT`` buffers they
> +   originated from, the client can set the ``timestamp`` field of the
> +   :c:type:`v4l2_buffer` struct when queuing an ``OUTPUT`` buffer. The
> +   ``CAPTURE`` buffer(s), which resulted from encoding that ``OUTPUT`` buffer
> +   will have their ``timestamp`` field set to the same value when dequeued.
> +
> +   In addition to the straighforward case of one ``OUTPUT`` buffer producing
> +   one ``CAPTURE`` buffer, the following cases are defined:
> +
> +   * one ``OUTPUT`` buffer generates multiple ``CAPTURE`` buffers: the same
> +     ``OUTPUT`` timestamp will be copied to multiple ``CAPTURE`` buffers,
> +
> +   * the encoding order differs from the presentation order (i.e. the
> +     ``CAPTURE`` buffers are out-of-order compared to the ``OUTPUT`` buffers):
> +     ``CAPTURE`` timestamps will not retain the order of ``OUTPUT`` timestamps
> +     and thus monotonicity of the timestamps cannot be guaranteed.
> +
> +.. note::
> +
> +   To let the client distinguish between frame types (keyframes, intermediate
> +   frames; the exact list of types depends on the coded format), the
> +   ``CAPTURE`` buffers will have corresponding flag bits set in their
> +   :c:type:`v4l2_buffer` struct when dequeued. See the documentation of
> +   :c:type:`v4l2_buffer` and each coded pixel format for exact list of flags
> +   and their meanings.

Is this required? (I think it should, but it isn't the case today).

Is the current set of buffer flags (Key/B/P frame) sufficient for the current
set of codecs?

> +
> +Encoding parameter changes
> +==========================
> +
> +The client is allowed to use :c:func:`VIDIOC_S_CTRL` to change encoder
> +parameters at any time. The availability of parameters is encoder-specific
> +and the client must query the encoder to find the set of available controls.
> +
> +The ability to change each parameter during encoding is encoder-specific, as per
> +the standard semantics of the V4L2 control interface. The client may attempt
> +setting a control of its interest during encoding and if the operation fails

I'd simplify this:

The client may attempt to set a control during encoding...

> +with the -EBUSY error code, the ``CAPTURE`` queue needs to be stopped for the
> +configuration change to be allowed (following the `Drain` sequence will be
> +needed to avoid losing the already queued/encoded frames).

Rephrase:

...to be allowed. To do this follow the `Drain` sequence to avoid losing the
already queued/encoded frames.

> +
> +The timing of parameter updates is encoder-specific, as per the standard
> +semantics of the V4L2 control interface. If the client needs to apply the
> +parameters exactly at specific frame, using the Request API should be

Change this to a reference to the Request API section.

> +considered, if supported by the encoder.
> +
> +Drain
> +=====
> +
> +To ensure that all the queued ``OUTPUT`` buffers have been processed and the
> +related ``CAPTURE`` buffers output to the client, the client must follow the

output -> are output

or perhaps better (up to you): are given

> +drain sequence described below. After the drain sequence ends, the client has
> +received all encoded frames for all ``OUTPUT`` buffers queued before the
> +sequence was started.
> +
> +1. Begin the drain sequence by issuing :c:func:`VIDIOC_ENCODER_CMD`.
> +
> +   * **Required fields:**
> +
> +     ``cmd``
> +         set to ``V4L2_ENC_CMD_STOP``
> +
> +     ``flags``
> +         set to 0
> +
> +     ``pts``
> +         set to 0
> +
> +   .. warning::
> +
> +   The sentence can be only initiated if both ``OUTPUT`` and ``CAPTURE`` queues
> +   are streaming. For compatibility reasons, the call to
> +   :c:func:`VIDIOC_ENCODER_CMD` will not fail even if any of the queues is not
> +   streaming, but at the same time it will not initiate the `Drain` sequence
> +   and so the steps described below would not be applicable.
> +
> +2. Any ``OUTPUT`` buffers queued by the client before the
> +   :c:func:`VIDIOC_ENCODER_CMD` was issued will be processed and encoded as
> +   normal. The client must continue to handle both queues independently,
> +   similarly to normal encode operation. This includes,
> +
> +   * queuing and dequeuing ``CAPTURE`` buffers, until a buffer marked with the
> +     ``V4L2_BUF_FLAG_LAST`` flag is dequeued,
> +
> +     .. warning::
> +
> +        The last buffer may be empty (with :c:type:`v4l2_buffer`
> +        ``bytesused`` = 0) and in such case it must be ignored by the client,

such -> that

Check the previous patch as well if you used the phrase 'such case' and replace
it with 'that case'.

> +        as it does not contain an encoded frame.
> +
> +     .. note::
> +
> +        Any attempt to dequeue more buffers beyond the buffer marked with
> +        ``V4L2_BUF_FLAG_LAST`` will result in a -EPIPE error from
> +        :c:func:`VIDIOC_DQBUF`.
> +
> +   * dequeuing processed ``OUTPUT`` buffers, until all the buffers queued
> +     before the ``V4L2_ENC_CMD_STOP`` command are dequeued,
> +
> +   * dequeuing the ``V4L2_EVENT_EOS`` event, if the client subscribes to it.
> +
> +   .. note::
> +
> +      For backwards compatibility, the encoder will signal a ``V4L2_EVENT_EOS``
> +      event when the last the last frame has been decoded and all frames are

the last the last -> the last

> +      ready to be dequeued. It is a deprecated behavior and the client must not

is a -> is

> +      rely on it. The ``V4L2_BUF_FLAG_LAST`` buffer flag should be used
> +      instead.

Question: should new codec drivers still implement the EOS event?

> +
> +3. Once all ``OUTPUT`` buffers queued before the ``V4L2_ENC_CMD_STOP`` call and
> +   the last ``CAPTURE`` buffer are dequeued, the encoder is stopped and it will
> +   accept, but not process any newly queued ``OUTPUT`` buffers until the client
> +   issues any of the following operations:
> +
> +   * ``V4L2_ENC_CMD_START`` - the encoder will resume operation normally,

Perhaps mention that this does not reset the encoder? It's not immediately clear
when reading this.

> +
> +   * a pair of :c:func:`VIDIOC_STREAMOFF` and :c:func:`VIDIOC_STREAMON` on the
> +     ``CAPTURE`` queue - the encoder will be reset (see the `Reset` sequence)
> +     and then resume encoding,
> +
> +   * a pair of :c:func:`VIDIOC_STREAMOFF` and :c:func:`VIDIOC_STREAMON` on the
> +     ``OUTPUT`` queue - the encoder will resume operation normally, however any
> +     source frames queued to the ``OUTPUT`` queue between ``V4L2_ENC_CMD_STOP``
> +     and :c:func:`VIDIOC_STREAMOFF` will be discarded.
> +
> +.. note::
> +
> +   Once the drain sequence is initiated, the client needs to drive it to
> +   completion, as described by the steps above, unless it aborts the process by
> +   issuing :c:func:`VIDIOC_STREAMOFF` on any of the ``OUTPUT`` or ``CAPTURE``
> +   queues.  The client is not allowed to issue ``V4L2_ENC_CMD_START`` or
> +   ``V4L2_ENC_CMD_STOP`` again while the drain sequence is in progress and they
> +   will fail with -EBUSY error code if attempted.
> +
> +   Although mandatory, the availability of encoder commands may be queried
> +   using :c:func:`VIDIOC_TRY_ENCODER_CMD`.
> +
> +Reset
> +=====
> +
> +The client may want to request the encoder to reinitialize the encoding, so
> +that the following stream data becomes independent from the stream data
> +generated before. Depending on the coded format, that may imply that,

that, -> that:

> +
> +* encoded frames produced after the restart must not reference any frames
> +  produced before the stop, e.g. no long term references for H.264,
> +
> +* any headers that must be included in a standalone stream must be produced
> +  again, e.g. SPS and PPS for H.264.
> +
> +This can be achieved by performing the reset sequence.
> +
> +1. Perform the `Drain` sequence to ensure all the in-flight encoding finishes
> +   and respective buffers are dequeued.
> +
> +2. Stop streaming on the ``CAPTURE`` queue via :c:func:`VIDIOC_STREAMOFF`. This
> +   will return all currently queued ``CAPTURE`` buffers to the client, without
> +   valid frame data.
> +
> +3. Start streaming on the ``CAPTURE`` queue via :c:func:`VIDIOC_STREAMON` and
> +   continue with regular encoding sequence. The encoded frames produced into
> +   ``CAPTURE`` buffers from now on will contain a standalone stream that can be
> +   decoded without the need for frames encoded before the reset sequence,
> +   starting at the first ``OUTPUT`` buffer queued after issuing the
> +   `V4L2_ENC_CMD_STOP` of the `Drain` sequence.
> +
> +This sequence may be also used to change encoding parameters for encoders
> +without the ability to change the parameters on the fly.
> +
> +Commit points
> +=============
> +
> +Setting formats and allocating buffers triggers changes in the behavior of the
> +encoder.
> +
> +1. Setting the format on the ``CAPTURE`` queue may change the set of formats
> +   supported/advertised on the ``OUTPUT`` queue. In particular, it also means
> +   that the ``OUTPUT`` format may be reset and the client must not rely on the
> +   previously set format being preserved.
> +
> +2. Enumerating formats on the ``OUTPUT`` queue always returns only formats
> +   supported for the current ``CAPTURE`` format.
> +
> +3. Setting the format on the ``OUTPUT`` queue does not change the list of
> +   formats available on the ``CAPTURE`` queue. An attempt to set the ``OUTPUT``
> +   format that is not supported for the currently selected ``CAPTURE`` format
> +   will result in the encoder adjusting the requested ``OUTPUT`` format to a
> +   supported one.
> +
> +4. Enumerating formats on the ``CAPTURE`` queue always returns the full set of
> +   supported coded formats, irrespectively of the current ``OUTPUT`` format.
> +
> +5. While buffers are allocated on the ``CAPTURE`` queue, the client must not
> +   change the format on the queue. Drivers will return the -EBUSY error code
> +   for any such format change attempt.
> +
> +To summarize, setting formats and allocation must always start with the
> +``CAPTURE`` queue and the ``CAPTURE`` queue is the master that governs the
> +set of supported formats for the ``OUTPUT`` queue.
> diff --git a/Documentation/media/uapi/v4l/devices.rst b/Documentation/media/uapi/v4l/devices.rst
> index 12d43fe711cf..1822c66c2154 100644
> --- a/Documentation/media/uapi/v4l/devices.rst
> +++ b/Documentation/media/uapi/v4l/devices.rst
> @@ -16,6 +16,7 @@ Interfaces
>      dev-osd
>      dev-codec
>      dev-decoder
> +    dev-encoder
>      dev-effect
>      dev-raw-vbi
>      dev-sliced-vbi
> diff --git a/Documentation/media/uapi/v4l/pixfmt-v4l2.rst b/Documentation/media/uapi/v4l/pixfmt-v4l2.rst
> index ca5f2270a829..085089cd9577 100644
> --- a/Documentation/media/uapi/v4l/pixfmt-v4l2.rst
> +++ b/Documentation/media/uapi/v4l/pixfmt-v4l2.rst
> @@ -37,6 +37,11 @@ Single-planar format structure
>  	inside the stream, when fed to a stateful mem2mem decoder, the fields
>  	may be zero to rely on the decoder to detect the right values. For more
>  	details see :ref:`decoder` and format descriptions.
> +
> +	For compressed formats on the CAPTURE side of a stateful mem2mem
> +	encoder, the fields must be zero, since the coded size is expected to
> +	be calculated internally by the encoder itself, based on the OUTPUT
> +	side. For more details see :ref:`encoder` and format descriptions.

The encoder document doesn't actually mention this. I think it should, though.

I'm a bit uncertain about this: the expected resolution might impact the
sizeimage value: i.e. encoding 640x480 requires much less memory then
encoding 4k video. If this is required to be 0x0, then the driver has to
fill in a worst-case sizeimage value. It might make more sense to say that
if a non-zero resolution is given, then the driver will attempt to
calculate a sensible sizeimage value.

>      * - __u32
>        - ``pixelformat``
>        - The pixel format or type of compression, set by the application.
> diff --git a/Documentation/media/uapi/v4l/v4l2.rst b/Documentation/media/uapi/v4l/v4l2.rst
> index 65dc096199ad..2ef6693b9499 100644
> --- a/Documentation/media/uapi/v4l/v4l2.rst
> +++ b/Documentation/media/uapi/v4l/v4l2.rst
> @@ -56,6 +56,7 @@ Authors, in alphabetical order:
>  - Figa, Tomasz <tfiga@chromium.org>
>  
>    - Documented the memory-to-memory decoder interface.
> +  - Documented the memory-to-memory encoder interface.
>  
>  - H Schimek, Michael <mschimek@gmx.at>
>  
> @@ -68,6 +69,7 @@ Authors, in alphabetical order:
>  - Osciak, Pawel <posciak@chromium.org>
>  
>    - Documented the memory-to-memory decoder interface.
> +  - Documented the memory-to-memory encoder interface.
>  
>  - Osciak, Pawel <pawel@osciak.com>
>  
> diff --git a/Documentation/media/uapi/v4l/vidioc-encoder-cmd.rst b/Documentation/media/uapi/v4l/vidioc-encoder-cmd.rst
> index 5ae8c933b1b9..d571c53e761a 100644
> --- a/Documentation/media/uapi/v4l/vidioc-encoder-cmd.rst
> +++ b/Documentation/media/uapi/v4l/vidioc-encoder-cmd.rst
> @@ -50,19 +50,23 @@ currently only used by the STOP command and contains one bit: If the
>  until the end of the current *Group Of Pictures*, otherwise it will stop
>  immediately.
>  
> -A :ref:`read() <func-read>` or :ref:`VIDIOC_STREAMON <VIDIOC_STREAMON>`
> -call sends an implicit START command to the encoder if it has not been
> -started yet. After a STOP command, :ref:`read() <func-read>` calls will read
> +After a STOP command, :ref:`read() <func-read>` calls will read
>  the remaining data buffered by the driver. When the buffer is empty,
>  :ref:`read() <func-read>` will return zero and the next :ref:`read() <func-read>`
>  call will restart the encoder.
>  
> +A :ref:`read() <func-read>` or :ref:`VIDIOC_STREAMON <VIDIOC_STREAMON>`
> +call sends an implicit START command to the encoder if it has not been
> +started yet. Applies to both queues of mem2mem encoders.
> +
>  A :ref:`close() <func-close>` or :ref:`VIDIOC_STREAMOFF <VIDIOC_STREAMON>`
>  call of a streaming file descriptor sends an implicit immediate STOP to
> -the encoder, and all buffered data is discarded.
> +the encoder, and all buffered data is discarded. Applies to both queues of
> +mem2mem encoders.
>  
>  These ioctls are optional, not all drivers may support them. They were
> -introduced in Linux 2.6.21.
> +introduced in Linux 2.6.21. They are, however, mandatory for stateful mem2mem
> +encoders (as further documented in :ref:`encoder`).
>  
>  
>  .. tabularcolumns:: |p{4.4cm}|p{4.4cm}|p{8.7cm}|
> @@ -107,16 +111,20 @@ introduced in Linux 2.6.21.
>        - Stop the encoder. When the ``V4L2_ENC_CMD_STOP_AT_GOP_END`` flag
>  	is set, encoding will continue until the end of the current *Group
>  	Of Pictures*, otherwise encoding will stop immediately. When the
> -	encoder is already stopped, this command does nothing. mem2mem
> -	encoders will send a ``V4L2_EVENT_EOS`` event when the last frame
> -	has been encoded and all frames are ready to be dequeued and will
> -	set the ``V4L2_BUF_FLAG_LAST`` buffer flag on the last buffer of
> -	the capture queue to indicate there will be no new buffers
> -	produced to dequeue. This buffer may be empty, indicated by the
> -	driver setting the ``bytesused`` field to 0. Once the
> -	``V4L2_BUF_FLAG_LAST`` flag was set, the
> -	:ref:`VIDIOC_DQBUF <VIDIOC_QBUF>` ioctl will not block anymore,
> -	but return an ``EPIPE`` error code.
> +	encoder is already stopped, this command does nothing.
> +
> +	A stateful mem2mem encoder will proceed with encoding the source
> +	buffers pending before the command is issued and then stop producing
> +	new frames. It will send a ``V4L2_EVENT_EOS`` event when the last frame
> +	has been encoded and all frames are ready to be dequeued and will set
> +	the ``V4L2_BUF_FLAG_LAST`` buffer flag on the last buffer of the
> +	capture queue to indicate there will be no new buffers produced to
> +	dequeue. This buffer may be empty, indicated by the driver setting the
> +	``bytesused`` field to 0. Once the buffer with the
> +	``V4L2_BUF_FLAG_LAST`` flag set was dequeued, the :ref:`VIDIOC_DQBUF
> +	<VIDIOC_QBUF>` ioctl will not block anymore, but return an ``EPIPE``
> +	error code. No flags or other arguments are accepted in case of mem2mem
> +	encoders.  See :ref:`encoder` for more details.
>      * - ``V4L2_ENC_CMD_PAUSE``
>        - 2
>        - Pause the encoder. When the encoder has not been started yet, the
> 

Regards,

	Hans
Nicolas Dufresne Nov. 17, 2018, 4:18 a.m. UTC | #2
Le lundi 12 novembre 2018 à 14:23 +0100, Hans Verkuil a écrit :
> On 10/22/2018 04:49 PM, Tomasz Figa wrote:
> > Due to complexity of the video encoding process, the V4L2 drivers of
> > stateful encoder hardware require specific sequences of V4L2 API calls
> > to be followed. These include capability enumeration, initialization,
> > encoding, encode parameters change, drain and reset.
> > 
> > Specifics of the above have been discussed during Media Workshops at
> > LinuxCon Europe 2012 in Barcelona and then later Embedded Linux
> > Conference Europe 2014 in Düsseldorf. The de facto Codec API that
> > originated at those events was later implemented by the drivers we already
> > have merged in mainline, such as s5p-mfc or coda.
> > 
> > The only thing missing was the real specification included as a part of
> > Linux Media documentation. Fix it now and document the encoder part of
> > the Codec API.
> > 
> > Signed-off-by: Tomasz Figa <tfiga@chromium.org>
> > ---
> >  Documentation/media/uapi/v4l/dev-encoder.rst  | 579 ++++++++++++++++++
> >  Documentation/media/uapi/v4l/devices.rst      |   1 +
> >  Documentation/media/uapi/v4l/pixfmt-v4l2.rst  |   5 +
> >  Documentation/media/uapi/v4l/v4l2.rst         |   2 +
> >  .../media/uapi/v4l/vidioc-encoder-cmd.rst     |  38 +-
> >  5 files changed, 610 insertions(+), 15 deletions(-)
> >  create mode 100644 Documentation/media/uapi/v4l/dev-encoder.rst
> > 
> > diff --git a/Documentation/media/uapi/v4l/dev-encoder.rst b/Documentation/media/uapi/v4l/dev-encoder.rst
> > new file mode 100644
> > index 000000000000..41139e5e48eb
> > --- /dev/null
> > +++ b/Documentation/media/uapi/v4l/dev-encoder.rst
> > @@ -0,0 +1,579 @@
> > +.. -*- coding: utf-8; mode: rst -*-
> > +
> > +.. _encoder:
> > +
> > +*************************************************
> > +Memory-to-memory Stateful Video Encoder Interface
> > +*************************************************
> > +
> > +A stateful video encoder takes raw video frames in display order and encodes
> > +them into a bitstream. It generates complete chunks of the bitstream, including
> > +all metadata, headers, etc. The resulting bitstream does not require any
> > +further post-processing by the client.
> > +
> > +Performing software stream processing, header generation etc. in the driver
> > +in order to support this interface is strongly discouraged. In case such
> > +operations are needed, use of the Stateless Video Encoder Interface (in
> > +development) is strongly advised.
> > +
> > +Conventions and notation used in this document
> > +==============================================
> > +
> > +1. The general V4L2 API rules apply if not specified in this document
> > +   otherwise.
> > +
> > +2. The meaning of words "must", "may", "should", etc. is as per RFC
> > +   2119.
> > +
> > +3. All steps not marked "optional" are required.
> > +
> > +4. :c:func:`VIDIOC_G_EXT_CTRLS`, :c:func:`VIDIOC_S_EXT_CTRLS` may be used
> > +   interchangeably with :c:func:`VIDIOC_G_CTRL`, :c:func:`VIDIOC_S_CTRL`,
> > +   unless specified otherwise.
> > +
> > +5. Single-plane API (see spec) and applicable structures may be used
> > +   interchangeably with Multi-plane API, unless specified otherwise,
> > +   depending on encoder capabilities and following the general V4L2
> > +   guidelines.
> > +
> > +6. i = [a..b]: sequence of integers from a to b, inclusive, i.e. i =
> > +   [0..2]: i = 0, 1, 2.
> > +
> > +7. Given an ``OUTPUT`` buffer A, A' represents a buffer on the ``CAPTURE``
> > +   queue containing data (encoded frame/stream) that resulted from processing
> > +   buffer A.
> 
> The same comments as I mentioned for the previous patch apply to this section.
> 
> > +
> > +Glossary
> > +========
> > +
> > +Refer to :ref:`decoder-glossary`.
> 
> Ah, you refer to the same glossary. Then my comment about the source resolution
> terms is obviously wrong.
> 
> I wonder if it wouldn't be better to split off the sections above into a separate
> HW codec intro section where you explain the differences between stateful/stateless
> encoders and decoders, and add the conventions and glossary.
> 
> After that you have the three documents for each variant (later four when we get
> stateless encoders).
> 
> Up to you, and it can be done later in a follow-up patch.
> 
> > +
> > +State machine
> > +=============
> > +
> > +.. kernel-render:: DOT
> > +   :alt: DOT digraph of encoder state machine
> > +   :caption: Encoder state machine
> > +
> > +   digraph encoder_state_machine {
> > +       node [shape = doublecircle, label="Encoding"] Encoding;
> > +
> > +       node [shape = circle, label="Initialization"] Initialization;
> > +       node [shape = circle, label="Stopped"] Stopped;
> > +       node [shape = circle, label="Drain"] Drain;
> > +       node [shape = circle, label="Reset"] Reset;
> > +
> > +       node [shape = point]; qi
> > +       qi -> Initialization [ label = "open()" ];
> > +
> > +       Initialization -> Encoding [ label = "Both queues streaming" ];
> > +
> > +       Encoding -> Drain [ label = "V4L2_DEC_CMD_STOP" ];
> > +       Encoding -> Reset [ label = "VIDIOC_STREAMOFF(CAPTURE)" ];
> > +       Encoding -> Stopped [ label = "VIDIOC_STREAMOFF(OUTPUT)" ];
> > +       Encoding -> Encoding;
> > +
> > +       Drain -> Stopped [ label = "All CAPTURE\nbuffers dequeued\nor\nVIDIOC_STREAMOFF(CAPTURE)" ];
> > +       Drain -> Reset [ label = "VIDIOC_STREAMOFF(CAPTURE)" ];
> > +
> > +       Reset -> Encoding [ label = "VIDIOC_STREAMON(CAPTURE)" ];
> > +       Reset -> Initialization [ label = "VIDIOC_REQBUFS(OUTPUT, 0)" ];
> > +
> > +       Stopped -> Encoding [ label = "V4L2_DEC_CMD_START\nor\nVIDIOC_STREAMON(OUTPUT)" ];
> > +       Stopped -> Reset [ label = "VIDIOC_STREAMOFF(CAPTURE)" ];
> > +   }
> > +
> > +Querying capabilities
> > +=====================
> > +
> > +1. To enumerate the set of coded formats supported by the encoder, the
> > +   client may call :c:func:`VIDIOC_ENUM_FMT` on ``CAPTURE``.
> > +
> > +   * The full set of supported formats will be returned, regardless of the
> > +     format set on ``OUTPUT``.
> > +
> > +2. To enumerate the set of supported raw formats, the client may call
> > +   :c:func:`VIDIOC_ENUM_FMT` on ``OUTPUT``.
> > +
> > +   * Only the formats supported for the format currently active on ``CAPTURE``
> > +     will be returned.
> > +
> > +   * In order to enumerate raw formats supported by a given coded format,
> > +     the client must first set that coded format on ``CAPTURE`` and then
> > +     enumerate the formats on ``OUTPUT``.
> > +
> > +3. The client may use :c:func:`VIDIOC_ENUM_FRAMESIZES` to detect supported
> > +   resolutions for a given format, passing desired pixel format in
> > +   :c:type:`v4l2_frmsizeenum` ``pixel_format``.
> > +
> > +   * Values returned by :c:func:`VIDIOC_ENUM_FRAMESIZES` for a coded pixel
> > +     format will include all possible coded resolutions supported by the
> > +     encoder for given coded pixel format.
> > +
> > +   * Values returned by :c:func:`VIDIOC_ENUM_FRAMESIZES` for a raw pixel format
> > +     will include all possible frame buffer resolutions supported by the
> > +     encoder for given raw pixel format and coded format currently set on
> > +     ``CAPTURE``.
> > +
> > +4. Supported profiles and levels for given format, if applicable, may be
> 
> format -> the coded format currently set on ``CAPTURE``
> 
> > +   queried using their respective controls via :c:func:`VIDIOC_QUERYCTRL`.
> > +
> > +5. Any additional encoder capabilities may be discovered by querying
> > +   their respective controls.
> > +
> > +Initialization
> > +==============
> > +
> > +1. **Optional.** Enumerate supported formats and resolutions. See
> > +   `Querying capabilities` above.
> 
> Can be dropped IMHO.
> 
> > +
> > +2. Set a coded format on the ``CAPTURE`` queue via :c:func:`VIDIOC_S_FMT`
> > +
> > +   * **Required fields:**
> > +
> > +     ``type``
> > +         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``CAPTURE``
> > +
> > +     ``pixelformat``
> > +         the coded format to be produced
> > +
> > +     ``sizeimage``
> > +         desired size of ``CAPTURE`` buffers; the encoder may adjust it to
> > +         match hardware requirements
> > +
> > +     other fields
> > +         follow standard semantics
> > +
> > +   * **Return fields:**
> > +
> > +     ``sizeimage``
> > +         adjusted size of ``CAPTURE`` buffers
> > +
> > +   .. warning::
> > +
> > +      Changing the ``CAPTURE`` format may change the currently set ``OUTPUT``
> > +      format. The encoder will derive a new ``OUTPUT`` format from the
> > +      ``CAPTURE`` format being set, including resolution, colorimetry
> > +      parameters, etc. If the client needs a specific ``OUTPUT`` format, it
> > +      must adjust it afterwards.
> > +
> > +3. **Optional.** Enumerate supported ``OUTPUT`` formats (raw formats for
> > +   source) for the selected coded format via :c:func:`VIDIOC_ENUM_FMT`.
> 
> Does this return the same set of formats as in the 'Querying Capabilities' phase?
> 
> > +
> > +   * **Required fields:**
> > +
> > +     ``type``
> > +         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``OUTPUT``
> > +
> > +     other fields
> > +         follow standard semantics
> > +
> > +   * **Return fields:**
> > +
> > +     ``pixelformat``
> > +         raw format supported for the coded format currently selected on
> > +         the ``OUTPUT`` queue.
> 
> OUTPUT -> CAPTURE
> 
> > +
> > +     other fields
> > +         follow standard semantics
> > +
> > +4. Set the raw source format on the ``OUTPUT`` queue via
> > +   :c:func:`VIDIOC_S_FMT`.
> > +
> > +   * **Required fields:**
> > +
> > +     ``type``
> > +         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``OUTPUT``
> > +
> > +     ``pixelformat``
> > +         raw format of the source
> > +
> > +     ``width``, ``height``
> > +         source resolution
> > +
> > +     other fields
> > +         follow standard semantics
> > +
> > +   * **Return fields:**
> > +
> > +     ``width``, ``height``
> > +         may be adjusted by encoder to match alignment requirements, as
> > +         required by the currently selected formats
> > +
> > +     other fields
> > +         follow standard semantics
> > +
> > +   * Setting the source resolution will reset the selection rectangles to their
> > +     default values, based on the new resolution, as described in the step 5
> > +     below.
> > +
> > +5. **Optional.** Set the visible resolution for the stream metadata via
> > +   :c:func:`VIDIOC_S_SELECTION` on the ``OUTPUT`` queue.
> > +
> > +   * **Required fields:**
> > +
> > +     ``type``
> > +         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``OUTPUT``
> > +
> > +     ``target``
> > +         set to ``V4L2_SEL_TGT_CROP``
> > +
> > +     ``r.left``, ``r.top``, ``r.width``, ``r.height``
> > +         visible rectangle; this must fit within the `V4L2_SEL_TGT_CROP_BOUNDS`
> > +         rectangle and may be subject to adjustment to match codec and
> > +         hardware constraints
> > +
> > +   * **Return fields:**
> > +
> > +     ``r.left``, ``r.top``, ``r.width``, ``r.height``
> > +         visible rectangle adjusted by the encoder
> > +
> > +   * The following selection targets are supported on ``OUTPUT``:
> > +
> > +     ``V4L2_SEL_TGT_CROP_BOUNDS``
> > +         equal to the full source frame, matching the active ``OUTPUT``
> > +         format
> > +
> > +     ``V4L2_SEL_TGT_CROP_DEFAULT``
> > +         equal to ``V4L2_SEL_TGT_CROP_BOUNDS``
> > +
> > +     ``V4L2_SEL_TGT_CROP``
> > +         rectangle within the source buffer to be encoded into the
> > +         ``CAPTURE`` stream; defaults to ``V4L2_SEL_TGT_CROP_DEFAULT``
> 
> Since this defaults to the CROP_DEFAULT rectangle this means that if you have
> a 16x16 macroblock size and you want to encode 1080p, you will always have to
> explicitly set the CROP rectangle to 1920x1080, right? Since the default will
> be 1088 instead of 1080.
> 
> It is probably wise to explicitly mention this.
> 
> > +
> > +     ``V4L2_SEL_TGT_COMPOSE_BOUNDS``
> > +         maximum rectangle within the coded resolution, which the cropped
> > +         source frame can be output into; if the hardware does not support
> 
> output -> composed
> 
> > +         composition or scaling, then this is always equal to the rectangle of
> > +         width and height matching ``V4L2_SEL_TGT_CROP`` and located at (0, 0)
> > +
> > +     ``V4L2_SEL_TGT_COMPOSE_DEFAULT``
> > +         equal to a rectangle of width and height matching
> > +         ``V4L2_SEL_TGT_CROP`` and located at (0, 0)
> > +
> > +     ``V4L2_SEL_TGT_COMPOSE``
> > +         rectangle within the coded frame, which the cropped source frame
> > +         is to be output into; defaults to
> 
> output -> composed
> 
> > +         ``V4L2_SEL_TGT_COMPOSE_DEFAULT``; read-only on hardware without
> > +         additional compose/scaling capabilities; resulting stream will
> > +         have this rectangle encoded as the visible rectangle in its
> > +         metadata
> > +
> > +   .. warning::
> > +
> > +      The encoder may adjust the crop/compose rectangles to the nearest
> > +      supported ones to meet codec and hardware requirements. The client needs
> > +      to check the adjusted rectangle returned by :c:func:`VIDIOC_S_SELECTION`.
> > +
> > +6. Allocate buffers for both ``OUTPUT`` and ``CAPTURE`` via
> > +   :c:func:`VIDIOC_REQBUFS`. This may be performed in any order.
> > +
> > +   * **Required fields:**
> > +
> > +     ``count``
> > +         requested number of buffers to allocate; greater than zero
> > +
> > +     ``type``
> > +         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``OUTPUT`` or
> > +         ``CAPTURE``
> > +
> > +     other fields
> > +         follow standard semantics
> > +
> > +   * **Return fields:**
> > +
> > +     ``count``
> > +          actual number of buffers allocated
> > +
> > +   .. warning::
> > +
> > +      The actual number of allocated buffers may differ from the ``count``
> > +      given. The client must check the updated value of ``count`` after the
> > +      call returns.
> > +
> > +   .. note::
> > +
> > +      To allocate more than the minimum number of buffers (for pipeline depth),
> > +      the client may query the ``V4L2_CID_MIN_BUFFERS_FOR_OUTPUT`` or
> > +      ``V4L2_CID_MIN_BUFFERS_FOR_CAPTURE`` control respectively, to get the
> > +      minimum number of buffers required by the encoder/format, and pass the
> > +      obtained value plus the number of additional buffers needed in the
> > +      ``count`` field to :c:func:`VIDIOC_REQBUFS`.
> 
> Does V4L2_CID_MIN_BUFFERS_FOR_CAPTURE make any sense for encoders?

We do account for it in GStreamer (the capture/output handling is
generic), but I don't know if it's being used anywhere. 

> 
> V4L2_CID_MIN_BUFFERS_FOR_OUTPUT can make sense depending on GOP size etc.

Not really the GOP size. In video conference we often run the encoder
with an open GOP and do key frame request on demand. Mostly the DPB
size. DPB is specific term use in certain CODEC, but they nearly all
keep some backlog, except for key frame only codecs (jpeg, png, etc.).

> 
> > +
> > +   Alternatively, :c:func:`VIDIOC_CREATE_BUFS` can be used to have more
> > +   control over buffer allocation.
> > +
> > +   * **Required fields:**
> > +
> > +     ``count``
> > +         requested number of buffers to allocate; greater than zero
> > +
> > +     ``type``
> > +         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``OUTPUT``
> > +
> > +     other fields
> > +         follow standard semantics
> > +
> > +   * **Return fields:**
> > +
> > +     ``count``
> > +         adjusted to the number of allocated buffers
> > +
> > +7. Begin streaming on both ``OUTPUT`` and ``CAPTURE`` queues via
> > +   :c:func:`VIDIOC_STREAMON`. This may be performed in any order. The actual
> > +   encoding process starts when both queues start streaming.
> > +
> > +.. note::
> > +
> > +   If the client stops the ``CAPTURE`` queue during the encode process and then
> > +   restarts it again, the encoder will begin generating a stream independent
> > +   from the stream generated before the stop. The exact constraints depend
> > +   on the coded format, but may include the following implications:
> > +
> > +   * encoded frames produced after the restart must not reference any
> > +     frames produced before the stop, e.g. no long term references for
> > +     H.264,
> > +
> > +   * any headers that must be included in a standalone stream must be
> > +     produced again, e.g. SPS and PPS for H.264.
> > +
> > +Encoding
> > +========
> > +
> > +This state is reached after the `Initialization` sequence finishes succesfully.
> 
> successfully
> 
> > +In this state, client queues and dequeues buffers to both queues via
> 
> client -> the client
> 
> > +:c:func:`VIDIOC_QBUF` and :c:func:`VIDIOC_DQBUF`, following the standard
> > +semantics.
> > +
> > +The contents of encoded ``CAPTURE`` buffers depend on the active coded pixel
> 
> contents ... depend -> content ... depends
> 
> > +format and may be affected by codec-specific extended controls, as stated
> > +in the documentation of each format.
> > +
> > +Both queues operate independently, following standard behavior of V4L2 buffer
> > +queues and memory-to-memory devices. In addition, the order of encoded frames
> > +dequeued from the ``CAPTURE`` queue may differ from the order of queuing raw
> > +frames to the ``OUTPUT`` queue, due to properties of the selected coded format,
> > +e.g. frame reordering.
> > +
> > +The client must not assume any direct relationship between ``CAPTURE`` and
> > +``OUTPUT`` buffers and any specific timing of buffers becoming
> > +available to dequeue. Specifically,
> > +
> > +* a buffer queued to ``OUTPUT`` may result in more than 1 buffer produced on
> > +  ``CAPTURE`` (if returning an encoded frame allowed the encoder to return a
> > +  frame that preceded it in display, but succeeded it in the decode order),
> > +
> > +* a buffer queued to ``OUTPUT`` may result in a buffer being produced on
> > +  ``CAPTURE`` later into encode process, and/or after processing further
> > +  ``OUTPUT`` buffers, or be returned out of order, e.g. if display
> > +  reordering is used,
> > +
> > +* buffers may become available on the ``CAPTURE`` queue without additional
> > +  buffers queued to ``OUTPUT`` (e.g. during drain or ``EOS``), because of the
> > +  ``OUTPUT`` buffers queued in the past whose decoding results are only
> > +  available at later time, due to specifics of the decoding process,
> > +
> > +* buffers queued to ``OUTPUT`` may not become available to dequeue instantly
> > +  after being encoded into a corresponding ``CATPURE`` buffer, e.g. if the
> > +  encoder needs to use the frame as a reference for encoding further frames.
> > +
> > +.. note::
> > +
> > +   To allow matching encoded ``CAPTURE`` buffers with ``OUTPUT`` buffers they
> > +   originated from, the client can set the ``timestamp`` field of the
> > +   :c:type:`v4l2_buffer` struct when queuing an ``OUTPUT`` buffer. The
> > +   ``CAPTURE`` buffer(s), which resulted from encoding that ``OUTPUT`` buffer
> > +   will have their ``timestamp`` field set to the same value when dequeued.
> > +
> > +   In addition to the straighforward case of one ``OUTPUT`` buffer producing
> > +   one ``CAPTURE`` buffer, the following cases are defined:
> > +
> > +   * one ``OUTPUT`` buffer generates multiple ``CAPTURE`` buffers: the same
> > +     ``OUTPUT`` timestamp will be copied to multiple ``CAPTURE`` buffers,
> > +
> > +   * the encoding order differs from the presentation order (i.e. the
> > +     ``CAPTURE`` buffers are out-of-order compared to the ``OUTPUT`` buffers):
> > +     ``CAPTURE`` timestamps will not retain the order of ``OUTPUT`` timestamps
> > +     and thus monotonicity of the timestamps cannot be guaranteed.
> > +
> > +.. note::
> > +
> > +   To let the client distinguish between frame types (keyframes, intermediate
> > +   frames; the exact list of types depends on the coded format), the
> > +   ``CAPTURE`` buffers will have corresponding flag bits set in their
> > +   :c:type:`v4l2_buffer` struct when dequeued. See the documentation of
> > +   :c:type:`v4l2_buffer` and each coded pixel format for exact list of flags
> > +   and their meanings.
> 
> Is this required? (I think it should, but it isn't the case today).

Most CODEC do provide this metadata, if it's absent, placing the stream
into a container may require more parsing.

> 
> Is the current set of buffer flags (Key/B/P frame) sufficient for the current
> set of codecs?

Doe it matter ? It can be extended when new codec get added. There is a
lot things in AV1 as an example, but we can add these when HW and
drivers starts shipping.

> 
> > +
> > +Encoding parameter changes
> > +==========================
> > +
> > +The client is allowed to use :c:func:`VIDIOC_S_CTRL` to change encoder
> > +parameters at any time. The availability of parameters is encoder-specific
> > +and the client must query the encoder to find the set of available controls.
> > +
> > +The ability to change each parameter during encoding is encoder-specific, as per
> > +the standard semantics of the V4L2 control interface. The client may attempt
> > +setting a control of its interest during encoding and if the operation fails
> 
> I'd simplify this:
> 
> The client may attempt to set a control during encoding...
> 
> > +with the -EBUSY error code, the ``CAPTURE`` queue needs to be stopped for the
> > +configuration change to be allowed (following the `Drain` sequence will be
> > +needed to avoid losing the already queued/encoded frames).
> 
> Rephrase:
> 
> ...to be allowed. To do this follow the `Drain` sequence to avoid losing the
> already queued/encoded frames.
> 
> > +
> > +The timing of parameter updates is encoder-specific, as per the standard
> > +semantics of the V4L2 control interface. If the client needs to apply the
> > +parameters exactly at specific frame, using the Request API should be
> 
> Change this to a reference to the Request API section.
> 
> > +considered, if supported by the encoder.
> > +
> > +Drain
> > +=====
> > +
> > +To ensure that all the queued ``OUTPUT`` buffers have been processed and the
> > +related ``CAPTURE`` buffers output to the client, the client must follow the
> 
> output -> are output
> 
> or perhaps better (up to you): are given
> 
> > +drain sequence described below. After the drain sequence ends, the client has
> > +received all encoded frames for all ``OUTPUT`` buffers queued before the
> > +sequence was started.
> > +
> > +1. Begin the drain sequence by issuing :c:func:`VIDIOC_ENCODER_CMD`.
> > +
> > +   * **Required fields:**
> > +
> > +     ``cmd``
> > +         set to ``V4L2_ENC_CMD_STOP``
> > +
> > +     ``flags``
> > +         set to 0
> > +
> > +     ``pts``
> > +         set to 0
> > +
> > +   .. warning::
> > +
> > +   The sentence can be only initiated if both ``OUTPUT`` and ``CAPTURE`` queues
> > +   are streaming. For compatibility reasons, the call to
> > +   :c:func:`VIDIOC_ENCODER_CMD` will not fail even if any of the queues is not
> > +   streaming, but at the same time it will not initiate the `Drain` sequence
> > +   and so the steps described below would not be applicable.
> > +
> > +2. Any ``OUTPUT`` buffers queued by the client before the
> > +   :c:func:`VIDIOC_ENCODER_CMD` was issued will be processed and encoded as
> > +   normal. The client must continue to handle both queues independently,
> > +   similarly to normal encode operation. This includes,
> > +
> > +   * queuing and dequeuing ``CAPTURE`` buffers, until a buffer marked with the
> > +     ``V4L2_BUF_FLAG_LAST`` flag is dequeued,
> > +
> > +     .. warning::
> > +
> > +        The last buffer may be empty (with :c:type:`v4l2_buffer`
> > +        ``bytesused`` = 0) and in such case it must be ignored by the client,
> 
> such -> that
> 
> Check the previous patch as well if you used the phrase 'such case' and replace
> it with 'that case'.
> 
> > +        as it does not contain an encoded frame.
> > +
> > +     .. note::
> > +
> > +        Any attempt to dequeue more buffers beyond the buffer marked with
> > +        ``V4L2_BUF_FLAG_LAST`` will result in a -EPIPE error from
> > +        :c:func:`VIDIOC_DQBUF`.
> > +
> > +   * dequeuing processed ``OUTPUT`` buffers, until all the buffers queued
> > +     before the ``V4L2_ENC_CMD_STOP`` command are dequeued,
> > +
> > +   * dequeuing the ``V4L2_EVENT_EOS`` event, if the client subscribes to it.
> > +
> > +   .. note::
> > +
> > +      For backwards compatibility, the encoder will signal a ``V4L2_EVENT_EOS``
> > +      event when the last the last frame has been decoded and all frames are
> 
> the last the last -> the last
> 
> > +      ready to be dequeued. It is a deprecated behavior and the client must not
> 
> is a -> is
> 
> > +      rely on it. The ``V4L2_BUF_FLAG_LAST`` buffer flag should be used
> > +      instead.
> 
> Question: should new codec drivers still implement the EOS event?

I'm been asking around, but I think here is a good place. Do we really
need the FLAG_LAST in userspace ? Userspace can also wait for the first
EPIPE return from DQBUF.

> 
> > +
> > +3. Once all ``OUTPUT`` buffers queued before the ``V4L2_ENC_CMD_STOP`` call and
> > +   the last ``CAPTURE`` buffer are dequeued, the encoder is stopped and it will
> > +   accept, but not process any newly queued ``OUTPUT`` buffers until the client
> > +   issues any of the following operations:
> > +
> > +   * ``V4L2_ENC_CMD_START`` - the encoder will resume operation normally,
> 
> Perhaps mention that this does not reset the encoder? It's not immediately clear
> when reading this.

Which drivers supports this ? I believe I tried with Exynos in the
past, and that didn't work. How do we know if a driver supports this or
not. Do we make it mandatory ? When it's not supported, it basically
mean userspace need to cache and resend the header in userspace, and
also need to skip to some sync point.

> 
> > +
> > +   * a pair of :c:func:`VIDIOC_STREAMOFF` and :c:func:`VIDIOC_STREAMON` on the
> > +     ``CAPTURE`` queue - the encoder will be reset (see the `Reset` sequence)
> > +     and then resume encoding,
> > +
> > +   * a pair of :c:func:`VIDIOC_STREAMOFF` and :c:func:`VIDIOC_STREAMON` on the
> > +     ``OUTPUT`` queue - the encoder will resume operation normally, however any
> > +     source frames queued to the ``OUTPUT`` queue between ``V4L2_ENC_CMD_STOP``
> > +     and :c:func:`VIDIOC_STREAMOFF` will be discarded.
> > +
> > +.. note::
> > +
> > +   Once the drain sequence is initiated, the client needs to drive it to
> > +   completion, as described by the steps above, unless it aborts the process by
> > +   issuing :c:func:`VIDIOC_STREAMOFF` on any of the ``OUTPUT`` or ``CAPTURE``
> > +   queues.  The client is not allowed to issue ``V4L2_ENC_CMD_START`` or
> > +   ``V4L2_ENC_CMD_STOP`` again while the drain sequence is in progress and they
> > +   will fail with -EBUSY error code if attempted.
> > +
> > +   Although mandatory, the availability of encoder commands may be queried
> > +   using :c:func:`VIDIOC_TRY_ENCODER_CMD`.
> > +
> > +Reset
> > +=====
> > +
> > +The client may want to request the encoder to reinitialize the encoding, so
> > +that the following stream data becomes independent from the stream data
> > +generated before. Depending on the coded format, that may imply that,
> 
> that, -> that:
> 
> > +
> > +* encoded frames produced after the restart must not reference any frames
> > +  produced before the stop, e.g. no long term references for H.264,
> > +
> > +* any headers that must be included in a standalone stream must be produced
> > +  again, e.g. SPS and PPS for H.264.
> > +
> > +This can be achieved by performing the reset sequence.
> > +
> > +1. Perform the `Drain` sequence to ensure all the in-flight encoding finishes
> > +   and respective buffers are dequeued.
> > +
> > +2. Stop streaming on the ``CAPTURE`` queue via :c:func:`VIDIOC_STREAMOFF`. This
> > +   will return all currently queued ``CAPTURE`` buffers to the client, without
> > +   valid frame data.
> > +
> > +3. Start streaming on the ``CAPTURE`` queue via :c:func:`VIDIOC_STREAMON` and
> > +   continue with regular encoding sequence. The encoded frames produced into
> > +   ``CAPTURE`` buffers from now on will contain a standalone stream that can be
> > +   decoded without the need for frames encoded before the reset sequence,
> > +   starting at the first ``OUTPUT`` buffer queued after issuing the
> > +   `V4L2_ENC_CMD_STOP` of the `Drain` sequence.
> > +
> > +This sequence may be also used to change encoding parameters for encoders
> > +without the ability to change the parameters on the fly.
> > +
> > +Commit points
> > +=============
> > +
> > +Setting formats and allocating buffers triggers changes in the behavior of the
> > +encoder.
> > +
> > +1. Setting the format on the ``CAPTURE`` queue may change the set of formats
> > +   supported/advertised on the ``OUTPUT`` queue. In particular, it also means
> > +   that the ``OUTPUT`` format may be reset and the client must not rely on the
> > +   previously set format being preserved.
> > +
> > +2. Enumerating formats on the ``OUTPUT`` queue always returns only formats
> > +   supported for the current ``CAPTURE`` format.
> > +
> > +3. Setting the format on the ``OUTPUT`` queue does not change the list of
> > +   formats available on the ``CAPTURE`` queue. An attempt to set the ``OUTPUT``
> > +   format that is not supported for the currently selected ``CAPTURE`` format
> > +   will result in the encoder adjusting the requested ``OUTPUT`` format to a
> > +   supported one.
> > +
> > +4. Enumerating formats on the ``CAPTURE`` queue always returns the full set of
> > +   supported coded formats, irrespectively of the current ``OUTPUT`` format.
> > +
> > +5. While buffers are allocated on the ``CAPTURE`` queue, the client must not
> > +   change the format on the queue. Drivers will return the -EBUSY error code
> > +   for any such format change attempt.
> > +
> > +To summarize, setting formats and allocation must always start with the
> > +``CAPTURE`` queue and the ``CAPTURE`` queue is the master that governs the
> > +set of supported formats for the ``OUTPUT`` queue.
> > diff --git a/Documentation/media/uapi/v4l/devices.rst b/Documentation/media/uapi/v4l/devices.rst
> > index 12d43fe711cf..1822c66c2154 100644
> > --- a/Documentation/media/uapi/v4l/devices.rst
> > +++ b/Documentation/media/uapi/v4l/devices.rst
> > @@ -16,6 +16,7 @@ Interfaces
> >      dev-osd
> >      dev-codec
> >      dev-decoder
> > +    dev-encoder
> >      dev-effect
> >      dev-raw-vbi
> >      dev-sliced-vbi
> > diff --git a/Documentation/media/uapi/v4l/pixfmt-v4l2.rst b/Documentation/media/uapi/v4l/pixfmt-v4l2.rst
> > index ca5f2270a829..085089cd9577 100644
> > --- a/Documentation/media/uapi/v4l/pixfmt-v4l2.rst
> > +++ b/Documentation/media/uapi/v4l/pixfmt-v4l2.rst
> > @@ -37,6 +37,11 @@ Single-planar format structure
> >  	inside the stream, when fed to a stateful mem2mem decoder, the fields
> >  	may be zero to rely on the decoder to detect the right values. For more
> >  	details see :ref:`decoder` and format descriptions.
> > +
> > +	For compressed formats on the CAPTURE side of a stateful mem2mem
> > +	encoder, the fields must be zero, since the coded size is expected to
> > +	be calculated internally by the encoder itself, based on the OUTPUT
> > +	side. For more details see :ref:`encoder` and format descriptions.
> 
> The encoder document doesn't actually mention this. I think it should, though.
> 
> I'm a bit uncertain about this: the expected resolution might impact the
> sizeimage value: i.e. encoding 640x480 requires much less memory then
> encoding 4k video. If this is required to be 0x0, then the driver has to
> fill in a worst-case sizeimage value. It might make more sense to say that
> if a non-zero resolution is given, then the driver will attempt to
> calculate a sensible sizeimage value.
> 
> >      * - __u32
> >        - ``pixelformat``
> >        - The pixel format or type of compression, set by the application.
> > diff --git a/Documentation/media/uapi/v4l/v4l2.rst b/Documentation/media/uapi/v4l/v4l2.rst
> > index 65dc096199ad..2ef6693b9499 100644
> > --- a/Documentation/media/uapi/v4l/v4l2.rst
> > +++ b/Documentation/media/uapi/v4l/v4l2.rst
> > @@ -56,6 +56,7 @@ Authors, in alphabetical order:
> >  - Figa, Tomasz <tfiga@chromium.org>
> >  
> >    - Documented the memory-to-memory decoder interface.
> > +  - Documented the memory-to-memory encoder interface.
> >  
> >  - H Schimek, Michael <mschimek@gmx.at>
> >  
> > @@ -68,6 +69,7 @@ Authors, in alphabetical order:
> >  - Osciak, Pawel <posciak@chromium.org>
> >  
> >    - Documented the memory-to-memory decoder interface.
> > +  - Documented the memory-to-memory encoder interface.
> >  
> >  - Osciak, Pawel <pawel@osciak.com>
> >  
> > diff --git a/Documentation/media/uapi/v4l/vidioc-encoder-cmd.rst b/Documentation/media/uapi/v4l/vidioc-encoder-cmd.rst
> > index 5ae8c933b1b9..d571c53e761a 100644
> > --- a/Documentation/media/uapi/v4l/vidioc-encoder-cmd.rst
> > +++ b/Documentation/media/uapi/v4l/vidioc-encoder-cmd.rst
> > @@ -50,19 +50,23 @@ currently only used by the STOP command and contains one bit: If the
> >  until the end of the current *Group Of Pictures*, otherwise it will stop
> >  immediately.
> >  
> > -A :ref:`read() <func-read>` or :ref:`VIDIOC_STREAMON <VIDIOC_STREAMON>`
> > -call sends an implicit START command to the encoder if it has not been
> > -started yet. After a STOP command, :ref:`read() <func-read>` calls will read
> > +After a STOP command, :ref:`read() <func-read>` calls will read
> >  the remaining data buffered by the driver. When the buffer is empty,
> >  :ref:`read() <func-read>` will return zero and the next :ref:`read() <func-read>`
> >  call will restart the encoder.
> >  
> > +A :ref:`read() <func-read>` or :ref:`VIDIOC_STREAMON <VIDIOC_STREAMON>`
> > +call sends an implicit START command to the encoder if it has not been
> > +started yet. Applies to both queues of mem2mem encoders.
> > +
> >  A :ref:`close() <func-close>` or :ref:`VIDIOC_STREAMOFF <VIDIOC_STREAMON>`
> >  call of a streaming file descriptor sends an implicit immediate STOP to
> > -the encoder, and all buffered data is discarded.
> > +the encoder, and all buffered data is discarded. Applies to both queues of
> > +mem2mem encoders.
> >  
> >  These ioctls are optional, not all drivers may support them. They were
> > -introduced in Linux 2.6.21.
> > +introduced in Linux 2.6.21. They are, however, mandatory for stateful mem2mem
> > +encoders (as further documented in :ref:`encoder`).
> >  
> >  
> >  .. tabularcolumns:: |p{4.4cm}|p{4.4cm}|p{8.7cm}|
> > @@ -107,16 +111,20 @@ introduced in Linux 2.6.21.
> >        - Stop the encoder. When the ``V4L2_ENC_CMD_STOP_AT_GOP_END`` flag
> >  	is set, encoding will continue until the end of the current *Group
> >  	Of Pictures*, otherwise encoding will stop immediately. When the
> > -	encoder is already stopped, this command does nothing. mem2mem
> > -	encoders will send a ``V4L2_EVENT_EOS`` event when the last frame
> > -	has been encoded and all frames are ready to be dequeued and will
> > -	set the ``V4L2_BUF_FLAG_LAST`` buffer flag on the last buffer of
> > -	the capture queue to indicate there will be no new buffers
> > -	produced to dequeue. This buffer may be empty, indicated by the
> > -	driver setting the ``bytesused`` field to 0. Once the
> > -	``V4L2_BUF_FLAG_LAST`` flag was set, the
> > -	:ref:`VIDIOC_DQBUF <VIDIOC_QBUF>` ioctl will not block anymore,
> > -	but return an ``EPIPE`` error code.
> > +	encoder is already stopped, this command does nothing.
> > +
> > +	A stateful mem2mem encoder will proceed with encoding the source
> > +	buffers pending before the command is issued and then stop producing
> > +	new frames. It will send a ``V4L2_EVENT_EOS`` event when the last frame
> > +	has been encoded and all frames are ready to be dequeued and will set
> > +	the ``V4L2_BUF_FLAG_LAST`` buffer flag on the last buffer of the
> > +	capture queue to indicate there will be no new buffers produced to
> > +	dequeue. This buffer may be empty, indicated by the driver setting the
> > +	``bytesused`` field to 0. Once the buffer with the
> > +	``V4L2_BUF_FLAG_LAST`` flag set was dequeued, the :ref:`VIDIOC_DQBUF
> > +	<VIDIOC_QBUF>` ioctl will not block anymore, but return an ``EPIPE``
> > +	error code. No flags or other arguments are accepted in case of mem2mem
> > +	encoders.  See :ref:`encoder` for more details.
> >      * - ``V4L2_ENC_CMD_PAUSE``
> >        - 2
> >        - Pause the encoder. When the encoder has not been started yet, the
> > 
> 
> Regards,
> 
> 	Hans
Hans Verkuil Nov. 17, 2018, 11:37 a.m. UTC | #3
On 11/17/2018 05:18 AM, Nicolas Dufresne wrote:
> Le lundi 12 novembre 2018 à 14:23 +0100, Hans Verkuil a écrit :
>> On 10/22/2018 04:49 PM, Tomasz Figa wrote:
>>> Due to complexity of the video encoding process, the V4L2 drivers of
>>> stateful encoder hardware require specific sequences of V4L2 API calls
>>> to be followed. These include capability enumeration, initialization,
>>> encoding, encode parameters change, drain and reset.
>>>
>>> Specifics of the above have been discussed during Media Workshops at
>>> LinuxCon Europe 2012 in Barcelona and then later Embedded Linux
>>> Conference Europe 2014 in Düsseldorf. The de facto Codec API that
>>> originated at those events was later implemented by the drivers we already
>>> have merged in mainline, such as s5p-mfc or coda.
>>>
>>> The only thing missing was the real specification included as a part of
>>> Linux Media documentation. Fix it now and document the encoder part of
>>> the Codec API.
>>>
>>> Signed-off-by: Tomasz Figa <tfiga@chromium.org>
>>> ---
>>>  Documentation/media/uapi/v4l/dev-encoder.rst  | 579 ++++++++++++++++++
>>>  Documentation/media/uapi/v4l/devices.rst      |   1 +
>>>  Documentation/media/uapi/v4l/pixfmt-v4l2.rst  |   5 +
>>>  Documentation/media/uapi/v4l/v4l2.rst         |   2 +
>>>  .../media/uapi/v4l/vidioc-encoder-cmd.rst     |  38 +-
>>>  5 files changed, 610 insertions(+), 15 deletions(-)
>>>  create mode 100644 Documentation/media/uapi/v4l/dev-encoder.rst
>>>
>>> diff --git a/Documentation/media/uapi/v4l/dev-encoder.rst b/Documentation/media/uapi/v4l/dev-encoder.rst
>>> new file mode 100644
>>> index 000000000000..41139e5e48eb
>>> --- /dev/null
>>> +++ b/Documentation/media/uapi/v4l/dev-encoder.rst

<snip>

>>> +6. Allocate buffers for both ``OUTPUT`` and ``CAPTURE`` via
>>> +   :c:func:`VIDIOC_REQBUFS`. This may be performed in any order.
>>> +
>>> +   * **Required fields:**
>>> +
>>> +     ``count``
>>> +         requested number of buffers to allocate; greater than zero
>>> +
>>> +     ``type``
>>> +         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``OUTPUT`` or
>>> +         ``CAPTURE``
>>> +
>>> +     other fields
>>> +         follow standard semantics
>>> +
>>> +   * **Return fields:**
>>> +
>>> +     ``count``
>>> +          actual number of buffers allocated
>>> +
>>> +   .. warning::
>>> +
>>> +      The actual number of allocated buffers may differ from the ``count``
>>> +      given. The client must check the updated value of ``count`` after the
>>> +      call returns.
>>> +
>>> +   .. note::
>>> +
>>> +      To allocate more than the minimum number of buffers (for pipeline depth),
>>> +      the client may query the ``V4L2_CID_MIN_BUFFERS_FOR_OUTPUT`` or
>>> +      ``V4L2_CID_MIN_BUFFERS_FOR_CAPTURE`` control respectively, to get the
>>> +      minimum number of buffers required by the encoder/format, and pass the
>>> +      obtained value plus the number of additional buffers needed in the
>>> +      ``count`` field to :c:func:`VIDIOC_REQBUFS`.
>>
>> Does V4L2_CID_MIN_BUFFERS_FOR_CAPTURE make any sense for encoders?
> 
> We do account for it in GStreamer (the capture/output handling is
> generic), but I don't know if it's being used anywhere. 

Do you use this value directly for REQBUFS, or do you use it as the minimum
value but in practice use more buffers?

> 
>>
>> V4L2_CID_MIN_BUFFERS_FOR_OUTPUT can make sense depending on GOP size etc.
> 
> Not really the GOP size. In video conference we often run the encoder
> with an open GOP and do key frame request on demand. Mostly the DPB
> size. DPB is specific term use in certain CODEC, but they nearly all
> keep some backlog, except for key frame only codecs (jpeg, png, etc.).
> 
>>

<snip>

>>> +.. note::
>>> +
>>> +   To let the client distinguish between frame types (keyframes, intermediate
>>> +   frames; the exact list of types depends on the coded format), the
>>> +   ``CAPTURE`` buffers will have corresponding flag bits set in their
>>> +   :c:type:`v4l2_buffer` struct when dequeued. See the documentation of
>>> +   :c:type:`v4l2_buffer` and each coded pixel format for exact list of flags
>>> +   and their meanings.
>>
>> Is this required? (I think it should, but it isn't the case today).
> 
> Most CODEC do provide this metadata, if it's absent, placing the stream
> into a container may require more parsing.
> 
>>
>> Is the current set of buffer flags (Key/B/P frame) sufficient for the current
>> set of codecs?
> 
> Doe it matter ? It can be extended when new codec get added. There is a
> lot things in AV1 as an example, but we can add these when HW and
> drivers starts shipping.

I was just curious :-) It doesn't really matter, I agree.

<snip>

>>> +      rely on it. The ``V4L2_BUF_FLAG_LAST`` buffer flag should be used
>>> +      instead.
>>
>> Question: should new codec drivers still implement the EOS event?
> 
> I'm been asking around, but I think here is a good place. Do we really
> need the FLAG_LAST in userspace ? Userspace can also wait for the first
> EPIPE return from DQBUF.

I'm interested in hearing Tomasz' opinion. This flag is used already, so there
definitely is a backwards compatibility issue here.

> 
>>
>>> +
>>> +3. Once all ``OUTPUT`` buffers queued before the ``V4L2_ENC_CMD_STOP`` call and
>>> +   the last ``CAPTURE`` buffer are dequeued, the encoder is stopped and it will
>>> +   accept, but not process any newly queued ``OUTPUT`` buffers until the client
>>> +   issues any of the following operations:
>>> +
>>> +   * ``V4L2_ENC_CMD_START`` - the encoder will resume operation normally,
>>
>> Perhaps mention that this does not reset the encoder? It's not immediately clear
>> when reading this.
> 
> Which drivers supports this ? I believe I tried with Exynos in the
> past, and that didn't work. How do we know if a driver supports this or
> not. Do we make it mandatory ? When it's not supported, it basically
> mean userspace need to cache and resend the header in userspace, and
> also need to skip to some sync point.

Once we agree on the spec, then the next step will be to add good compliance
checks and update drivers that fail the tests.

To check if the driver support this ioctl you can call VIDIOC_TRY_ENCODER_CMD
to see if the functionality is supported.

Regards,

	Hans
Nicolas Dufresne Nov. 18, 2018, 1:34 a.m. UTC | #4
Le samedi 17 novembre 2018 à 12:37 +0100, Hans Verkuil a écrit :
> > > Does V4L2_CID_MIN_BUFFERS_FOR_CAPTURE make any sense for encoders?
> > 
> > We do account for it in GStreamer (the capture/output handling is
> > generic), but I don't know if it's being used anywhere. 
> 
> Do you use this value directly for REQBUFS, or do you use it as the minimum
> value but in practice use more buffers?

We add more buffers to that value. We assume this value is what will be
held by the driver, hence without adding some buffers, the driver would
go idle as soon as one is dequeued. We also need to allocate for the
importing driver.

In general, if we have a pipeline with Driver A sending to Driver B,
both driver will require a certain amount of buffers to operate. E.g.
with DRM display, the driver will hold on 1 buffer (the scannout
buffer).

In GStreamer, it's implemented generically, so we do:

  MIN_BUFFERS_FOR + remote_min + 1

If only MIN_BUFFERS_FOR was allocated, ignoring remote driver
requirement, the streaming will likely get stuck.

regards,
Nicolas
Tomasz Figa Jan. 23, 2019, 9:52 a.m. UTC | #5
On Mon, Nov 12, 2018 at 10:23 PM Hans Verkuil <hverkuil@xs4all.nl> wrote:
>
> On 10/22/2018 04:49 PM, Tomasz Figa wrote:
> > Due to complexity of the video encoding process, the V4L2 drivers of
> > stateful encoder hardware require specific sequences of V4L2 API calls
> > to be followed. These include capability enumeration, initialization,
> > encoding, encode parameters change, drain and reset.
> >
> > Specifics of the above have been discussed during Media Workshops at
> > LinuxCon Europe 2012 in Barcelona and then later Embedded Linux
> > Conference Europe 2014 in Düsseldorf. The de facto Codec API that
> > originated at those events was later implemented by the drivers we already
> > have merged in mainline, such as s5p-mfc or coda.
> >
> > The only thing missing was the real specification included as a part of
> > Linux Media documentation. Fix it now and document the encoder part of
> > the Codec API.
> >
> > Signed-off-by: Tomasz Figa <tfiga@chromium.org>
> > ---
> >  Documentation/media/uapi/v4l/dev-encoder.rst  | 579 ++++++++++++++++++
> >  Documentation/media/uapi/v4l/devices.rst      |   1 +
> >  Documentation/media/uapi/v4l/pixfmt-v4l2.rst  |   5 +
> >  Documentation/media/uapi/v4l/v4l2.rst         |   2 +
> >  .../media/uapi/v4l/vidioc-encoder-cmd.rst     |  38 +-
> >  5 files changed, 610 insertions(+), 15 deletions(-)
> >  create mode 100644 Documentation/media/uapi/v4l/dev-encoder.rst
> >
> > diff --git a/Documentation/media/uapi/v4l/dev-encoder.rst b/Documentation/media/uapi/v4l/dev-encoder.rst
> > new file mode 100644
> > index 000000000000..41139e5e48eb
> > --- /dev/null
> > +++ b/Documentation/media/uapi/v4l/dev-encoder.rst
> > @@ -0,0 +1,579 @@
> > +.. -*- coding: utf-8; mode: rst -*-
> > +
> > +.. _encoder:
> > +
> > +*************************************************
> > +Memory-to-memory Stateful Video Encoder Interface
> > +*************************************************
> > +
> > +A stateful video encoder takes raw video frames in display order and encodes
> > +them into a bitstream. It generates complete chunks of the bitstream, including
> > +all metadata, headers, etc. The resulting bitstream does not require any
> > +further post-processing by the client.
> > +
> > +Performing software stream processing, header generation etc. in the driver
> > +in order to support this interface is strongly discouraged. In case such
> > +operations are needed, use of the Stateless Video Encoder Interface (in
> > +development) is strongly advised.
> > +
> > +Conventions and notation used in this document
> > +==============================================
> > +
> > +1. The general V4L2 API rules apply if not specified in this document
> > +   otherwise.
> > +
> > +2. The meaning of words "must", "may", "should", etc. is as per RFC
> > +   2119.
> > +
> > +3. All steps not marked "optional" are required.
> > +
> > +4. :c:func:`VIDIOC_G_EXT_CTRLS`, :c:func:`VIDIOC_S_EXT_CTRLS` may be used
> > +   interchangeably with :c:func:`VIDIOC_G_CTRL`, :c:func:`VIDIOC_S_CTRL`,
> > +   unless specified otherwise.
> > +
> > +5. Single-plane API (see spec) and applicable structures may be used
> > +   interchangeably with Multi-plane API, unless specified otherwise,
> > +   depending on encoder capabilities and following the general V4L2
> > +   guidelines.
> > +
> > +6. i = [a..b]: sequence of integers from a to b, inclusive, i.e. i =
> > +   [0..2]: i = 0, 1, 2.
> > +
> > +7. Given an ``OUTPUT`` buffer A, A' represents a buffer on the ``CAPTURE``
> > +   queue containing data (encoded frame/stream) that resulted from processing
> > +   buffer A.
>
> The same comments as I mentioned for the previous patch apply to this section.
>

I suppose you mean the decoder patch, right? (Will address those.)

> > +
> > +Glossary
> > +========
> > +
> > +Refer to :ref:`decoder-glossary`.
>
> Ah, you refer to the same glossary. Then my comment about the source resolution
> terms is obviously wrong.
>
> I wonder if it wouldn't be better to split off the sections above into a separate
> HW codec intro section where you explain the differences between stateful/stateless
> encoders and decoders, and add the conventions and glossary.
>
> After that you have the three documents for each variant (later four when we get
> stateless encoders).
>
> Up to you, and it can be done later in a follow-up patch.
>

I'd indeed prefer to do it in a follow up patch, to avoid distracting
this series too much. Agreed that such an intro section would make
sense, though.

> > +
> > +State machine
> > +=============
> > +
> > +.. kernel-render:: DOT
> > +   :alt: DOT digraph of encoder state machine
> > +   :caption: Encoder state machine
> > +
> > +   digraph encoder_state_machine {
> > +       node [shape = doublecircle, label="Encoding"] Encoding;
> > +
> > +       node [shape = circle, label="Initialization"] Initialization;
> > +       node [shape = circle, label="Stopped"] Stopped;
> > +       node [shape = circle, label="Drain"] Drain;
> > +       node [shape = circle, label="Reset"] Reset;
> > +
> > +       node [shape = point]; qi
> > +       qi -> Initialization [ label = "open()" ];
> > +
> > +       Initialization -> Encoding [ label = "Both queues streaming" ];
> > +
> > +       Encoding -> Drain [ label = "V4L2_DEC_CMD_STOP" ];
> > +       Encoding -> Reset [ label = "VIDIOC_STREAMOFF(CAPTURE)" ];
> > +       Encoding -> Stopped [ label = "VIDIOC_STREAMOFF(OUTPUT)" ];
> > +       Encoding -> Encoding;
> > +
> > +       Drain -> Stopped [ label = "All CAPTURE\nbuffers dequeued\nor\nVIDIOC_STREAMOFF(CAPTURE)" ];
> > +       Drain -> Reset [ label = "VIDIOC_STREAMOFF(CAPTURE)" ];
> > +
> > +       Reset -> Encoding [ label = "VIDIOC_STREAMON(CAPTURE)" ];
> > +       Reset -> Initialization [ label = "VIDIOC_REQBUFS(OUTPUT, 0)" ];
> > +
> > +       Stopped -> Encoding [ label = "V4L2_DEC_CMD_START\nor\nVIDIOC_STREAMON(OUTPUT)" ];
> > +       Stopped -> Reset [ label = "VIDIOC_STREAMOFF(CAPTURE)" ];
> > +   }
> > +
> > +Querying capabilities
> > +=====================
> > +
> > +1. To enumerate the set of coded formats supported by the encoder, the
> > +   client may call :c:func:`VIDIOC_ENUM_FMT` on ``CAPTURE``.
> > +
> > +   * The full set of supported formats will be returned, regardless of the
> > +     format set on ``OUTPUT``.
> > +
> > +2. To enumerate the set of supported raw formats, the client may call
> > +   :c:func:`VIDIOC_ENUM_FMT` on ``OUTPUT``.
> > +
> > +   * Only the formats supported for the format currently active on ``CAPTURE``
> > +     will be returned.
> > +
> > +   * In order to enumerate raw formats supported by a given coded format,
> > +     the client must first set that coded format on ``CAPTURE`` and then
> > +     enumerate the formats on ``OUTPUT``.
> > +
> > +3. The client may use :c:func:`VIDIOC_ENUM_FRAMESIZES` to detect supported
> > +   resolutions for a given format, passing desired pixel format in
> > +   :c:type:`v4l2_frmsizeenum` ``pixel_format``.
> > +
> > +   * Values returned by :c:func:`VIDIOC_ENUM_FRAMESIZES` for a coded pixel
> > +     format will include all possible coded resolutions supported by the
> > +     encoder for given coded pixel format.
> > +
> > +   * Values returned by :c:func:`VIDIOC_ENUM_FRAMESIZES` for a raw pixel format
> > +     will include all possible frame buffer resolutions supported by the
> > +     encoder for given raw pixel format and coded format currently set on
> > +     ``CAPTURE``.
> > +
> > +4. Supported profiles and levels for given format, if applicable, may be
>
> format -> the coded format currently set on ``CAPTURE``
>

Ack.

> > +   queried using their respective controls via :c:func:`VIDIOC_QUERYCTRL`.
> > +
> > +5. Any additional encoder capabilities may be discovered by querying
> > +   their respective controls.
> > +
> > +Initialization
> > +==============
> > +
> > +1. **Optional.** Enumerate supported formats and resolutions. See
> > +   `Querying capabilities` above.
>
> Can be dropped IMHO.
>

Ack.

> > +
> > +2. Set a coded format on the ``CAPTURE`` queue via :c:func:`VIDIOC_S_FMT`
> > +
> > +   * **Required fields:**
> > +
> > +     ``type``
> > +         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``CAPTURE``
> > +
> > +     ``pixelformat``
> > +         the coded format to be produced
> > +
> > +     ``sizeimage``
> > +         desired size of ``CAPTURE`` buffers; the encoder may adjust it to
> > +         match hardware requirements
> > +
> > +     other fields
> > +         follow standard semantics
> > +
> > +   * **Return fields:**
> > +
> > +     ``sizeimage``
> > +         adjusted size of ``CAPTURE`` buffers
> > +
> > +   .. warning::
> > +
> > +      Changing the ``CAPTURE`` format may change the currently set ``OUTPUT``
> > +      format. The encoder will derive a new ``OUTPUT`` format from the
> > +      ``CAPTURE`` format being set, including resolution, colorimetry
> > +      parameters, etc. If the client needs a specific ``OUTPUT`` format, it
> > +      must adjust it afterwards.
> > +
> > +3. **Optional.** Enumerate supported ``OUTPUT`` formats (raw formats for
> > +   source) for the selected coded format via :c:func:`VIDIOC_ENUM_FMT`.
>
> Does this return the same set of formats as in the 'Querying Capabilities' phase?
>

It's actually an interesting question. At this point we wouldn't have
the OUTPUT resolution set yet, so that would be the same set as in the
initial query. If we set the resolution (with some arbitrary
pixelformat), it may become a subset...

> > +
> > +   * **Required fields:**
> > +
> > +     ``type``
> > +         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``OUTPUT``
> > +
> > +     other fields
> > +         follow standard semantics
> > +
> > +   * **Return fields:**
> > +
> > +     ``pixelformat``
> > +         raw format supported for the coded format currently selected on
> > +         the ``OUTPUT`` queue.
>
> OUTPUT -> CAPTURE
>

Ack.

> > +
> > +     other fields
> > +         follow standard semantics
> > +
> > +4. Set the raw source format on the ``OUTPUT`` queue via
> > +   :c:func:`VIDIOC_S_FMT`.
> > +
> > +   * **Required fields:**
> > +
> > +     ``type``
> > +         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``OUTPUT``
> > +
> > +     ``pixelformat``
> > +         raw format of the source
> > +
> > +     ``width``, ``height``
> > +         source resolution
> > +
> > +     other fields
> > +         follow standard semantics
> > +
> > +   * **Return fields:**
> > +
> > +     ``width``, ``height``
> > +         may be adjusted by encoder to match alignment requirements, as
> > +         required by the currently selected formats
> > +
> > +     other fields
> > +         follow standard semantics
> > +
> > +   * Setting the source resolution will reset the selection rectangles to their
> > +     default values, based on the new resolution, as described in the step 5
> > +     below.
> > +
> > +5. **Optional.** Set the visible resolution for the stream metadata via
> > +   :c:func:`VIDIOC_S_SELECTION` on the ``OUTPUT`` queue.
> > +
> > +   * **Required fields:**
> > +
> > +     ``type``
> > +         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``OUTPUT``
> > +
> > +     ``target``
> > +         set to ``V4L2_SEL_TGT_CROP``
> > +
> > +     ``r.left``, ``r.top``, ``r.width``, ``r.height``
> > +         visible rectangle; this must fit within the `V4L2_SEL_TGT_CROP_BOUNDS`
> > +         rectangle and may be subject to adjustment to match codec and
> > +         hardware constraints
> > +
> > +   * **Return fields:**
> > +
> > +     ``r.left``, ``r.top``, ``r.width``, ``r.height``
> > +         visible rectangle adjusted by the encoder
> > +
> > +   * The following selection targets are supported on ``OUTPUT``:
> > +
> > +     ``V4L2_SEL_TGT_CROP_BOUNDS``
> > +         equal to the full source frame, matching the active ``OUTPUT``
> > +         format
> > +
> > +     ``V4L2_SEL_TGT_CROP_DEFAULT``
> > +         equal to ``V4L2_SEL_TGT_CROP_BOUNDS``
> > +
> > +     ``V4L2_SEL_TGT_CROP``
> > +         rectangle within the source buffer to be encoded into the
> > +         ``CAPTURE`` stream; defaults to ``V4L2_SEL_TGT_CROP_DEFAULT``
>
> Since this defaults to the CROP_DEFAULT rectangle this means that if you have
> a 16x16 macroblock size and you want to encode 1080p, you will always have to
> explicitly set the CROP rectangle to 1920x1080, right? Since the default will
> be 1088 instead of 1080.

Not necessarily. It depends on whether the encoder needs the source
buffers to be aligned to macroblocks or not.

>
> It is probably wise to explicitly mention this.
>

Sounds reasonable to add a note.

> > +
> > +     ``V4L2_SEL_TGT_COMPOSE_BOUNDS``
> > +         maximum rectangle within the coded resolution, which the cropped
> > +         source frame can be output into; if the hardware does not support
>
> output -> composed
>

Ack.

> > +         composition or scaling, then this is always equal to the rectangle of
> > +         width and height matching ``V4L2_SEL_TGT_CROP`` and located at (0, 0)
> > +
> > +     ``V4L2_SEL_TGT_COMPOSE_DEFAULT``
> > +         equal to a rectangle of width and height matching
> > +         ``V4L2_SEL_TGT_CROP`` and located at (0, 0)
> > +
> > +     ``V4L2_SEL_TGT_COMPOSE``
> > +         rectangle within the coded frame, which the cropped source frame
> > +         is to be output into; defaults to
>
> output -> composed
>

Ack.

> > +         ``V4L2_SEL_TGT_COMPOSE_DEFAULT``; read-only on hardware without
> > +         additional compose/scaling capabilities; resulting stream will
> > +         have this rectangle encoded as the visible rectangle in its
> > +         metadata
> > +
> > +   .. warning::
> > +
> > +      The encoder may adjust the crop/compose rectangles to the nearest
> > +      supported ones to meet codec and hardware requirements. The client needs
> > +      to check the adjusted rectangle returned by :c:func:`VIDIOC_S_SELECTION`.
> > +
> > +6. Allocate buffers for both ``OUTPUT`` and ``CAPTURE`` via
> > +   :c:func:`VIDIOC_REQBUFS`. This may be performed in any order.
> > +
> > +   * **Required fields:**
> > +
> > +     ``count``
> > +         requested number of buffers to allocate; greater than zero
> > +
> > +     ``type``
> > +         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``OUTPUT`` or
> > +         ``CAPTURE``
> > +
> > +     other fields
> > +         follow standard semantics
> > +
> > +   * **Return fields:**
> > +
> > +     ``count``
> > +          actual number of buffers allocated
> > +
> > +   .. warning::
> > +
> > +      The actual number of allocated buffers may differ from the ``count``
> > +      given. The client must check the updated value of ``count`` after the
> > +      call returns.
> > +
> > +   .. note::
> > +
> > +      To allocate more than the minimum number of buffers (for pipeline depth),
> > +      the client may query the ``V4L2_CID_MIN_BUFFERS_FOR_OUTPUT`` or
> > +      ``V4L2_CID_MIN_BUFFERS_FOR_CAPTURE`` control respectively, to get the
> > +      minimum number of buffers required by the encoder/format, and pass the
> > +      obtained value plus the number of additional buffers needed in the
> > +      ``count`` field to :c:func:`VIDIOC_REQBUFS`.
>
> Does V4L2_CID_MIN_BUFFERS_FOR_CAPTURE make any sense for encoders?
>

Indeed, I don't think it makes much sense. Let me drop it.

> V4L2_CID_MIN_BUFFERS_FOR_OUTPUT can make sense depending on GOP size etc.
>

Yep.


> > +
> > +   Alternatively, :c:func:`VIDIOC_CREATE_BUFS` can be used to have more
> > +   control over buffer allocation.
> > +
> > +   * **Required fields:**
> > +
> > +     ``count``
> > +         requested number of buffers to allocate; greater than zero
> > +
> > +     ``type``
> > +         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``OUTPUT``
> > +
> > +     other fields
> > +         follow standard semantics
> > +
> > +   * **Return fields:**
> > +
> > +     ``count``
> > +         adjusted to the number of allocated buffers
> > +
> > +7. Begin streaming on both ``OUTPUT`` and ``CAPTURE`` queues via
> > +   :c:func:`VIDIOC_STREAMON`. This may be performed in any order. The actual
> > +   encoding process starts when both queues start streaming.
> > +
> > +.. note::
> > +
> > +   If the client stops the ``CAPTURE`` queue during the encode process and then
> > +   restarts it again, the encoder will begin generating a stream independent
> > +   from the stream generated before the stop. The exact constraints depend
> > +   on the coded format, but may include the following implications:
> > +
> > +   * encoded frames produced after the restart must not reference any
> > +     frames produced before the stop, e.g. no long term references for
> > +     H.264,
> > +
> > +   * any headers that must be included in a standalone stream must be
> > +     produced again, e.g. SPS and PPS for H.264.
> > +
> > +Encoding
> > +========
> > +
> > +This state is reached after the `Initialization` sequence finishes succesfully.
>
> successfully
>

Ack.

> > +In this state, client queues and dequeues buffers to both queues via
>
> client -> the client
>

Ack.

> > +:c:func:`VIDIOC_QBUF` and :c:func:`VIDIOC_DQBUF`, following the standard
> > +semantics.
> > +
> > +The contents of encoded ``CAPTURE`` buffers depend on the active coded pixel
>
> contents ... depend -> content ... depends
>

The plural sounds more natural to me to be honest.

> > +format and may be affected by codec-specific extended controls, as stated
> > +in the documentation of each format.
> > +
> > +Both queues operate independently, following standard behavior of V4L2 buffer
> > +queues and memory-to-memory devices. In addition, the order of encoded frames
> > +dequeued from the ``CAPTURE`` queue may differ from the order of queuing raw
> > +frames to the ``OUTPUT`` queue, due to properties of the selected coded format,
> > +e.g. frame reordering.
> > +
> > +The client must not assume any direct relationship between ``CAPTURE`` and
> > +``OUTPUT`` buffers and any specific timing of buffers becoming
> > +available to dequeue. Specifically,
> > +
> > +* a buffer queued to ``OUTPUT`` may result in more than 1 buffer produced on
> > +  ``CAPTURE`` (if returning an encoded frame allowed the encoder to return a
> > +  frame that preceded it in display, but succeeded it in the decode order),
> > +
> > +* a buffer queued to ``OUTPUT`` may result in a buffer being produced on
> > +  ``CAPTURE`` later into encode process, and/or after processing further
> > +  ``OUTPUT`` buffers, or be returned out of order, e.g. if display
> > +  reordering is used,
> > +
> > +* buffers may become available on the ``CAPTURE`` queue without additional
> > +  buffers queued to ``OUTPUT`` (e.g. during drain or ``EOS``), because of the
> > +  ``OUTPUT`` buffers queued in the past whose decoding results are only
> > +  available at later time, due to specifics of the decoding process,
> > +
> > +* buffers queued to ``OUTPUT`` may not become available to dequeue instantly
> > +  after being encoded into a corresponding ``CATPURE`` buffer, e.g. if the
> > +  encoder needs to use the frame as a reference for encoding further frames.
> > +
> > +.. note::
> > +
> > +   To allow matching encoded ``CAPTURE`` buffers with ``OUTPUT`` buffers they
> > +   originated from, the client can set the ``timestamp`` field of the
> > +   :c:type:`v4l2_buffer` struct when queuing an ``OUTPUT`` buffer. The
> > +   ``CAPTURE`` buffer(s), which resulted from encoding that ``OUTPUT`` buffer
> > +   will have their ``timestamp`` field set to the same value when dequeued.
> > +
> > +   In addition to the straighforward case of one ``OUTPUT`` buffer producing
> > +   one ``CAPTURE`` buffer, the following cases are defined:
> > +
> > +   * one ``OUTPUT`` buffer generates multiple ``CAPTURE`` buffers: the same
> > +     ``OUTPUT`` timestamp will be copied to multiple ``CAPTURE`` buffers,
> > +
> > +   * the encoding order differs from the presentation order (i.e. the
> > +     ``CAPTURE`` buffers are out-of-order compared to the ``OUTPUT`` buffers):
> > +     ``CAPTURE`` timestamps will not retain the order of ``OUTPUT`` timestamps
> > +     and thus monotonicity of the timestamps cannot be guaranteed.
> > +
> > +.. note::
> > +
> > +   To let the client distinguish between frame types (keyframes, intermediate
> > +   frames; the exact list of types depends on the coded format), the
> > +   ``CAPTURE`` buffers will have corresponding flag bits set in their
> > +   :c:type:`v4l2_buffer` struct when dequeued. See the documentation of
> > +   :c:type:`v4l2_buffer` and each coded pixel format for exact list of flags
> > +   and their meanings.
>
> Is this required? (I think it should, but it isn't the case today).
>

At least V4L2_BUF_FLAG_KEYFRAME has been always required for Chromium,
as it's indispensable for any kind of real time streaming, e.g.
Hangouts. (Although technically the streaming layer could parse the
stream and pick it up itself...) I can see s5p-mfc, coda, venus and
mtk-vcodec supporting at least this one. mtk-vcodec doesn't seem to
support the other ones.

Should we make only the V4L2_BUF_FLAG_KEYFRAME mandatory, at least for now?

> Is the current set of buffer flags (Key/B/P frame) sufficient for the current
> set of codecs?
>

No, there is no way to distinguish between I and IDR frames in H.264.
You can see more details on our issue tracker: crbug.com/868792.

> > +
> > +Encoding parameter changes
> > +==========================
> > +
> > +The client is allowed to use :c:func:`VIDIOC_S_CTRL` to change encoder
> > +parameters at any time. The availability of parameters is encoder-specific
> > +and the client must query the encoder to find the set of available controls.
> > +
> > +The ability to change each parameter during encoding is encoder-specific, as per
> > +the standard semantics of the V4L2 control interface. The client may attempt
> > +setting a control of its interest during encoding and if the operation fails
>
> I'd simplify this:
>
> The client may attempt to set a control during encoding...

Ack.

>
> > +with the -EBUSY error code, the ``CAPTURE`` queue needs to be stopped for the
> > +configuration change to be allowed (following the `Drain` sequence will be
> > +needed to avoid losing the already queued/encoded frames).
>
> Rephrase:
>
> ...to be allowed. To do this follow the `Drain` sequence to avoid losing the
> already queued/encoded frames.
>

How about "To do this, it may follow...", to keep client as the third
person convention?

> > +
> > +The timing of parameter updates is encoder-specific, as per the standard
> > +semantics of the V4L2 control interface. If the client needs to apply the
> > +parameters exactly at specific frame, using the Request API should be
>
> Change this to a reference to the Request API section.
>

Do you mean just adding a reference or some other changes too?

> > +considered, if supported by the encoder.
> > +
> > +Drain
> > +=====
> > +
> > +To ensure that all the queued ``OUTPUT`` buffers have been processed and the
> > +related ``CAPTURE`` buffers output to the client, the client must follow the
>
> output -> are output
>
> or perhaps better (up to you): are given
>

Ack (went with the latter and updated the decoder too).

> > +drain sequence described below. After the drain sequence ends, the client has
> > +received all encoded frames for all ``OUTPUT`` buffers queued before the
> > +sequence was started.
> > +
> > +1. Begin the drain sequence by issuing :c:func:`VIDIOC_ENCODER_CMD`.
> > +
> > +   * **Required fields:**
> > +
> > +     ``cmd``
> > +         set to ``V4L2_ENC_CMD_STOP``
> > +
> > +     ``flags``
> > +         set to 0
> > +
> > +     ``pts``
> > +         set to 0
> > +
> > +   .. warning::
> > +
> > +   The sentence can be only initiated if both ``OUTPUT`` and ``CAPTURE`` queues
> > +   are streaming. For compatibility reasons, the call to
> > +   :c:func:`VIDIOC_ENCODER_CMD` will not fail even if any of the queues is not
> > +   streaming, but at the same time it will not initiate the `Drain` sequence
> > +   and so the steps described below would not be applicable.
> > +
> > +2. Any ``OUTPUT`` buffers queued by the client before the
> > +   :c:func:`VIDIOC_ENCODER_CMD` was issued will be processed and encoded as
> > +   normal. The client must continue to handle both queues independently,
> > +   similarly to normal encode operation. This includes,
> > +
> > +   * queuing and dequeuing ``CAPTURE`` buffers, until a buffer marked with the
> > +     ``V4L2_BUF_FLAG_LAST`` flag is dequeued,
> > +
> > +     .. warning::
> > +
> > +        The last buffer may be empty (with :c:type:`v4l2_buffer`
> > +        ``bytesused`` = 0) and in such case it must be ignored by the client,
>
> such -> that
>
> Check the previous patch as well if you used the phrase 'such case' and replace
> it with 'that case'.
>

Ack.

> > +        as it does not contain an encoded frame.
> > +
> > +     .. note::
> > +
> > +        Any attempt to dequeue more buffers beyond the buffer marked with
> > +        ``V4L2_BUF_FLAG_LAST`` will result in a -EPIPE error from
> > +        :c:func:`VIDIOC_DQBUF`.
> > +
> > +   * dequeuing processed ``OUTPUT`` buffers, until all the buffers queued
> > +     before the ``V4L2_ENC_CMD_STOP`` command are dequeued,
> > +
> > +   * dequeuing the ``V4L2_EVENT_EOS`` event, if the client subscribes to it.
> > +
> > +   .. note::
> > +
> > +      For backwards compatibility, the encoder will signal a ``V4L2_EVENT_EOS``
> > +      event when the last the last frame has been decoded and all frames are
>
> the last the last -> the last
>

Ack.

> > +      ready to be dequeued. It is a deprecated behavior and the client must not
>
> is a -> is

Ack.

>
> > +      rely on it. The ``V4L2_BUF_FLAG_LAST`` buffer flag should be used
> > +      instead.
>
> Question: should new codec drivers still implement the EOS event?
>

Good question. It's a userspace compatibility issue, so if we intend a
new codec driver to work with old userspace, it must do so. Perhaps up
to the driver author/maintainer?

> > +
> > +3. Once all ``OUTPUT`` buffers queued before the ``V4L2_ENC_CMD_STOP`` call and
> > +   the last ``CAPTURE`` buffer are dequeued, the encoder is stopped and it will
> > +   accept, but not process any newly queued ``OUTPUT`` buffers until the client
> > +   issues any of the following operations:
> > +
> > +   * ``V4L2_ENC_CMD_START`` - the encoder will resume operation normally,
>
> Perhaps mention that this does not reset the encoder? It's not immediately clear
> when reading this.
>

Ack.

> > +
> > +   * a pair of :c:func:`VIDIOC_STREAMOFF` and :c:func:`VIDIOC_STREAMON` on the
> > +     ``CAPTURE`` queue - the encoder will be reset (see the `Reset` sequence)
> > +     and then resume encoding,
> > +
> > +   * a pair of :c:func:`VIDIOC_STREAMOFF` and :c:func:`VIDIOC_STREAMON` on the
> > +     ``OUTPUT`` queue - the encoder will resume operation normally, however any
> > +     source frames queued to the ``OUTPUT`` queue between ``V4L2_ENC_CMD_STOP``
> > +     and :c:func:`VIDIOC_STREAMOFF` will be discarded.
> > +
> > +.. note::
> > +
> > +   Once the drain sequence is initiated, the client needs to drive it to
> > +   completion, as described by the steps above, unless it aborts the process by
> > +   issuing :c:func:`VIDIOC_STREAMOFF` on any of the ``OUTPUT`` or ``CAPTURE``
> > +   queues.  The client is not allowed to issue ``V4L2_ENC_CMD_START`` or
> > +   ``V4L2_ENC_CMD_STOP`` again while the drain sequence is in progress and they
> > +   will fail with -EBUSY error code if attempted.
> > +
> > +   Although mandatory, the availability of encoder commands may be queried
> > +   using :c:func:`VIDIOC_TRY_ENCODER_CMD`.
> > +
> > +Reset
> > +=====
> > +
> > +The client may want to request the encoder to reinitialize the encoding, so
> > +that the following stream data becomes independent from the stream data
> > +generated before. Depending on the coded format, that may imply that,
>
> that, -> that:
>

Ack. (And also few other places.)

> > +
> > +* encoded frames produced after the restart must not reference any frames
> > +  produced before the stop, e.g. no long term references for H.264,
> > +
> > +* any headers that must be included in a standalone stream must be produced
> > +  again, e.g. SPS and PPS for H.264.
> > +
> > +This can be achieved by performing the reset sequence.
> > +
> > +1. Perform the `Drain` sequence to ensure all the in-flight encoding finishes
> > +   and respective buffers are dequeued.
> > +
> > +2. Stop streaming on the ``CAPTURE`` queue via :c:func:`VIDIOC_STREAMOFF`. This
> > +   will return all currently queued ``CAPTURE`` buffers to the client, without
> > +   valid frame data.
> > +
> > +3. Start streaming on the ``CAPTURE`` queue via :c:func:`VIDIOC_STREAMON` and
> > +   continue with regular encoding sequence. The encoded frames produced into
> > +   ``CAPTURE`` buffers from now on will contain a standalone stream that can be
> > +   decoded without the need for frames encoded before the reset sequence,
> > +   starting at the first ``OUTPUT`` buffer queued after issuing the
> > +   `V4L2_ENC_CMD_STOP` of the `Drain` sequence.
> > +
> > +This sequence may be also used to change encoding parameters for encoders
> > +without the ability to change the parameters on the fly.
> > +
> > +Commit points
> > +=============
> > +
> > +Setting formats and allocating buffers triggers changes in the behavior of the
> > +encoder.
> > +
> > +1. Setting the format on the ``CAPTURE`` queue may change the set of formats
> > +   supported/advertised on the ``OUTPUT`` queue. In particular, it also means
> > +   that the ``OUTPUT`` format may be reset and the client must not rely on the
> > +   previously set format being preserved.
> > +
> > +2. Enumerating formats on the ``OUTPUT`` queue always returns only formats
> > +   supported for the current ``CAPTURE`` format.
> > +
> > +3. Setting the format on the ``OUTPUT`` queue does not change the list of
> > +   formats available on the ``CAPTURE`` queue. An attempt to set the ``OUTPUT``
> > +   format that is not supported for the currently selected ``CAPTURE`` format
> > +   will result in the encoder adjusting the requested ``OUTPUT`` format to a
> > +   supported one.
> > +
> > +4. Enumerating formats on the ``CAPTURE`` queue always returns the full set of
> > +   supported coded formats, irrespectively of the current ``OUTPUT`` format.
> > +
> > +5. While buffers are allocated on the ``CAPTURE`` queue, the client must not
> > +   change the format on the queue. Drivers will return the -EBUSY error code
> > +   for any such format change attempt.
> > +
> > +To summarize, setting formats and allocation must always start with the
> > +``CAPTURE`` queue and the ``CAPTURE`` queue is the master that governs the
> > +set of supported formats for the ``OUTPUT`` queue.
> > diff --git a/Documentation/media/uapi/v4l/devices.rst b/Documentation/media/uapi/v4l/devices.rst
> > index 12d43fe711cf..1822c66c2154 100644
> > --- a/Documentation/media/uapi/v4l/devices.rst
> > +++ b/Documentation/media/uapi/v4l/devices.rst
> > @@ -16,6 +16,7 @@ Interfaces
> >      dev-osd
> >      dev-codec
> >      dev-decoder
> > +    dev-encoder
> >      dev-effect
> >      dev-raw-vbi
> >      dev-sliced-vbi
> > diff --git a/Documentation/media/uapi/v4l/pixfmt-v4l2.rst b/Documentation/media/uapi/v4l/pixfmt-v4l2.rst
> > index ca5f2270a829..085089cd9577 100644
> > --- a/Documentation/media/uapi/v4l/pixfmt-v4l2.rst
> > +++ b/Documentation/media/uapi/v4l/pixfmt-v4l2.rst
> > @@ -37,6 +37,11 @@ Single-planar format structure
> >       inside the stream, when fed to a stateful mem2mem decoder, the fields
> >       may be zero to rely on the decoder to detect the right values. For more
> >       details see :ref:`decoder` and format descriptions.
> > +
> > +     For compressed formats on the CAPTURE side of a stateful mem2mem
> > +     encoder, the fields must be zero, since the coded size is expected to
> > +     be calculated internally by the encoder itself, based on the OUTPUT
> > +     side. For more details see :ref:`encoder` and format descriptions.
>
> The encoder document doesn't actually mention this. I think it should, though.

Indeed. I'll make it say that the fields are "ignored (always zero)".

To be honest, I wanted to define them in a way that they would be
hardwired to the internal coded size selected by the driver, but it
just complicated things, since one would need to set the CAPTURE
format first, then the OUTPUT format, selection rectangles and only
then could read back the coded resolution from the CAPTURE format. It
would have also violated the assumption that CAPTURE format was not
expected to be altered by changing OUTPUT format.

>
> I'm a bit uncertain about this: the expected resolution might impact the
> sizeimage value: i.e. encoding 640x480 requires much less memory then
> encoding 4k video. If this is required to be 0x0, then the driver has to
> fill in a worst-case sizeimage value. It might make more sense to say that
> if a non-zero resolution is given, then the driver will attempt to
> calculate a sensible sizeimage value.

The driver would still be able to determine the sizeimage by the
internally known coded size, which it calculated based on OUTPUT
format, selection, codec constraints, etc. It's not something for the
userspace to provide (nor it would be able to provide).

Best regards,
Tomasz
Tomasz Figa Jan. 23, 2019, 10 a.m. UTC | #6
On Sat, Nov 17, 2018 at 8:37 PM Hans Verkuil <hverkuil@xs4all.nl> wrote:
>
> On 11/17/2018 05:18 AM, Nicolas Dufresne wrote:
> > Le lundi 12 novembre 2018 à 14:23 +0100, Hans Verkuil a écrit :
> >> On 10/22/2018 04:49 PM, Tomasz Figa wrote:
[snip]
> >>> +      rely on it. The ``V4L2_BUF_FLAG_LAST`` buffer flag should be used
> >>> +      instead.
> >>
> >> Question: should new codec drivers still implement the EOS event?
> >
> > I'm been asking around, but I think here is a good place. Do we really
> > need the FLAG_LAST in userspace ? Userspace can also wait for the first
> > EPIPE return from DQBUF.
>
> I'm interested in hearing Tomasz' opinion. This flag is used already, so there
> definitely is a backwards compatibility issue here.
>

FWIW, it would add the overhead of 1 more system call, although I
don't think it's of our concern.

My personal feeling is that using error codes for signaling normal
conditions isn't very elegant, though.

> >
> >>
> >>> +
> >>> +3. Once all ``OUTPUT`` buffers queued before the ``V4L2_ENC_CMD_STOP`` call and
> >>> +   the last ``CAPTURE`` buffer are dequeued, the encoder is stopped and it will
> >>> +   accept, but not process any newly queued ``OUTPUT`` buffers until the client
> >>> +   issues any of the following operations:
> >>> +
> >>> +   * ``V4L2_ENC_CMD_START`` - the encoder will resume operation normally,
> >>
> >> Perhaps mention that this does not reset the encoder? It's not immediately clear
> >> when reading this.
> >
> > Which drivers supports this ? I believe I tried with Exynos in the
> > past, and that didn't work. How do we know if a driver supports this or
> > not. Do we make it mandatory ? When it's not supported, it basically
> > mean userspace need to cache and resend the header in userspace, and
> > also need to skip to some sync point.
>
> Once we agree on the spec, then the next step will be to add good compliance
> checks and update drivers that fail the tests.
>
> To check if the driver support this ioctl you can call VIDIOC_TRY_ENCODER_CMD
> to see if the functionality is supported.

There is nothing here for the hardware to support. It's an entirely
driver thing, since it just needs to wait for the encoder to complete
all the pending frames and stop enqueuing more frames to the decoder
until V4L2_ENC_CMD_START is called. Any driver that can't do it must
be fixed, since otherwise you have no way to ensure that you got all
the encoded output.

Best regards,
Tomasz
Tomasz Figa Jan. 23, 2019, 10:02 a.m. UTC | #7
On Sun, Nov 18, 2018 at 10:34 AM Nicolas Dufresne <nicolas@ndufresne.ca> wrote:
>
> Le samedi 17 novembre 2018 à 12:37 +0100, Hans Verkuil a écrit :
> > > > Does V4L2_CID_MIN_BUFFERS_FOR_CAPTURE make any sense for encoders?
> > >
> > > We do account for it in GStreamer (the capture/output handling is
> > > generic), but I don't know if it's being used anywhere.
> >
> > Do you use this value directly for REQBUFS, or do you use it as the minimum
> > value but in practice use more buffers?
>
> We add more buffers to that value. We assume this value is what will be
> held by the driver, hence without adding some buffers, the driver would
> go idle as soon as one is dequeued. We also need to allocate for the
> importing driver.
>
> In general, if we have a pipeline with Driver A sending to Driver B,
> both driver will require a certain amount of buffers to operate. E.g.
> with DRM display, the driver will hold on 1 buffer (the scannout
> buffer).
>
> In GStreamer, it's implemented generically, so we do:
>
>   MIN_BUFFERS_FOR + remote_min + 1
>
> If only MIN_BUFFERS_FOR was allocated, ignoring remote driver
> requirement, the streaming will likely get stuck.

What happens if the driver doesn't report it?

Best regards,
Tomasz
Hans Verkuil Jan. 23, 2019, 11:28 a.m. UTC | #8
On 01/23/19 11:00, Tomasz Figa wrote:
> On Sat, Nov 17, 2018 at 8:37 PM Hans Verkuil <hverkuil@xs4all.nl> wrote:
>>
>> On 11/17/2018 05:18 AM, Nicolas Dufresne wrote:
>>> Le lundi 12 novembre 2018 à 14:23 +0100, Hans Verkuil a écrit :
>>>> On 10/22/2018 04:49 PM, Tomasz Figa wrote:
> [snip]
>>>>> +      rely on it. The ``V4L2_BUF_FLAG_LAST`` buffer flag should be used
>>>>> +      instead.
>>>>
>>>> Question: should new codec drivers still implement the EOS event?
>>>
>>> I'm been asking around, but I think here is a good place. Do we really
>>> need the FLAG_LAST in userspace ? Userspace can also wait for the first
>>> EPIPE return from DQBUF.
>>
>> I'm interested in hearing Tomasz' opinion. This flag is used already, so there
>> definitely is a backwards compatibility issue here.
>>
> 
> FWIW, it would add the overhead of 1 more system call, although I
> don't think it's of our concern.
> 
> My personal feeling is that using error codes for signaling normal
> conditions isn't very elegant, though.

I agree. Let's keep this flag.

Regards,

	Hans

> 
>>>
>>>>
>>>>> +
>>>>> +3. Once all ``OUTPUT`` buffers queued before the ``V4L2_ENC_CMD_STOP`` call and
>>>>> +   the last ``CAPTURE`` buffer are dequeued, the encoder is stopped and it will
>>>>> +   accept, but not process any newly queued ``OUTPUT`` buffers until the client
>>>>> +   issues any of the following operations:
>>>>> +
>>>>> +   * ``V4L2_ENC_CMD_START`` - the encoder will resume operation normally,
>>>>
>>>> Perhaps mention that this does not reset the encoder? It's not immediately clear
>>>> when reading this.
>>>
>>> Which drivers supports this ? I believe I tried with Exynos in the
>>> past, and that didn't work. How do we know if a driver supports this or
>>> not. Do we make it mandatory ? When it's not supported, it basically
>>> mean userspace need to cache and resend the header in userspace, and
>>> also need to skip to some sync point.
>>
>> Once we agree on the spec, then the next step will be to add good compliance
>> checks and update drivers that fail the tests.
>>
>> To check if the driver support this ioctl you can call VIDIOC_TRY_ENCODER_CMD
>> to see if the functionality is supported.
> 
> There is nothing here for the hardware to support. It's an entirely
> driver thing, since it just needs to wait for the encoder to complete
> all the pending frames and stop enqueuing more frames to the decoder
> until V4L2_ENC_CMD_START is called. Any driver that can't do it must
> be fixed, since otherwise you have no way to ensure that you got all
> the encoded output.
> 
> Best regards,
> Tomasz
>
Hans Verkuil Jan. 23, 2019, 1:04 p.m. UTC | #9
On 01/23/19 10:52, Tomasz Figa wrote:
> On Mon, Nov 12, 2018 at 10:23 PM Hans Verkuil <hverkuil@xs4all.nl> wrote:
>>
>> On 10/22/2018 04:49 PM, Tomasz Figa wrote:
>>> Due to complexity of the video encoding process, the V4L2 drivers of
>>> stateful encoder hardware require specific sequences of V4L2 API calls
>>> to be followed. These include capability enumeration, initialization,
>>> encoding, encode parameters change, drain and reset.
>>>
>>> Specifics of the above have been discussed during Media Workshops at
>>> LinuxCon Europe 2012 in Barcelona and then later Embedded Linux
>>> Conference Europe 2014 in Düsseldorf. The de facto Codec API that
>>> originated at those events was later implemented by the drivers we already
>>> have merged in mainline, such as s5p-mfc or coda.
>>>
>>> The only thing missing was the real specification included as a part of
>>> Linux Media documentation. Fix it now and document the encoder part of
>>> the Codec API.
>>>
>>> Signed-off-by: Tomasz Figa <tfiga@chromium.org>
>>> ---
>>>  Documentation/media/uapi/v4l/dev-encoder.rst  | 579 ++++++++++++++++++
>>>  Documentation/media/uapi/v4l/devices.rst      |   1 +
>>>  Documentation/media/uapi/v4l/pixfmt-v4l2.rst  |   5 +
>>>  Documentation/media/uapi/v4l/v4l2.rst         |   2 +
>>>  .../media/uapi/v4l/vidioc-encoder-cmd.rst     |  38 +-
>>>  5 files changed, 610 insertions(+), 15 deletions(-)
>>>  create mode 100644 Documentation/media/uapi/v4l/dev-encoder.rst
>>>
>>> diff --git a/Documentation/media/uapi/v4l/dev-encoder.rst b/Documentation/media/uapi/v4l/dev-encoder.rst
>>> new file mode 100644
>>> index 000000000000..41139e5e48eb
>>> --- /dev/null
>>> +++ b/Documentation/media/uapi/v4l/dev-encoder.rst
>>> @@ -0,0 +1,579 @@
>>> +.. -*- coding: utf-8; mode: rst -*-
>>> +
>>> +.. _encoder:
>>> +
>>> +*************************************************
>>> +Memory-to-memory Stateful Video Encoder Interface
>>> +*************************************************
>>> +
>>> +A stateful video encoder takes raw video frames in display order and encodes
>>> +them into a bitstream. It generates complete chunks of the bitstream, including
>>> +all metadata, headers, etc. The resulting bitstream does not require any
>>> +further post-processing by the client.
>>> +
>>> +Performing software stream processing, header generation etc. in the driver
>>> +in order to support this interface is strongly discouraged. In case such
>>> +operations are needed, use of the Stateless Video Encoder Interface (in
>>> +development) is strongly advised.
>>> +
>>> +Conventions and notation used in this document
>>> +==============================================
>>> +
>>> +1. The general V4L2 API rules apply if not specified in this document
>>> +   otherwise.
>>> +
>>> +2. The meaning of words "must", "may", "should", etc. is as per RFC
>>> +   2119.
>>> +
>>> +3. All steps not marked "optional" are required.
>>> +
>>> +4. :c:func:`VIDIOC_G_EXT_CTRLS`, :c:func:`VIDIOC_S_EXT_CTRLS` may be used
>>> +   interchangeably with :c:func:`VIDIOC_G_CTRL`, :c:func:`VIDIOC_S_CTRL`,
>>> +   unless specified otherwise.
>>> +
>>> +5. Single-plane API (see spec) and applicable structures may be used
>>> +   interchangeably with Multi-plane API, unless specified otherwise,
>>> +   depending on encoder capabilities and following the general V4L2
>>> +   guidelines.
>>> +
>>> +6. i = [a..b]: sequence of integers from a to b, inclusive, i.e. i =
>>> +   [0..2]: i = 0, 1, 2.
>>> +
>>> +7. Given an ``OUTPUT`` buffer A, A' represents a buffer on the ``CAPTURE``
>>> +   queue containing data (encoded frame/stream) that resulted from processing
>>> +   buffer A.
>>
>> The same comments as I mentioned for the previous patch apply to this section.
>>
> 
> I suppose you mean the decoder patch, right? (Will address those.)

Yes.

> 
>>> +
>>> +Glossary
>>> +========
>>> +
>>> +Refer to :ref:`decoder-glossary`.
>>
>> Ah, you refer to the same glossary. Then my comment about the source resolution
>> terms is obviously wrong.
>>
>> I wonder if it wouldn't be better to split off the sections above into a separate
>> HW codec intro section where you explain the differences between stateful/stateless
>> encoders and decoders, and add the conventions and glossary.
>>
>> After that you have the three documents for each variant (later four when we get
>> stateless encoders).
>>
>> Up to you, and it can be done later in a follow-up patch.
>>
> 
> I'd indeed prefer to do it in a follow up patch, to avoid distracting
> this series too much. Agreed that such an intro section would make
> sense, though.

A follow-up patch is fine.

> 
>>> +
>>> +State machine
>>> +=============
>>> +
>>> +.. kernel-render:: DOT
>>> +   :alt: DOT digraph of encoder state machine
>>> +   :caption: Encoder state machine
>>> +
>>> +   digraph encoder_state_machine {
>>> +       node [shape = doublecircle, label="Encoding"] Encoding;
>>> +
>>> +       node [shape = circle, label="Initialization"] Initialization;
>>> +       node [shape = circle, label="Stopped"] Stopped;
>>> +       node [shape = circle, label="Drain"] Drain;
>>> +       node [shape = circle, label="Reset"] Reset;
>>> +
>>> +       node [shape = point]; qi
>>> +       qi -> Initialization [ label = "open()" ];
>>> +
>>> +       Initialization -> Encoding [ label = "Both queues streaming" ];
>>> +
>>> +       Encoding -> Drain [ label = "V4L2_DEC_CMD_STOP" ];
>>> +       Encoding -> Reset [ label = "VIDIOC_STREAMOFF(CAPTURE)" ];
>>> +       Encoding -> Stopped [ label = "VIDIOC_STREAMOFF(OUTPUT)" ];
>>> +       Encoding -> Encoding;
>>> +
>>> +       Drain -> Stopped [ label = "All CAPTURE\nbuffers dequeued\nor\nVIDIOC_STREAMOFF(CAPTURE)" ];
>>> +       Drain -> Reset [ label = "VIDIOC_STREAMOFF(CAPTURE)" ];
>>> +
>>> +       Reset -> Encoding [ label = "VIDIOC_STREAMON(CAPTURE)" ];
>>> +       Reset -> Initialization [ label = "VIDIOC_REQBUFS(OUTPUT, 0)" ];
>>> +
>>> +       Stopped -> Encoding [ label = "V4L2_DEC_CMD_START\nor\nVIDIOC_STREAMON(OUTPUT)" ];
>>> +       Stopped -> Reset [ label = "VIDIOC_STREAMOFF(CAPTURE)" ];
>>> +   }
>>> +
>>> +Querying capabilities
>>> +=====================
>>> +
>>> +1. To enumerate the set of coded formats supported by the encoder, the
>>> +   client may call :c:func:`VIDIOC_ENUM_FMT` on ``CAPTURE``.
>>> +
>>> +   * The full set of supported formats will be returned, regardless of the
>>> +     format set on ``OUTPUT``.
>>> +
>>> +2. To enumerate the set of supported raw formats, the client may call
>>> +   :c:func:`VIDIOC_ENUM_FMT` on ``OUTPUT``.
>>> +
>>> +   * Only the formats supported for the format currently active on ``CAPTURE``
>>> +     will be returned.
>>> +
>>> +   * In order to enumerate raw formats supported by a given coded format,
>>> +     the client must first set that coded format on ``CAPTURE`` and then
>>> +     enumerate the formats on ``OUTPUT``.
>>> +
>>> +3. The client may use :c:func:`VIDIOC_ENUM_FRAMESIZES` to detect supported
>>> +   resolutions for a given format, passing desired pixel format in
>>> +   :c:type:`v4l2_frmsizeenum` ``pixel_format``.
>>> +
>>> +   * Values returned by :c:func:`VIDIOC_ENUM_FRAMESIZES` for a coded pixel
>>> +     format will include all possible coded resolutions supported by the
>>> +     encoder for given coded pixel format.
>>> +
>>> +   * Values returned by :c:func:`VIDIOC_ENUM_FRAMESIZES` for a raw pixel format
>>> +     will include all possible frame buffer resolutions supported by the
>>> +     encoder for given raw pixel format and coded format currently set on
>>> +     ``CAPTURE``.
>>> +
>>> +4. Supported profiles and levels for given format, if applicable, may be
>>
>> format -> the coded format currently set on ``CAPTURE``
>>
> 
> Ack.
> 
>>> +   queried using their respective controls via :c:func:`VIDIOC_QUERYCTRL`.
>>> +
>>> +5. Any additional encoder capabilities may be discovered by querying
>>> +   their respective controls.
>>> +
>>> +Initialization
>>> +==============
>>> +
>>> +1. **Optional.** Enumerate supported formats and resolutions. See
>>> +   `Querying capabilities` above.
>>
>> Can be dropped IMHO.
>>
> 
> Ack.
> 
>>> +
>>> +2. Set a coded format on the ``CAPTURE`` queue via :c:func:`VIDIOC_S_FMT`
>>> +
>>> +   * **Required fields:**
>>> +
>>> +     ``type``
>>> +         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``CAPTURE``
>>> +
>>> +     ``pixelformat``
>>> +         the coded format to be produced
>>> +
>>> +     ``sizeimage``
>>> +         desired size of ``CAPTURE`` buffers; the encoder may adjust it to
>>> +         match hardware requirements
>>> +
>>> +     other fields
>>> +         follow standard semantics
>>> +
>>> +   * **Return fields:**
>>> +
>>> +     ``sizeimage``
>>> +         adjusted size of ``CAPTURE`` buffers
>>> +
>>> +   .. warning::
>>> +
>>> +      Changing the ``CAPTURE`` format may change the currently set ``OUTPUT``
>>> +      format. The encoder will derive a new ``OUTPUT`` format from the
>>> +      ``CAPTURE`` format being set, including resolution, colorimetry
>>> +      parameters, etc. If the client needs a specific ``OUTPUT`` format, it
>>> +      must adjust it afterwards.
>>> +
>>> +3. **Optional.** Enumerate supported ``OUTPUT`` formats (raw formats for
>>> +   source) for the selected coded format via :c:func:`VIDIOC_ENUM_FMT`.
>>
>> Does this return the same set of formats as in the 'Querying Capabilities' phase?
>>
> 
> It's actually an interesting question. At this point we wouldn't have
> the OUTPUT resolution set yet, so that would be the same set as in the
> initial query. If we set the resolution (with some arbitrary
> pixelformat), it may become a subset...

But doesn't setting the capture format also set the resolution?

To quote from the text above:

"The encoder will derive a new ``OUTPUT`` format from the ``CAPTURE`` format
 being set, including resolution, colorimetry parameters, etc."

So you set the capture format with a resolution (you know that), then
ENUM_FMT will return the subset for that codec and resolution.

But see also the comment at the end of this email.

> 
>>> +
>>> +   * **Required fields:**
>>> +
>>> +     ``type``
>>> +         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``OUTPUT``
>>> +
>>> +     other fields
>>> +         follow standard semantics
>>> +
>>> +   * **Return fields:**
>>> +
>>> +     ``pixelformat``
>>> +         raw format supported for the coded format currently selected on
>>> +         the ``OUTPUT`` queue.
>>
>> OUTPUT -> CAPTURE
>>
> 
> Ack.
> 
>>> +
>>> +     other fields
>>> +         follow standard semantics
>>> +
>>> +4. Set the raw source format on the ``OUTPUT`` queue via
>>> +   :c:func:`VIDIOC_S_FMT`.
>>> +
>>> +   * **Required fields:**
>>> +
>>> +     ``type``
>>> +         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``OUTPUT``
>>> +
>>> +     ``pixelformat``
>>> +         raw format of the source
>>> +
>>> +     ``width``, ``height``
>>> +         source resolution
>>> +
>>> +     other fields
>>> +         follow standard semantics
>>> +
>>> +   * **Return fields:**
>>> +
>>> +     ``width``, ``height``
>>> +         may be adjusted by encoder to match alignment requirements, as
>>> +         required by the currently selected formats
>>> +
>>> +     other fields
>>> +         follow standard semantics
>>> +
>>> +   * Setting the source resolution will reset the selection rectangles to their
>>> +     default values, based on the new resolution, as described in the step 5
>>> +     below.
>>> +
>>> +5. **Optional.** Set the visible resolution for the stream metadata via
>>> +   :c:func:`VIDIOC_S_SELECTION` on the ``OUTPUT`` queue.
>>> +
>>> +   * **Required fields:**
>>> +
>>> +     ``type``
>>> +         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``OUTPUT``
>>> +
>>> +     ``target``
>>> +         set to ``V4L2_SEL_TGT_CROP``
>>> +
>>> +     ``r.left``, ``r.top``, ``r.width``, ``r.height``
>>> +         visible rectangle; this must fit within the `V4L2_SEL_TGT_CROP_BOUNDS`
>>> +         rectangle and may be subject to adjustment to match codec and
>>> +         hardware constraints
>>> +
>>> +   * **Return fields:**
>>> +
>>> +     ``r.left``, ``r.top``, ``r.width``, ``r.height``
>>> +         visible rectangle adjusted by the encoder
>>> +
>>> +   * The following selection targets are supported on ``OUTPUT``:
>>> +
>>> +     ``V4L2_SEL_TGT_CROP_BOUNDS``
>>> +         equal to the full source frame, matching the active ``OUTPUT``
>>> +         format
>>> +
>>> +     ``V4L2_SEL_TGT_CROP_DEFAULT``
>>> +         equal to ``V4L2_SEL_TGT_CROP_BOUNDS``
>>> +
>>> +     ``V4L2_SEL_TGT_CROP``
>>> +         rectangle within the source buffer to be encoded into the
>>> +         ``CAPTURE`` stream; defaults to ``V4L2_SEL_TGT_CROP_DEFAULT``
>>
>> Since this defaults to the CROP_DEFAULT rectangle this means that if you have
>> a 16x16 macroblock size and you want to encode 1080p, you will always have to
>> explicitly set the CROP rectangle to 1920x1080, right? Since the default will
>> be 1088 instead of 1080.
> 
> Not necessarily. It depends on whether the encoder needs the source
> buffers to be aligned to macroblocks or not.
> 
>>
>> It is probably wise to explicitly mention this.
>>
> 
> Sounds reasonable to add a note.
> 
>>> +
>>> +     ``V4L2_SEL_TGT_COMPOSE_BOUNDS``
>>> +         maximum rectangle within the coded resolution, which the cropped
>>> +         source frame can be output into; if the hardware does not support
>>
>> output -> composed
>>
> 
> Ack.
> 
>>> +         composition or scaling, then this is always equal to the rectangle of
>>> +         width and height matching ``V4L2_SEL_TGT_CROP`` and located at (0, 0)
>>> +
>>> +     ``V4L2_SEL_TGT_COMPOSE_DEFAULT``
>>> +         equal to a rectangle of width and height matching
>>> +         ``V4L2_SEL_TGT_CROP`` and located at (0, 0)
>>> +
>>> +     ``V4L2_SEL_TGT_COMPOSE``
>>> +         rectangle within the coded frame, which the cropped source frame
>>> +         is to be output into; defaults to
>>
>> output -> composed
>>
> 
> Ack.
> 
>>> +         ``V4L2_SEL_TGT_COMPOSE_DEFAULT``; read-only on hardware without
>>> +         additional compose/scaling capabilities; resulting stream will
>>> +         have this rectangle encoded as the visible rectangle in its
>>> +         metadata
>>> +
>>> +   .. warning::
>>> +
>>> +      The encoder may adjust the crop/compose rectangles to the nearest
>>> +      supported ones to meet codec and hardware requirements. The client needs
>>> +      to check the adjusted rectangle returned by :c:func:`VIDIOC_S_SELECTION`.
>>> +
>>> +6. Allocate buffers for both ``OUTPUT`` and ``CAPTURE`` via
>>> +   :c:func:`VIDIOC_REQBUFS`. This may be performed in any order.
>>> +
>>> +   * **Required fields:**
>>> +
>>> +     ``count``
>>> +         requested number of buffers to allocate; greater than zero
>>> +
>>> +     ``type``
>>> +         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``OUTPUT`` or
>>> +         ``CAPTURE``
>>> +
>>> +     other fields
>>> +         follow standard semantics
>>> +
>>> +   * **Return fields:**
>>> +
>>> +     ``count``
>>> +          actual number of buffers allocated
>>> +
>>> +   .. warning::
>>> +
>>> +      The actual number of allocated buffers may differ from the ``count``
>>> +      given. The client must check the updated value of ``count`` after the
>>> +      call returns.
>>> +
>>> +   .. note::
>>> +
>>> +      To allocate more than the minimum number of buffers (for pipeline depth),
>>> +      the client may query the ``V4L2_CID_MIN_BUFFERS_FOR_OUTPUT`` or
>>> +      ``V4L2_CID_MIN_BUFFERS_FOR_CAPTURE`` control respectively, to get the
>>> +      minimum number of buffers required by the encoder/format, and pass the
>>> +      obtained value plus the number of additional buffers needed in the
>>> +      ``count`` field to :c:func:`VIDIOC_REQBUFS`.
>>
>> Does V4L2_CID_MIN_BUFFERS_FOR_CAPTURE make any sense for encoders?
>>
> 
> Indeed, I don't think it makes much sense. Let me drop it.
> 
>> V4L2_CID_MIN_BUFFERS_FOR_OUTPUT can make sense depending on GOP size etc.
>>
> 
> Yep.
> 
> 
>>> +
>>> +   Alternatively, :c:func:`VIDIOC_CREATE_BUFS` can be used to have more
>>> +   control over buffer allocation.
>>> +
>>> +   * **Required fields:**
>>> +
>>> +     ``count``
>>> +         requested number of buffers to allocate; greater than zero
>>> +
>>> +     ``type``
>>> +         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``OUTPUT``
>>> +
>>> +     other fields
>>> +         follow standard semantics
>>> +
>>> +   * **Return fields:**
>>> +
>>> +     ``count``
>>> +         adjusted to the number of allocated buffers
>>> +
>>> +7. Begin streaming on both ``OUTPUT`` and ``CAPTURE`` queues via
>>> +   :c:func:`VIDIOC_STREAMON`. This may be performed in any order. The actual
>>> +   encoding process starts when both queues start streaming.
>>> +
>>> +.. note::
>>> +
>>> +   If the client stops the ``CAPTURE`` queue during the encode process and then
>>> +   restarts it again, the encoder will begin generating a stream independent
>>> +   from the stream generated before the stop. The exact constraints depend
>>> +   on the coded format, but may include the following implications:
>>> +
>>> +   * encoded frames produced after the restart must not reference any
>>> +     frames produced before the stop, e.g. no long term references for
>>> +     H.264,
>>> +
>>> +   * any headers that must be included in a standalone stream must be
>>> +     produced again, e.g. SPS and PPS for H.264.
>>> +
>>> +Encoding
>>> +========
>>> +
>>> +This state is reached after the `Initialization` sequence finishes succesfully.
>>
>> successfully
>>
> 
> Ack.
> 
>>> +In this state, client queues and dequeues buffers to both queues via
>>
>> client -> the client
>>
> 
> Ack.
> 
>>> +:c:func:`VIDIOC_QBUF` and :c:func:`VIDIOC_DQBUF`, following the standard
>>> +semantics.
>>> +
>>> +The contents of encoded ``CAPTURE`` buffers depend on the active coded pixel
>>
>> contents ... depend -> content ... depends
>>
> 
> The plural sounds more natural to me to be honest.
> 
>>> +format and may be affected by codec-specific extended controls, as stated
>>> +in the documentation of each format.
>>> +
>>> +Both queues operate independently, following standard behavior of V4L2 buffer
>>> +queues and memory-to-memory devices. In addition, the order of encoded frames
>>> +dequeued from the ``CAPTURE`` queue may differ from the order of queuing raw
>>> +frames to the ``OUTPUT`` queue, due to properties of the selected coded format,
>>> +e.g. frame reordering.
>>> +
>>> +The client must not assume any direct relationship between ``CAPTURE`` and
>>> +``OUTPUT`` buffers and any specific timing of buffers becoming
>>> +available to dequeue. Specifically,
>>> +
>>> +* a buffer queued to ``OUTPUT`` may result in more than 1 buffer produced on
>>> +  ``CAPTURE`` (if returning an encoded frame allowed the encoder to return a
>>> +  frame that preceded it in display, but succeeded it in the decode order),
>>> +
>>> +* a buffer queued to ``OUTPUT`` may result in a buffer being produced on
>>> +  ``CAPTURE`` later into encode process, and/or after processing further
>>> +  ``OUTPUT`` buffers, or be returned out of order, e.g. if display
>>> +  reordering is used,
>>> +
>>> +* buffers may become available on the ``CAPTURE`` queue without additional
>>> +  buffers queued to ``OUTPUT`` (e.g. during drain or ``EOS``), because of the
>>> +  ``OUTPUT`` buffers queued in the past whose decoding results are only
>>> +  available at later time, due to specifics of the decoding process,
>>> +
>>> +* buffers queued to ``OUTPUT`` may not become available to dequeue instantly
>>> +  after being encoded into a corresponding ``CATPURE`` buffer, e.g. if the
>>> +  encoder needs to use the frame as a reference for encoding further frames.
>>> +
>>> +.. note::
>>> +
>>> +   To allow matching encoded ``CAPTURE`` buffers with ``OUTPUT`` buffers they
>>> +   originated from, the client can set the ``timestamp`` field of the
>>> +   :c:type:`v4l2_buffer` struct when queuing an ``OUTPUT`` buffer. The
>>> +   ``CAPTURE`` buffer(s), which resulted from encoding that ``OUTPUT`` buffer
>>> +   will have their ``timestamp`` field set to the same value when dequeued.
>>> +
>>> +   In addition to the straighforward case of one ``OUTPUT`` buffer producing
>>> +   one ``CAPTURE`` buffer, the following cases are defined:
>>> +
>>> +   * one ``OUTPUT`` buffer generates multiple ``CAPTURE`` buffers: the same
>>> +     ``OUTPUT`` timestamp will be copied to multiple ``CAPTURE`` buffers,
>>> +
>>> +   * the encoding order differs from the presentation order (i.e. the
>>> +     ``CAPTURE`` buffers are out-of-order compared to the ``OUTPUT`` buffers):
>>> +     ``CAPTURE`` timestamps will not retain the order of ``OUTPUT`` timestamps
>>> +     and thus monotonicity of the timestamps cannot be guaranteed.
>>> +
>>> +.. note::
>>> +
>>> +   To let the client distinguish between frame types (keyframes, intermediate
>>> +   frames; the exact list of types depends on the coded format), the
>>> +   ``CAPTURE`` buffers will have corresponding flag bits set in their
>>> +   :c:type:`v4l2_buffer` struct when dequeued. See the documentation of
>>> +   :c:type:`v4l2_buffer` and each coded pixel format for exact list of flags
>>> +   and their meanings.
>>
>> Is this required? (I think it should, but it isn't the case today).
>>
> 
> At least V4L2_BUF_FLAG_KEYFRAME has been always required for Chromium,
> as it's indispensable for any kind of real time streaming, e.g.
> Hangouts. (Although technically the streaming layer could parse the
> stream and pick it up itself...) I can see s5p-mfc, coda, venus and
> mtk-vcodec supporting at least this one. mtk-vcodec doesn't seem to
> support the other ones.
> 
> Should we make only the V4L2_BUF_FLAG_KEYFRAME mandatory, at least for now?
> 
>> Is the current set of buffer flags (Key/B/P frame) sufficient for the current
>> set of codecs?
>>
> 
> No, there is no way to distinguish between I and IDR frames in H.264.
> You can see more details on our issue tracker: crbug.com/868792.

OK. How about adding an IDRFRAME flag? If hardware can distinguish between
I and IDR frames, then it will set IDRFRAME in addition to KEYFRAME.

I think that is consistent with our spec. And it makes sense to make KEYFRAME
mandatory.

> 
>>> +
>>> +Encoding parameter changes
>>> +==========================
>>> +
>>> +The client is allowed to use :c:func:`VIDIOC_S_CTRL` to change encoder
>>> +parameters at any time. The availability of parameters is encoder-specific
>>> +and the client must query the encoder to find the set of available controls.
>>> +
>>> +The ability to change each parameter during encoding is encoder-specific, as per
>>> +the standard semantics of the V4L2 control interface. The client may attempt
>>> +setting a control of its interest during encoding and if the operation fails
>>
>> I'd simplify this:
>>
>> The client may attempt to set a control during encoding...
> 
> Ack.
> 
>>
>>> +with the -EBUSY error code, the ``CAPTURE`` queue needs to be stopped for the
>>> +configuration change to be allowed (following the `Drain` sequence will be
>>> +needed to avoid losing the already queued/encoded frames).
>>
>> Rephrase:
>>
>> ...to be allowed. To do this follow the `Drain` sequence to avoid losing the
>> already queued/encoded frames.
>>
> 
> How about "To do this, it may follow...", to keep client as the third
> person convention?

Ack.

> 
>>> +
>>> +The timing of parameter updates is encoder-specific, as per the standard
>>> +semantics of the V4L2 control interface. If the client needs to apply the
>>> +parameters exactly at specific frame, using the Request API should be
>>
>> Change this to a reference to the Request API section.
>>
> 
> Do you mean just adding a reference or some other changes too?

I meant: replace "Request API" by a reference (hyperlink) to the Request API
section. That way you can click on the link to get to the doc for that API.

> 
>>> +considered, if supported by the encoder.
>>> +
>>> +Drain
>>> +=====
>>> +
>>> +To ensure that all the queued ``OUTPUT`` buffers have been processed and the
>>> +related ``CAPTURE`` buffers output to the client, the client must follow the
>>
>> output -> are output
>>
>> or perhaps better (up to you): are given
>>
> 
> Ack (went with the latter and updated the decoder too).
> 
>>> +drain sequence described below. After the drain sequence ends, the client has
>>> +received all encoded frames for all ``OUTPUT`` buffers queued before the
>>> +sequence was started.
>>> +
>>> +1. Begin the drain sequence by issuing :c:func:`VIDIOC_ENCODER_CMD`.
>>> +
>>> +   * **Required fields:**
>>> +
>>> +     ``cmd``
>>> +         set to ``V4L2_ENC_CMD_STOP``
>>> +
>>> +     ``flags``
>>> +         set to 0
>>> +
>>> +     ``pts``
>>> +         set to 0
>>> +
>>> +   .. warning::
>>> +
>>> +   The sentence can be only initiated if both ``OUTPUT`` and ``CAPTURE`` queues
>>> +   are streaming. For compatibility reasons, the call to
>>> +   :c:func:`VIDIOC_ENCODER_CMD` will not fail even if any of the queues is not
>>> +   streaming, but at the same time it will not initiate the `Drain` sequence
>>> +   and so the steps described below would not be applicable.
>>> +
>>> +2. Any ``OUTPUT`` buffers queued by the client before the
>>> +   :c:func:`VIDIOC_ENCODER_CMD` was issued will be processed and encoded as
>>> +   normal. The client must continue to handle both queues independently,
>>> +   similarly to normal encode operation. This includes,
>>> +
>>> +   * queuing and dequeuing ``CAPTURE`` buffers, until a buffer marked with the
>>> +     ``V4L2_BUF_FLAG_LAST`` flag is dequeued,
>>> +
>>> +     .. warning::
>>> +
>>> +        The last buffer may be empty (with :c:type:`v4l2_buffer`
>>> +        ``bytesused`` = 0) and in such case it must be ignored by the client,
>>
>> such -> that
>>
>> Check the previous patch as well if you used the phrase 'such case' and replace
>> it with 'that case'.
>>
> 
> Ack.
> 
>>> +        as it does not contain an encoded frame.
>>> +
>>> +     .. note::
>>> +
>>> +        Any attempt to dequeue more buffers beyond the buffer marked with
>>> +        ``V4L2_BUF_FLAG_LAST`` will result in a -EPIPE error from
>>> +        :c:func:`VIDIOC_DQBUF`.
>>> +
>>> +   * dequeuing processed ``OUTPUT`` buffers, until all the buffers queued
>>> +     before the ``V4L2_ENC_CMD_STOP`` command are dequeued,
>>> +
>>> +   * dequeuing the ``V4L2_EVENT_EOS`` event, if the client subscribes to it.
>>> +
>>> +   .. note::
>>> +
>>> +      For backwards compatibility, the encoder will signal a ``V4L2_EVENT_EOS``
>>> +      event when the last the last frame has been decoded and all frames are
>>
>> the last the last -> the last
>>
> 
> Ack.
> 
>>> +      ready to be dequeued. It is a deprecated behavior and the client must not
>>
>> is a -> is
> 
> Ack.
> 
>>
>>> +      rely on it. The ``V4L2_BUF_FLAG_LAST`` buffer flag should be used
>>> +      instead.
>>
>> Question: should new codec drivers still implement the EOS event?
>>
> 
> Good question. It's a userspace compatibility issue, so if we intend a
> new codec driver to work with old userspace, it must do so. Perhaps up
> to the driver author/maintainer?

I think we need to keep it for a bit, but make sure that we document that
userspace should not rely on it and use the LAST flag instead.

> 
>>> +
>>> +3. Once all ``OUTPUT`` buffers queued before the ``V4L2_ENC_CMD_STOP`` call and
>>> +   the last ``CAPTURE`` buffer are dequeued, the encoder is stopped and it will
>>> +   accept, but not process any newly queued ``OUTPUT`` buffers until the client
>>> +   issues any of the following operations:
>>> +
>>> +   * ``V4L2_ENC_CMD_START`` - the encoder will resume operation normally,
>>
>> Perhaps mention that this does not reset the encoder? It's not immediately clear
>> when reading this.
>>
> 
> Ack.
> 
>>> +
>>> +   * a pair of :c:func:`VIDIOC_STREAMOFF` and :c:func:`VIDIOC_STREAMON` on the
>>> +     ``CAPTURE`` queue - the encoder will be reset (see the `Reset` sequence)
>>> +     and then resume encoding,
>>> +
>>> +   * a pair of :c:func:`VIDIOC_STREAMOFF` and :c:func:`VIDIOC_STREAMON` on the
>>> +     ``OUTPUT`` queue - the encoder will resume operation normally, however any
>>> +     source frames queued to the ``OUTPUT`` queue between ``V4L2_ENC_CMD_STOP``
>>> +     and :c:func:`VIDIOC_STREAMOFF` will be discarded.
>>> +
>>> +.. note::
>>> +
>>> +   Once the drain sequence is initiated, the client needs to drive it to
>>> +   completion, as described by the steps above, unless it aborts the process by
>>> +   issuing :c:func:`VIDIOC_STREAMOFF` on any of the ``OUTPUT`` or ``CAPTURE``
>>> +   queues.  The client is not allowed to issue ``V4L2_ENC_CMD_START`` or
>>> +   ``V4L2_ENC_CMD_STOP`` again while the drain sequence is in progress and they
>>> +   will fail with -EBUSY error code if attempted.
>>> +
>>> +   Although mandatory, the availability of encoder commands may be queried
>>> +   using :c:func:`VIDIOC_TRY_ENCODER_CMD`.
>>> +
>>> +Reset
>>> +=====
>>> +
>>> +The client may want to request the encoder to reinitialize the encoding, so
>>> +that the following stream data becomes independent from the stream data
>>> +generated before. Depending on the coded format, that may imply that,
>>
>> that, -> that:
>>
> 
> Ack. (And also few other places.)
> 
>>> +
>>> +* encoded frames produced after the restart must not reference any frames
>>> +  produced before the stop, e.g. no long term references for H.264,
>>> +
>>> +* any headers that must be included in a standalone stream must be produced
>>> +  again, e.g. SPS and PPS for H.264.
>>> +
>>> +This can be achieved by performing the reset sequence.
>>> +
>>> +1. Perform the `Drain` sequence to ensure all the in-flight encoding finishes
>>> +   and respective buffers are dequeued.
>>> +
>>> +2. Stop streaming on the ``CAPTURE`` queue via :c:func:`VIDIOC_STREAMOFF`. This
>>> +   will return all currently queued ``CAPTURE`` buffers to the client, without
>>> +   valid frame data.
>>> +
>>> +3. Start streaming on the ``CAPTURE`` queue via :c:func:`VIDIOC_STREAMON` and
>>> +   continue with regular encoding sequence. The encoded frames produced into
>>> +   ``CAPTURE`` buffers from now on will contain a standalone stream that can be
>>> +   decoded without the need for frames encoded before the reset sequence,
>>> +   starting at the first ``OUTPUT`` buffer queued after issuing the
>>> +   `V4L2_ENC_CMD_STOP` of the `Drain` sequence.
>>> +
>>> +This sequence may be also used to change encoding parameters for encoders
>>> +without the ability to change the parameters on the fly.
>>> +
>>> +Commit points
>>> +=============
>>> +
>>> +Setting formats and allocating buffers triggers changes in the behavior of the
>>> +encoder.
>>> +
>>> +1. Setting the format on the ``CAPTURE`` queue may change the set of formats
>>> +   supported/advertised on the ``OUTPUT`` queue. In particular, it also means
>>> +   that the ``OUTPUT`` format may be reset and the client must not rely on the
>>> +   previously set format being preserved.
>>> +
>>> +2. Enumerating formats on the ``OUTPUT`` queue always returns only formats
>>> +   supported for the current ``CAPTURE`` format.
>>> +
>>> +3. Setting the format on the ``OUTPUT`` queue does not change the list of
>>> +   formats available on the ``CAPTURE`` queue. An attempt to set the ``OUTPUT``
>>> +   format that is not supported for the currently selected ``CAPTURE`` format
>>> +   will result in the encoder adjusting the requested ``OUTPUT`` format to a
>>> +   supported one.
>>> +
>>> +4. Enumerating formats on the ``CAPTURE`` queue always returns the full set of
>>> +   supported coded formats, irrespectively of the current ``OUTPUT`` format.
>>> +
>>> +5. While buffers are allocated on the ``CAPTURE`` queue, the client must not
>>> +   change the format on the queue. Drivers will return the -EBUSY error code
>>> +   for any such format change attempt.
>>> +
>>> +To summarize, setting formats and allocation must always start with the
>>> +``CAPTURE`` queue and the ``CAPTURE`` queue is the master that governs the
>>> +set of supported formats for the ``OUTPUT`` queue.
>>> diff --git a/Documentation/media/uapi/v4l/devices.rst b/Documentation/media/uapi/v4l/devices.rst
>>> index 12d43fe711cf..1822c66c2154 100644
>>> --- a/Documentation/media/uapi/v4l/devices.rst
>>> +++ b/Documentation/media/uapi/v4l/devices.rst
>>> @@ -16,6 +16,7 @@ Interfaces
>>>      dev-osd
>>>      dev-codec
>>>      dev-decoder
>>> +    dev-encoder
>>>      dev-effect
>>>      dev-raw-vbi
>>>      dev-sliced-vbi
>>> diff --git a/Documentation/media/uapi/v4l/pixfmt-v4l2.rst b/Documentation/media/uapi/v4l/pixfmt-v4l2.rst
>>> index ca5f2270a829..085089cd9577 100644
>>> --- a/Documentation/media/uapi/v4l/pixfmt-v4l2.rst
>>> +++ b/Documentation/media/uapi/v4l/pixfmt-v4l2.rst
>>> @@ -37,6 +37,11 @@ Single-planar format structure
>>>       inside the stream, when fed to a stateful mem2mem decoder, the fields
>>>       may be zero to rely on the decoder to detect the right values. For more
>>>       details see :ref:`decoder` and format descriptions.
>>> +
>>> +     For compressed formats on the CAPTURE side of a stateful mem2mem
>>> +     encoder, the fields must be zero, since the coded size is expected to
>>> +     be calculated internally by the encoder itself, based on the OUTPUT
>>> +     side. For more details see :ref:`encoder` and format descriptions.
>>
>> The encoder document doesn't actually mention this. I think it should, though.
> 
> Indeed. I'll make it say that the fields are "ignored (always zero)".
> 
> To be honest, I wanted to define them in a way that they would be
> hardwired to the internal coded size selected by the driver, but it
> just complicated things, since one would need to set the CAPTURE
> format first, then the OUTPUT format, selection rectangles and only
> then could read back the coded resolution from the CAPTURE format. It
> would have also violated the assumption that CAPTURE format was not
> expected to be altered by changing OUTPUT format.

Right. So this now contradicts what the spec said at the beginning
(about "The encoder will derive a new ``OUTPUT`` format from the ``CAPTURE``
format being set, including resolution, colorimetry parameters, etc.").

An alternative to this is to indeed ignored width/height for the capture
format and just have ENUM_FMT report the full list of formats for the
given capture pixelformat, and require ENUM_FRAMESIZES as well to list
the resolutions that are supported by each output pixelformat.

If the OUTPUT resolution set by userspace is too large, then the CROP
rectangle should be set to a valid size or (if cropping is not supported)
the resolution should be reduced.

> 
>>
>> I'm a bit uncertain about this: the expected resolution might impact the
>> sizeimage value: i.e. encoding 640x480 requires much less memory then
>> encoding 4k video. If this is required to be 0x0, then the driver has to
>> fill in a worst-case sizeimage value. It might make more sense to say that
>> if a non-zero resolution is given, then the driver will attempt to
>> calculate a sensible sizeimage value.
> 
> The driver would still be able to determine the sizeimage by the
> internally known coded size, which it calculated based on OUTPUT
> format, selection, codec constraints, etc. It's not something for the
> userspace to provide (nor it would be able to provide).

How can it determine this if the output resolution isn't known?

It would have to set sizeimage to the worst-case size in the absence of
that information.

Regards,

	Hans
Nicolas Dufresne Jan. 24, 2019, 8:02 p.m. UTC | #10
Le mercredi 23 janvier 2019 à 19:02 +0900, Tomasz Figa a écrit :
> On Sun, Nov 18, 2018 at 10:34 AM Nicolas Dufresne <nicolas@ndufresne.ca> wrote:
> > Le samedi 17 novembre 2018 à 12:37 +0100, Hans Verkuil a écrit :
> > > > > Does V4L2_CID_MIN_BUFFERS_FOR_CAPTURE make any sense for encoders?
> > > > 
> > > > We do account for it in GStreamer (the capture/output handling is
> > > > generic), but I don't know if it's being used anywhere.
> > > 
> > > Do you use this value directly for REQBUFS, or do you use it as the minimum
> > > value but in practice use more buffers?
> > 
> > We add more buffers to that value. We assume this value is what will be
> > held by the driver, hence without adding some buffers, the driver would
> > go idle as soon as one is dequeued. We also need to allocate for the
> > importing driver.
> > 
> > In general, if we have a pipeline with Driver A sending to Driver B,
> > both driver will require a certain amount of buffers to operate. E.g.
> > with DRM display, the driver will hold on 1 buffer (the scannout
> > buffer).
> > 
> > In GStreamer, it's implemented generically, so we do:
> > 
> >   MIN_BUFFERS_FOR + remote_min + 1
> > 
> > If only MIN_BUFFERS_FOR was allocated, ignoring remote driver
> > requirement, the streaming will likely get stuck.
> 
> What happens if the driver doesn't report it?

If the driver does not report it because it does not use it (I think
CODA decoder is like that), there is no issue. If the driver do not
report but needs extra, the driver will end up growing count in
REQBUFS, so the end result will be under-allocation since the remote
requirement won't be accounted. Streaming will hang.

A good example is transcoding, you encoder will never have enough
frames to reproduce an output, because the decoder is waiting for his
frame to come back. Only solution to that would be a memcpy(), or
double allocation (redoing REQBUFS later on). The MIN_BUFFERS_FOR
announcement is the optimal way, avoiding copies and allocating twice.

> 
> Best regards,
> Tomasz
Nicolas Dufresne Jan. 24, 2019, 8:04 p.m. UTC | #11
Le mercredi 23 janvier 2019 à 12:28 +0100, Hans Verkuil a écrit :
> On 01/23/19 11:00, Tomasz Figa wrote:
> > On Sat, Nov 17, 2018 at 8:37 PM Hans Verkuil <hverkuil@xs4all.nl> wrote:
> > > On 11/17/2018 05:18 AM, Nicolas Dufresne wrote:
> > > > Le lundi 12 novembre 2018 à 14:23 +0100, Hans Verkuil a écrit :
> > > > > On 10/22/2018 04:49 PM, Tomasz Figa wrote:
> > [snip]
> > > > > > +      rely on it. The ``V4L2_BUF_FLAG_LAST`` buffer flag should be used
> > > > > > +      instead.
> > > > > 
> > > > > Question: should new codec drivers still implement the EOS event?
> > > > 
> > > > I'm been asking around, but I think here is a good place. Do we really
> > > > need the FLAG_LAST in userspace ? Userspace can also wait for the first
> > > > EPIPE return from DQBUF.
> > > 
> > > I'm interested in hearing Tomasz' opinion. This flag is used already, so there
> > > definitely is a backwards compatibility issue here.
> > > 
> > 
> > FWIW, it would add the overhead of 1 more system call, although I
> > don't think it's of our concern.
> > 
> > My personal feeling is that using error codes for signaling normal
> > conditions isn't very elegant, though.
> 
> I agree. Let's keep this flag.

Agreed, though a reminder of the initial question, "do we keep the EOS
event ?", and I think the event can be dropped.

> 
> Regards,
> 
> 	Hans
> 
> > > > > > +
> > > > > > +3. Once all ``OUTPUT`` buffers queued before the ``V4L2_ENC_CMD_STOP`` call and
> > > > > > +   the last ``CAPTURE`` buffer are dequeued, the encoder is stopped and it will
> > > > > > +   accept, but not process any newly queued ``OUTPUT`` buffers until the client
> > > > > > +   issues any of the following operations:
> > > > > > +
> > > > > > +   * ``V4L2_ENC_CMD_START`` - the encoder will resume operation normally,
> > > > > 
> > > > > Perhaps mention that this does not reset the encoder? It's not immediately clear
> > > > > when reading this.
> > > > 
> > > > Which drivers supports this ? I believe I tried with Exynos in the
> > > > past, and that didn't work. How do we know if a driver supports this or
> > > > not. Do we make it mandatory ? When it's not supported, it basically
> > > > mean userspace need to cache and resend the header in userspace, and
> > > > also need to skip to some sync point.
> > > 
> > > Once we agree on the spec, then the next step will be to add good compliance
> > > checks and update drivers that fail the tests.
> > > 
> > > To check if the driver support this ioctl you can call VIDIOC_TRY_ENCODER_CMD
> > > to see if the functionality is supported.
> > 
> > There is nothing here for the hardware to support. It's an entirely
> > driver thing, since it just needs to wait for the encoder to complete
> > all the pending frames and stop enqueuing more frames to the decoder
> > until V4L2_ENC_CMD_START is called. Any driver that can't do it must
> > be fixed, since otherwise you have no way to ensure that you got all
> > the encoded output.
> > 
> > Best regards,
> > Tomasz
> >
Nicolas Dufresne Jan. 24, 2019, 8:14 p.m. UTC | #12
Le mercredi 23 janvier 2019 à 14:04 +0100, Hans Verkuil a écrit :
> > > Does this return the same set of formats as in the 'Querying Capabilities' phase?
> > > 
> > 
> > It's actually an interesting question. At this point we wouldn't have
> > the OUTPUT resolution set yet, so that would be the same set as in the
> > initial query. If we set the resolution (with some arbitrary
> > pixelformat), it may become a subset...
> 
> But doesn't setting the capture format also set the resolution?
> 
> To quote from the text above:
> 
> "The encoder will derive a new ``OUTPUT`` format from the ``CAPTURE`` format
>  being set, including resolution, colorimetry parameters, etc."
> 
> So you set the capture format with a resolution (you know that), then
> ENUM_FMT will return the subset for that codec and resolution.
> 
> But see also the comment at the end of this email.

I'm thinking that the fact that there is no "unset" value for pixel
format creates a certain ambiguity. Maybe we could create a new pixel
format, and all CODEC driver could have that set by default ? Then we
can just fail STREAMON if that format is set.

That being said, in GStreamer, I have split each elements per CODEC,
and now only enumerate the information "per-codec". That makes me think
this "global" enumeration was just a miss-use of the API / me learning
to use it. Not having to implement this rather complex thing in the
driver would be nice. Notably, the new Amlogic driver does not have
this "Querying Capabilities" phase, and with latest GStreamer works
just fine.

Nicolas
Tomasz Figa Jan. 25, 2019, 3:29 a.m. UTC | #13
On Fri, Jan 25, 2019 at 5:04 AM Nicolas Dufresne <nicolas@ndufresne.ca> wrote:
>
> Le mercredi 23 janvier 2019 à 12:28 +0100, Hans Verkuil a écrit :
> > On 01/23/19 11:00, Tomasz Figa wrote:
> > > On Sat, Nov 17, 2018 at 8:37 PM Hans Verkuil <hverkuil@xs4all.nl> wrote:
> > > > On 11/17/2018 05:18 AM, Nicolas Dufresne wrote:
> > > > > Le lundi 12 novembre 2018 à 14:23 +0100, Hans Verkuil a écrit :
> > > > > > On 10/22/2018 04:49 PM, Tomasz Figa wrote:
> > > [snip]
> > > > > > > +      rely on it. The ``V4L2_BUF_FLAG_LAST`` buffer flag should be used
> > > > > > > +      instead.
> > > > > >
> > > > > > Question: should new codec drivers still implement the EOS event?
> > > > >
> > > > > I'm been asking around, but I think here is a good place. Do we really
> > > > > need the FLAG_LAST in userspace ? Userspace can also wait for the first
> > > > > EPIPE return from DQBUF.
> > > >
> > > > I'm interested in hearing Tomasz' opinion. This flag is used already, so there
> > > > definitely is a backwards compatibility issue here.
> > > >
> > >
> > > FWIW, it would add the overhead of 1 more system call, although I
> > > don't think it's of our concern.
> > >
> > > My personal feeling is that using error codes for signaling normal
> > > conditions isn't very elegant, though.
> >
> > I agree. Let's keep this flag.
>
> Agreed, though a reminder of the initial question, "do we keep the EOS
> event ?", and I think the event can be dropped.
>

I would happily drop it, if we know that it wouldn't break any
userspace. Chromium doesn't use it either.

Best regards,
Tomasz

> >
> > Regards,
> >
> >       Hans
> >
> > > > > > > +
> > > > > > > +3. Once all ``OUTPUT`` buffers queued before the ``V4L2_ENC_CMD_STOP`` call and
> > > > > > > +   the last ``CAPTURE`` buffer are dequeued, the encoder is stopped and it will
> > > > > > > +   accept, but not process any newly queued ``OUTPUT`` buffers until the client
> > > > > > > +   issues any of the following operations:
> > > > > > > +
> > > > > > > +   * ``V4L2_ENC_CMD_START`` - the encoder will resume operation normally,
> > > > > >
> > > > > > Perhaps mention that this does not reset the encoder? It's not immediately clear
> > > > > > when reading this.
> > > > >
> > > > > Which drivers supports this ? I believe I tried with Exynos in the
> > > > > past, and that didn't work. How do we know if a driver supports this or
> > > > > not. Do we make it mandatory ? When it's not supported, it basically
> > > > > mean userspace need to cache and resend the header in userspace, and
> > > > > also need to skip to some sync point.
> > > >
> > > > Once we agree on the spec, then the next step will be to add good compliance
> > > > checks and update drivers that fail the tests.
> > > >
> > > > To check if the driver support this ioctl you can call VIDIOC_TRY_ENCODER_CMD
> > > > to see if the functionality is supported.
> > >
> > > There is nothing here for the hardware to support. It's an entirely
> > > driver thing, since it just needs to wait for the encoder to complete
> > > all the pending frames and stop enqueuing more frames to the decoder
> > > until V4L2_ENC_CMD_START is called. Any driver that can't do it must
> > > be fixed, since otherwise you have no way to ensure that you got all
> > > the encoded output.
> > >
> > > Best regards,
> > > Tomasz
> > >
>
Tomasz Figa Jan. 25, 2019, 3:59 a.m. UTC | #14
On Fri, Jan 25, 2019 at 5:14 AM Nicolas Dufresne <nicolas@ndufresne.ca> wrote:
>
> Le mercredi 23 janvier 2019 à 14:04 +0100, Hans Verkuil a écrit :
> > > > Does this return the same set of formats as in the 'Querying Capabilities' phase?
> > > >
> > >
> > > It's actually an interesting question. At this point we wouldn't have
> > > the OUTPUT resolution set yet, so that would be the same set as in the
> > > initial query. If we set the resolution (with some arbitrary
> > > pixelformat), it may become a subset...
> >
> > But doesn't setting the capture format also set the resolution?
> >
> > To quote from the text above:
> >
> > "The encoder will derive a new ``OUTPUT`` format from the ``CAPTURE`` format
> >  being set, including resolution, colorimetry parameters, etc."
> >
> > So you set the capture format with a resolution (you know that), then
> > ENUM_FMT will return the subset for that codec and resolution.
> >
> > But see also the comment at the end of this email.
>
> I'm thinking that the fact that there is no "unset" value for pixel
> format creates a certain ambiguity. Maybe we could create a new pixel
> format, and all CODEC driver could have that set by default ? Then we
> can just fail STREAMON if that format is set.

The state on the CAPTURE queue is actually not "unset". The queue is
simply not ready (yet) and any operations on it will error out.

Once the application sets the coded resolution on the OUTPUT queue or
the decoder parses the stream information, the CAPTURE queue becomes
ready and one can do the ioctls on it.

>
> That being said, in GStreamer, I have split each elements per CODEC,
> and now only enumerate the information "per-codec". That makes me think
> this "global" enumeration was just a miss-use of the API / me learning
> to use it. Not having to implement this rather complex thing in the
> driver would be nice. Notably, the new Amlogic driver does not have
> this "Querying Capabilities" phase, and with latest GStreamer works
> just fine.

What do you mean by "doesn't have"? Does it lack an implementation of
VIDIOC_ENUM_FMT and VIDIOC_ENUM_FRAMESIZES?

Best regards,
Tomasz
Nicolas Dufresne Jan. 30, 2019, 3:06 p.m. UTC | #15
Le vendredi 25 janvier 2019 à 12:59 +0900, Tomasz Figa a écrit :
> On Fri, Jan 25, 2019 at 5:14 AM Nicolas Dufresne <nicolas@ndufresne.ca> wrote:
> > Le mercredi 23 janvier 2019 à 14:04 +0100, Hans Verkuil a écrit :
> > > > > Does this return the same set of formats as in the 'Querying Capabilities' phase?
> > > > > 
> > > > 
> > > > It's actually an interesting question. At this point we wouldn't have
> > > > the OUTPUT resolution set yet, so that would be the same set as in the
> > > > initial query. If we set the resolution (with some arbitrary
> > > > pixelformat), it may become a subset...
> > > 
> > > But doesn't setting the capture format also set the resolution?
> > > 
> > > To quote from the text above:
> > > 
> > > "The encoder will derive a new ``OUTPUT`` format from the ``CAPTURE`` format
> > >  being set, including resolution, colorimetry parameters, etc."
> > > 
> > > So you set the capture format with a resolution (you know that), then
> > > ENUM_FMT will return the subset for that codec and resolution.
> > > 
> > > But see also the comment at the end of this email.
> > 
> > I'm thinking that the fact that there is no "unset" value for pixel
> > format creates a certain ambiguity. Maybe we could create a new pixel
> > format, and all CODEC driver could have that set by default ? Then we
> > can just fail STREAMON if that format is set.
> 
> The state on the CAPTURE queue is actually not "unset". The queue is
> simply not ready (yet) and any operations on it will error out.

My point was that it's just awkward to have this "not ready" state, in
which you cannot go back. And in which the enum-format will ignore the
format configured on the other side.

What I wanted to say is that this special case is not really needed.

> 
> Once the application sets the coded resolution on the OUTPUT queue or
> the decoder parses the stream information, the CAPTURE queue becomes
> ready and one can do the ioctls on it.
> 
> > That being said, in GStreamer, I have split each elements per CODEC,
> > and now only enumerate the information "per-codec". That makes me think
> > this "global" enumeration was just a miss-use of the API / me learning
> > to use it. Not having to implement this rather complex thing in the
> > driver would be nice. Notably, the new Amlogic driver does not have
> > this "Querying Capabilities" phase, and with latest GStreamer works
> > just fine.
> 
> What do you mean by "doesn't have"? Does it lack an implementation of
> VIDIOC_ENUM_FMT and VIDIOC_ENUM_FRAMESIZES?

What it does is that it sets a default value for the codec format, so
if you just open the device and do enum_fmt/framesizes, you get that is
possible for the default codec that was selected. And I thin it's
entirely correct, doing ENUM_FMT(capture) without doing an
S_FMT(output) can easily be documented as undefined behaviour.

For proper enumeration would be:

for formats on OUTPUT device:
  S_FMT(OUTPUT):
  for formats on CAPTURE device:
    ...

(the pseudo for look represent an enum operation)

> 
> Best regards,
> Tomasz
Tomasz Figa Feb. 6, 2019, 5:49 a.m. UTC | #16
On Thu, Jan 31, 2019 at 12:06 AM Nicolas Dufresne <nicolas@ndufresne.ca> wrote:
>
> Le vendredi 25 janvier 2019 à 12:59 +0900, Tomasz Figa a écrit :
> > On Fri, Jan 25, 2019 at 5:14 AM Nicolas Dufresne <nicolas@ndufresne.ca> wrote:
> > > Le mercredi 23 janvier 2019 à 14:04 +0100, Hans Verkuil a écrit :
> > > > > > Does this return the same set of formats as in the 'Querying Capabilities' phase?
> > > > > >
> > > > >
> > > > > It's actually an interesting question. At this point we wouldn't have
> > > > > the OUTPUT resolution set yet, so that would be the same set as in the
> > > > > initial query. If we set the resolution (with some arbitrary
> > > > > pixelformat), it may become a subset...
> > > >
> > > > But doesn't setting the capture format also set the resolution?
> > > >
> > > > To quote from the text above:
> > > >
> > > > "The encoder will derive a new ``OUTPUT`` format from the ``CAPTURE`` format
> > > >  being set, including resolution, colorimetry parameters, etc."
> > > >
> > > > So you set the capture format with a resolution (you know that), then
> > > > ENUM_FMT will return the subset for that codec and resolution.
> > > >
> > > > But see also the comment at the end of this email.
> > >
> > > I'm thinking that the fact that there is no "unset" value for pixel
> > > format creates a certain ambiguity. Maybe we could create a new pixel
> > > format, and all CODEC driver could have that set by default ? Then we
> > > can just fail STREAMON if that format is set.
> >
> > The state on the CAPTURE queue is actually not "unset". The queue is
> > simply not ready (yet) and any operations on it will error out.
>
> My point was that it's just awkward to have this "not ready" state, in
> which you cannot go back. And in which the enum-format will ignore the
> format configured on the other side.
>
> What I wanted to say is that this special case is not really needed.
>

Yeah, I think we may actually end up going in that direction, as you
probably noticed in the discussion over the "venus: dec: make decoder
compliant with stateful codec API" patch [1].

[1] https://patchwork.kernel.org/patch/10768539/#22462703

> >
> > Once the application sets the coded resolution on the OUTPUT queue or
> > the decoder parses the stream information, the CAPTURE queue becomes
> > ready and one can do the ioctls on it.
> >
> > > That being said, in GStreamer, I have split each elements per CODEC,
> > > and now only enumerate the information "per-codec". That makes me think
> > > this "global" enumeration was just a miss-use of the API / me learning
> > > to use it. Not having to implement this rather complex thing in the
> > > driver would be nice. Notably, the new Amlogic driver does not have
> > > this "Querying Capabilities" phase, and with latest GStreamer works
> > > just fine.
> >
> > What do you mean by "doesn't have"? Does it lack an implementation of
> > VIDIOC_ENUM_FMT and VIDIOC_ENUM_FRAMESIZES?
>
> What it does is that it sets a default value for the codec format, so
> if you just open the device and do enum_fmt/framesizes, you get that is
> possible for the default codec that was selected. And I thin it's
> entirely correct, doing ENUM_FMT(capture) without doing an
> S_FMT(output) can easily be documented as undefined behaviour.

Okay.

>
> For proper enumeration would be:
>
> for formats on OUTPUT device:
>   S_FMT(OUTPUT):
>   for formats on CAPTURE device:
>     ...
>
> (the pseudo for look represent an enum operation)

And that's how it's defined in v3. There is no default state without
any codec selected.

Best regards,
Tomasz

Patch
diff mbox series

diff --git a/Documentation/media/uapi/v4l/dev-encoder.rst b/Documentation/media/uapi/v4l/dev-encoder.rst
new file mode 100644
index 000000000000..41139e5e48eb
--- /dev/null
+++ b/Documentation/media/uapi/v4l/dev-encoder.rst
@@ -0,0 +1,579 @@ 
+.. -*- coding: utf-8; mode: rst -*-
+
+.. _encoder:
+
+*************************************************
+Memory-to-memory Stateful Video Encoder Interface
+*************************************************
+
+A stateful video encoder takes raw video frames in display order and encodes
+them into a bitstream. It generates complete chunks of the bitstream, including
+all metadata, headers, etc. The resulting bitstream does not require any
+further post-processing by the client.
+
+Performing software stream processing, header generation etc. in the driver
+in order to support this interface is strongly discouraged. In case such
+operations are needed, use of the Stateless Video Encoder Interface (in
+development) is strongly advised.
+
+Conventions and notation used in this document
+==============================================
+
+1. The general V4L2 API rules apply if not specified in this document
+   otherwise.
+
+2. The meaning of words "must", "may", "should", etc. is as per RFC
+   2119.
+
+3. All steps not marked "optional" are required.
+
+4. :c:func:`VIDIOC_G_EXT_CTRLS`, :c:func:`VIDIOC_S_EXT_CTRLS` may be used
+   interchangeably with :c:func:`VIDIOC_G_CTRL`, :c:func:`VIDIOC_S_CTRL`,
+   unless specified otherwise.
+
+5. Single-plane API (see spec) and applicable structures may be used
+   interchangeably with Multi-plane API, unless specified otherwise,
+   depending on encoder capabilities and following the general V4L2
+   guidelines.
+
+6. i = [a..b]: sequence of integers from a to b, inclusive, i.e. i =
+   [0..2]: i = 0, 1, 2.
+
+7. Given an ``OUTPUT`` buffer A, A' represents a buffer on the ``CAPTURE``
+   queue containing data (encoded frame/stream) that resulted from processing
+   buffer A.
+
+Glossary
+========
+
+Refer to :ref:`decoder-glossary`.
+
+State machine
+=============
+
+.. kernel-render:: DOT
+   :alt: DOT digraph of encoder state machine
+   :caption: Encoder state machine
+
+   digraph encoder_state_machine {
+       node [shape = doublecircle, label="Encoding"] Encoding;
+
+       node [shape = circle, label="Initialization"] Initialization;
+       node [shape = circle, label="Stopped"] Stopped;
+       node [shape = circle, label="Drain"] Drain;
+       node [shape = circle, label="Reset"] Reset;
+
+       node [shape = point]; qi
+       qi -> Initialization [ label = "open()" ];
+
+       Initialization -> Encoding [ label = "Both queues streaming" ];
+
+       Encoding -> Drain [ label = "V4L2_DEC_CMD_STOP" ];
+       Encoding -> Reset [ label = "VIDIOC_STREAMOFF(CAPTURE)" ];
+       Encoding -> Stopped [ label = "VIDIOC_STREAMOFF(OUTPUT)" ];
+       Encoding -> Encoding;
+
+       Drain -> Stopped [ label = "All CAPTURE\nbuffers dequeued\nor\nVIDIOC_STREAMOFF(CAPTURE)" ];
+       Drain -> Reset [ label = "VIDIOC_STREAMOFF(CAPTURE)" ];
+
+       Reset -> Encoding [ label = "VIDIOC_STREAMON(CAPTURE)" ];
+       Reset -> Initialization [ label = "VIDIOC_REQBUFS(OUTPUT, 0)" ];
+
+       Stopped -> Encoding [ label = "V4L2_DEC_CMD_START\nor\nVIDIOC_STREAMON(OUTPUT)" ];
+       Stopped -> Reset [ label = "VIDIOC_STREAMOFF(CAPTURE)" ];
+   }
+
+Querying capabilities
+=====================
+
+1. To enumerate the set of coded formats supported by the encoder, the
+   client may call :c:func:`VIDIOC_ENUM_FMT` on ``CAPTURE``.
+
+   * The full set of supported formats will be returned, regardless of the
+     format set on ``OUTPUT``.
+
+2. To enumerate the set of supported raw formats, the client may call
+   :c:func:`VIDIOC_ENUM_FMT` on ``OUTPUT``.
+
+   * Only the formats supported for the format currently active on ``CAPTURE``
+     will be returned.
+
+   * In order to enumerate raw formats supported by a given coded format,
+     the client must first set that coded format on ``CAPTURE`` and then
+     enumerate the formats on ``OUTPUT``.
+
+3. The client may use :c:func:`VIDIOC_ENUM_FRAMESIZES` to detect supported
+   resolutions for a given format, passing desired pixel format in
+   :c:type:`v4l2_frmsizeenum` ``pixel_format``.
+
+   * Values returned by :c:func:`VIDIOC_ENUM_FRAMESIZES` for a coded pixel
+     format will include all possible coded resolutions supported by the
+     encoder for given coded pixel format.
+
+   * Values returned by :c:func:`VIDIOC_ENUM_FRAMESIZES` for a raw pixel format
+     will include all possible frame buffer resolutions supported by the
+     encoder for given raw pixel format and coded format currently set on
+     ``CAPTURE``.
+
+4. Supported profiles and levels for given format, if applicable, may be
+   queried using their respective controls via :c:func:`VIDIOC_QUERYCTRL`.
+
+5. Any additional encoder capabilities may be discovered by querying
+   their respective controls.
+
+Initialization
+==============
+
+1. **Optional.** Enumerate supported formats and resolutions. See
+   `Querying capabilities` above.
+
+2. Set a coded format on the ``CAPTURE`` queue via :c:func:`VIDIOC_S_FMT`
+
+   * **Required fields:**
+
+     ``type``
+         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``CAPTURE``
+
+     ``pixelformat``
+         the coded format to be produced
+
+     ``sizeimage``
+         desired size of ``CAPTURE`` buffers; the encoder may adjust it to
+         match hardware requirements
+
+     other fields
+         follow standard semantics
+
+   * **Return fields:**
+
+     ``sizeimage``
+         adjusted size of ``CAPTURE`` buffers
+
+   .. warning::
+
+      Changing the ``CAPTURE`` format may change the currently set ``OUTPUT``
+      format. The encoder will derive a new ``OUTPUT`` format from the
+      ``CAPTURE`` format being set, including resolution, colorimetry
+      parameters, etc. If the client needs a specific ``OUTPUT`` format, it
+      must adjust it afterwards.
+
+3. **Optional.** Enumerate supported ``OUTPUT`` formats (raw formats for
+   source) for the selected coded format via :c:func:`VIDIOC_ENUM_FMT`.
+
+   * **Required fields:**
+
+     ``type``
+         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``OUTPUT``
+
+     other fields
+         follow standard semantics
+
+   * **Return fields:**
+
+     ``pixelformat``
+         raw format supported for the coded format currently selected on
+         the ``OUTPUT`` queue.
+
+     other fields
+         follow standard semantics
+
+4. Set the raw source format on the ``OUTPUT`` queue via
+   :c:func:`VIDIOC_S_FMT`.
+
+   * **Required fields:**
+
+     ``type``
+         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``OUTPUT``
+
+     ``pixelformat``
+         raw format of the source
+
+     ``width``, ``height``
+         source resolution
+
+     other fields
+         follow standard semantics
+
+   * **Return fields:**
+
+     ``width``, ``height``
+         may be adjusted by encoder to match alignment requirements, as
+         required by the currently selected formats
+
+     other fields
+         follow standard semantics
+
+   * Setting the source resolution will reset the selection rectangles to their
+     default values, based on the new resolution, as described in the step 5
+     below.
+
+5. **Optional.** Set the visible resolution for the stream metadata via
+   :c:func:`VIDIOC_S_SELECTION` on the ``OUTPUT`` queue.
+
+   * **Required fields:**
+
+     ``type``
+         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``OUTPUT``
+
+     ``target``
+         set to ``V4L2_SEL_TGT_CROP``
+
+     ``r.left``, ``r.top``, ``r.width``, ``r.height``
+         visible rectangle; this must fit within the `V4L2_SEL_TGT_CROP_BOUNDS`
+         rectangle and may be subject to adjustment to match codec and
+         hardware constraints
+
+   * **Return fields:**
+
+     ``r.left``, ``r.top``, ``r.width``, ``r.height``
+         visible rectangle adjusted by the encoder
+
+   * The following selection targets are supported on ``OUTPUT``:
+
+     ``V4L2_SEL_TGT_CROP_BOUNDS``
+         equal to the full source frame, matching the active ``OUTPUT``
+         format
+
+     ``V4L2_SEL_TGT_CROP_DEFAULT``
+         equal to ``V4L2_SEL_TGT_CROP_BOUNDS``
+
+     ``V4L2_SEL_TGT_CROP``
+         rectangle within the source buffer to be encoded into the
+         ``CAPTURE`` stream; defaults to ``V4L2_SEL_TGT_CROP_DEFAULT``
+
+     ``V4L2_SEL_TGT_COMPOSE_BOUNDS``
+         maximum rectangle within the coded resolution, which the cropped
+         source frame can be output into; if the hardware does not support
+         composition or scaling, then this is always equal to the rectangle of
+         width and height matching ``V4L2_SEL_TGT_CROP`` and located at (0, 0)
+
+     ``V4L2_SEL_TGT_COMPOSE_DEFAULT``
+         equal to a rectangle of width and height matching
+         ``V4L2_SEL_TGT_CROP`` and located at (0, 0)
+
+     ``V4L2_SEL_TGT_COMPOSE``
+         rectangle within the coded frame, which the cropped source frame
+         is to be output into; defaults to
+         ``V4L2_SEL_TGT_COMPOSE_DEFAULT``; read-only on hardware without
+         additional compose/scaling capabilities; resulting stream will
+         have this rectangle encoded as the visible rectangle in its
+         metadata
+
+   .. warning::
+
+      The encoder may adjust the crop/compose rectangles to the nearest
+      supported ones to meet codec and hardware requirements. The client needs
+      to check the adjusted rectangle returned by :c:func:`VIDIOC_S_SELECTION`.
+
+6. Allocate buffers for both ``OUTPUT`` and ``CAPTURE`` via
+   :c:func:`VIDIOC_REQBUFS`. This may be performed in any order.
+
+   * **Required fields:**
+
+     ``count``
+         requested number of buffers to allocate; greater than zero
+
+     ``type``
+         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``OUTPUT`` or
+         ``CAPTURE``
+
+     other fields
+         follow standard semantics
+
+   * **Return fields:**
+
+     ``count``
+          actual number of buffers allocated
+
+   .. warning::
+
+      The actual number of allocated buffers may differ from the ``count``
+      given. The client must check the updated value of ``count`` after the
+      call returns.
+
+   .. note::
+
+      To allocate more than the minimum number of buffers (for pipeline depth),
+      the client may query the ``V4L2_CID_MIN_BUFFERS_FOR_OUTPUT`` or
+      ``V4L2_CID_MIN_BUFFERS_FOR_CAPTURE`` control respectively, to get the
+      minimum number of buffers required by the encoder/format, and pass the
+      obtained value plus the number of additional buffers needed in the
+      ``count`` field to :c:func:`VIDIOC_REQBUFS`.
+
+   Alternatively, :c:func:`VIDIOC_CREATE_BUFS` can be used to have more
+   control over buffer allocation.
+
+   * **Required fields:**
+
+     ``count``
+         requested number of buffers to allocate; greater than zero
+
+     ``type``
+         a ``V4L2_BUF_TYPE_*`` enum appropriate for ``OUTPUT``
+
+     other fields
+         follow standard semantics
+
+   * **Return fields:**
+
+     ``count``
+         adjusted to the number of allocated buffers
+
+7. Begin streaming on both ``OUTPUT`` and ``CAPTURE`` queues via
+   :c:func:`VIDIOC_STREAMON`. This may be performed in any order. The actual
+   encoding process starts when both queues start streaming.
+
+.. note::
+
+   If the client stops the ``CAPTURE`` queue during the encode process and then
+   restarts it again, the encoder will begin generating a stream independent
+   from the stream generated before the stop. The exact constraints depend
+   on the coded format, but may include the following implications:
+
+   * encoded frames produced after the restart must not reference any
+     frames produced before the stop, e.g. no long term references for
+     H.264,
+
+   * any headers that must be included in a standalone stream must be
+     produced again, e.g. SPS and PPS for H.264.
+
+Encoding
+========
+
+This state is reached after the `Initialization` sequence finishes succesfully.
+In this state, client queues and dequeues buffers to both queues via
+:c:func:`VIDIOC_QBUF` and :c:func:`VIDIOC_DQBUF`, following the standard
+semantics.
+
+The contents of encoded ``CAPTURE`` buffers depend on the active coded pixel
+format and may be affected by codec-specific extended controls, as stated
+in the documentation of each format.
+
+Both queues operate independently, following standard behavior of V4L2 buffer
+queues and memory-to-memory devices. In addition, the order of encoded frames
+dequeued from the ``CAPTURE`` queue may differ from the order of queuing raw
+frames to the ``OUTPUT`` queue, due to properties of the selected coded format,
+e.g. frame reordering.
+
+The client must not assume any direct relationship between ``CAPTURE`` and
+``OUTPUT`` buffers and any specific timing of buffers becoming
+available to dequeue. Specifically,
+
+* a buffer queued to ``OUTPUT`` may result in more than 1 buffer produced on
+  ``CAPTURE`` (if returning an encoded frame allowed the encoder to return a
+  frame that preceded it in display, but succeeded it in the decode order),
+
+* a buffer queued to ``OUTPUT`` may result in a buffer being produced on
+  ``CAPTURE`` later into encode process, and/or after processing further
+  ``OUTPUT`` buffers, or be returned out of order, e.g. if display
+  reordering is used,
+
+* buffers may become available on the ``CAPTURE`` queue without additional
+  buffers queued to ``OUTPUT`` (e.g. during drain or ``EOS``), because of the
+  ``OUTPUT`` buffers queued in the past whose decoding results are only
+  available at later time, due to specifics of the decoding process,
+
+* buffers queued to ``OUTPUT`` may not become available to dequeue instantly
+  after being encoded into a corresponding ``CATPURE`` buffer, e.g. if the
+  encoder needs to use the frame as a reference for encoding further frames.
+
+.. note::
+
+   To allow matching encoded ``CAPTURE`` buffers with ``OUTPUT`` buffers they
+   originated from, the client can set the ``timestamp`` field of the
+   :c:type:`v4l2_buffer` struct when queuing an ``OUTPUT`` buffer. The
+   ``CAPTURE`` buffer(s), which resulted from encoding that ``OUTPUT`` buffer
+   will have their ``timestamp`` field set to the same value when dequeued.
+
+   In addition to the straighforward case of one ``OUTPUT`` buffer producing
+   one ``CAPTURE`` buffer, the following cases are defined:
+
+   * one ``OUTPUT`` buffer generates multiple ``CAPTURE`` buffers: the same
+     ``OUTPUT`` timestamp will be copied to multiple ``CAPTURE`` buffers,
+
+   * the encoding order differs from the presentation order (i.e. the
+     ``CAPTURE`` buffers are out-of-order compared to the ``OUTPUT`` buffers):
+     ``CAPTURE`` timestamps will not retain the order of ``OUTPUT`` timestamps
+     and thus monotonicity of the timestamps cannot be guaranteed.
+
+.. note::
+
+   To let the client distinguish between frame types (keyframes, intermediate
+   frames; the exact list of types depends on the coded format), the
+   ``CAPTURE`` buffers will have corresponding flag bits set in their
+   :c:type:`v4l2_buffer` struct when dequeued. See the documentation of
+   :c:type:`v4l2_buffer` and each coded pixel format for exact list of flags
+   and their meanings.
+
+Encoding parameter changes
+==========================
+
+The client is allowed to use :c:func:`VIDIOC_S_CTRL` to change encoder
+parameters at any time. The availability of parameters is encoder-specific
+and the client must query the encoder to find the set of available controls.
+
+The ability to change each parameter during encoding is encoder-specific, as per
+the standard semantics of the V4L2 control interface. The client may attempt
+setting a control of its interest during encoding and if the operation fails
+with the -EBUSY error code, the ``CAPTURE`` queue needs to be stopped for the
+configuration change to be allowed (following the `Drain` sequence will be
+needed to avoid losing the already queued/encoded frames).
+
+The timing of parameter updates is encoder-specific, as per the standard
+semantics of the V4L2 control interface. If the client needs to apply the
+parameters exactly at specific frame, using the Request API should be
+considered, if supported by the encoder.
+
+Drain
+=====
+
+To ensure that all the queued ``OUTPUT`` buffers have been processed and the
+related ``CAPTURE`` buffers output to the client, the client must follow the
+drain sequence described below. After the drain sequence ends, the client has
+received all encoded frames for all ``OUTPUT`` buffers queued before the
+sequence was started.
+
+1. Begin the drain sequence by issuing :c:func:`VIDIOC_ENCODER_CMD`.
+
+   * **Required fields:**
+
+     ``cmd``
+         set to ``V4L2_ENC_CMD_STOP``
+
+     ``flags``
+         set to 0
+
+     ``pts``
+         set to 0
+
+   .. warning::
+
+   The sentence can be only initiated if both ``OUTPUT`` and ``CAPTURE`` queues
+   are streaming. For compatibility reasons, the call to
+   :c:func:`VIDIOC_ENCODER_CMD` will not fail even if any of the queues is not
+   streaming, but at the same time it will not initiate the `Drain` sequence
+   and so the steps described below would not be applicable.
+
+2. Any ``OUTPUT`` buffers queued by the client before the
+   :c:func:`VIDIOC_ENCODER_CMD` was issued will be processed and encoded as
+   normal. The client must continue to handle both queues independently,
+   similarly to normal encode operation. This includes,
+
+   * queuing and dequeuing ``CAPTURE`` buffers, until a buffer marked with the
+     ``V4L2_BUF_FLAG_LAST`` flag is dequeued,
+
+     .. warning::
+
+        The last buffer may be empty (with :c:type:`v4l2_buffer`
+        ``bytesused`` = 0) and in such case it must be ignored by the client,
+        as it does not contain an encoded frame.
+
+     .. note::
+
+        Any attempt to dequeue more buffers beyond the buffer marked with
+        ``V4L2_BUF_FLAG_LAST`` will result in a -EPIPE error from
+        :c:func:`VIDIOC_DQBUF`.
+
+   * dequeuing processed ``OUTPUT`` buffers, until all the buffers queued
+     before the ``V4L2_ENC_CMD_STOP`` command are dequeued,
+
+   * dequeuing the ``V4L2_EVENT_EOS`` event, if the client subscribes to it.
+
+   .. note::
+
+      For backwards compatibility, the encoder will signal a ``V4L2_EVENT_EOS``
+      event when the last the last frame has been decoded and all frames are
+      ready to be dequeued. It is a deprecated behavior and the client must not
+      rely on it. The ``V4L2_BUF_FLAG_LAST`` buffer flag should be used
+      instead.
+
+3. Once all ``OUTPUT`` buffers queued before the ``V4L2_ENC_CMD_STOP`` call and
+   the last ``CAPTURE`` buffer are dequeued, the encoder is stopped and it will
+   accept, but not process any newly queued ``OUTPUT`` buffers until the client
+   issues any of the following operations:
+
+   * ``V4L2_ENC_CMD_START`` - the encoder will resume operation normally,
+
+   * a pair of :c:func:`VIDIOC_STREAMOFF` and :c:func:`VIDIOC_STREAMON` on the
+     ``CAPTURE`` queue - the encoder will be reset (see the `Reset` sequence)
+     and then resume encoding,
+
+   * a pair of :c:func:`VIDIOC_STREAMOFF` and :c:func:`VIDIOC_STREAMON` on the
+     ``OUTPUT`` queue - the encoder will resume operation normally, however any
+     source frames queued to the ``OUTPUT`` queue between ``V4L2_ENC_CMD_STOP``
+     and :c:func:`VIDIOC_STREAMOFF` will be discarded.
+
+.. note::
+
+   Once the drain sequence is initiated, the client needs to drive it to
+   completion, as described by the steps above, unless it aborts the process by
+   issuing :c:func:`VIDIOC_STREAMOFF` on any of the ``OUTPUT`` or ``CAPTURE``
+   queues.  The client is not allowed to issue ``V4L2_ENC_CMD_START`` or
+   ``V4L2_ENC_CMD_STOP`` again while the drain sequence is in progress and they
+   will fail with -EBUSY error code if attempted.
+
+   Although mandatory, the availability of encoder commands may be queried
+   using :c:func:`VIDIOC_TRY_ENCODER_CMD`.
+
+Reset
+=====
+
+The client may want to request the encoder to reinitialize the encoding, so
+that the following stream data becomes independent from the stream data
+generated before. Depending on the coded format, that may imply that,
+
+* encoded frames produced after the restart must not reference any frames
+  produced before the stop, e.g. no long term references for H.264,
+
+* any headers that must be included in a standalone stream must be produced
+  again, e.g. SPS and PPS for H.264.
+
+This can be achieved by performing the reset sequence.
+
+1. Perform the `Drain` sequence to ensure all the in-flight encoding finishes
+   and respective buffers are dequeued.
+
+2. Stop streaming on the ``CAPTURE`` queue via :c:func:`VIDIOC_STREAMOFF`. This
+   will return all currently queued ``CAPTURE`` buffers to the client, without
+   valid frame data.
+
+3. Start streaming on the ``CAPTURE`` queue via :c:func:`VIDIOC_STREAMON` and
+   continue with regular encoding sequence. The encoded frames produced into
+   ``CAPTURE`` buffers from now on will contain a standalone stream that can be
+   decoded without the need for frames encoded before the reset sequence,
+   starting at the first ``OUTPUT`` buffer queued after issuing the
+   `V4L2_ENC_CMD_STOP` of the `Drain` sequence.
+
+This sequence may be also used to change encoding parameters for encoders
+without the ability to change the parameters on the fly.
+
+Commit points
+=============
+
+Setting formats and allocating buffers triggers changes in the behavior of the
+encoder.
+
+1. Setting the format on the ``CAPTURE`` queue may change the set of formats
+   supported/advertised on the ``OUTPUT`` queue. In particular, it also means
+   that the ``OUTPUT`` format may be reset and the client must not rely on the
+   previously set format being preserved.
+
+2. Enumerating formats on the ``OUTPUT`` queue always returns only formats
+   supported for the current ``CAPTURE`` format.
+
+3. Setting the format on the ``OUTPUT`` queue does not change the list of
+   formats available on the ``CAPTURE`` queue. An attempt to set the ``OUTPUT``
+   format that is not supported for the currently selected ``CAPTURE`` format
+   will result in the encoder adjusting the requested ``OUTPUT`` format to a
+   supported one.
+
+4. Enumerating formats on the ``CAPTURE`` queue always returns the full set of
+   supported coded formats, irrespectively of the current ``OUTPUT`` format.
+
+5. While buffers are allocated on the ``CAPTURE`` queue, the client must not
+   change the format on the queue. Drivers will return the -EBUSY error code
+   for any such format change attempt.
+
+To summarize, setting formats and allocation must always start with the
+``CAPTURE`` queue and the ``CAPTURE`` queue is the master that governs the
+set of supported formats for the ``OUTPUT`` queue.
diff --git a/Documentation/media/uapi/v4l/devices.rst b/Documentation/media/uapi/v4l/devices.rst
index 12d43fe711cf..1822c66c2154 100644
--- a/Documentation/media/uapi/v4l/devices.rst
+++ b/Documentation/media/uapi/v4l/devices.rst
@@ -16,6 +16,7 @@  Interfaces
     dev-osd
     dev-codec
     dev-decoder
+    dev-encoder
     dev-effect
     dev-raw-vbi
     dev-sliced-vbi
diff --git a/Documentation/media/uapi/v4l/pixfmt-v4l2.rst b/Documentation/media/uapi/v4l/pixfmt-v4l2.rst
index ca5f2270a829..085089cd9577 100644
--- a/Documentation/media/uapi/v4l/pixfmt-v4l2.rst
+++ b/Documentation/media/uapi/v4l/pixfmt-v4l2.rst
@@ -37,6 +37,11 @@  Single-planar format structure
 	inside the stream, when fed to a stateful mem2mem decoder, the fields
 	may be zero to rely on the decoder to detect the right values. For more
 	details see :ref:`decoder` and format descriptions.
+
+	For compressed formats on the CAPTURE side of a stateful mem2mem
+	encoder, the fields must be zero, since the coded size is expected to
+	be calculated internally by the encoder itself, based on the OUTPUT
+	side. For more details see :ref:`encoder` and format descriptions.
     * - __u32
       - ``pixelformat``
       - The pixel format or type of compression, set by the application.
diff --git a/Documentation/media/uapi/v4l/v4l2.rst b/Documentation/media/uapi/v4l/v4l2.rst
index 65dc096199ad..2ef6693b9499 100644
--- a/Documentation/media/uapi/v4l/v4l2.rst
+++ b/Documentation/media/uapi/v4l/v4l2.rst
@@ -56,6 +56,7 @@  Authors, in alphabetical order:
 - Figa, Tomasz <tfiga@chromium.org>
 
   - Documented the memory-to-memory decoder interface.
+  - Documented the memory-to-memory encoder interface.
 
 - H Schimek, Michael <mschimek@gmx.at>
 
@@ -68,6 +69,7 @@  Authors, in alphabetical order:
 - Osciak, Pawel <posciak@chromium.org>
 
   - Documented the memory-to-memory decoder interface.
+  - Documented the memory-to-memory encoder interface.
 
 - Osciak, Pawel <pawel@osciak.com>
 
diff --git a/Documentation/media/uapi/v4l/vidioc-encoder-cmd.rst b/Documentation/media/uapi/v4l/vidioc-encoder-cmd.rst
index 5ae8c933b1b9..d571c53e761a 100644
--- a/Documentation/media/uapi/v4l/vidioc-encoder-cmd.rst
+++ b/Documentation/media/uapi/v4l/vidioc-encoder-cmd.rst
@@ -50,19 +50,23 @@  currently only used by the STOP command and contains one bit: If the
 until the end of the current *Group Of Pictures*, otherwise it will stop
 immediately.
 
-A :ref:`read() <func-read>` or :ref:`VIDIOC_STREAMON <VIDIOC_STREAMON>`
-call sends an implicit START command to the encoder if it has not been
-started yet. After a STOP command, :ref:`read() <func-read>` calls will read
+After a STOP command, :ref:`read() <func-read>` calls will read
 the remaining data buffered by the driver. When the buffer is empty,
 :ref:`read() <func-read>` will return zero and the next :ref:`read() <func-read>`
 call will restart the encoder.
 
+A :ref:`read() <func-read>` or :ref:`VIDIOC_STREAMON <VIDIOC_STREAMON>`
+call sends an implicit START command to the encoder if it has not been
+started yet. Applies to both queues of mem2mem encoders.
+
 A :ref:`close() <func-close>` or :ref:`VIDIOC_STREAMOFF <VIDIOC_STREAMON>`
 call of a streaming file descriptor sends an implicit immediate STOP to
-the encoder, and all buffered data is discarded.
+the encoder, and all buffered data is discarded. Applies to both queues of
+mem2mem encoders.
 
 These ioctls are optional, not all drivers may support them. They were
-introduced in Linux 2.6.21.
+introduced in Linux 2.6.21. They are, however, mandatory for stateful mem2mem
+encoders (as further documented in :ref:`encoder`).
 
 
 .. tabularcolumns:: |p{4.4cm}|p{4.4cm}|p{8.7cm}|
@@ -107,16 +111,20 @@  introduced in Linux 2.6.21.
       - Stop the encoder. When the ``V4L2_ENC_CMD_STOP_AT_GOP_END`` flag
 	is set, encoding will continue until the end of the current *Group
 	Of Pictures*, otherwise encoding will stop immediately. When the
-	encoder is already stopped, this command does nothing. mem2mem
-	encoders will send a ``V4L2_EVENT_EOS`` event when the last frame
-	has been encoded and all frames are ready to be dequeued and will
-	set the ``V4L2_BUF_FLAG_LAST`` buffer flag on the last buffer of
-	the capture queue to indicate there will be no new buffers
-	produced to dequeue. This buffer may be empty, indicated by the
-	driver setting the ``bytesused`` field to 0. Once the
-	``V4L2_BUF_FLAG_LAST`` flag was set, the
-	:ref:`VIDIOC_DQBUF <VIDIOC_QBUF>` ioctl will not block anymore,
-	but return an ``EPIPE`` error code.
+	encoder is already stopped, this command does nothing.
+
+	A stateful mem2mem encoder will proceed with encoding the source
+	buffers pending before the command is issued and then stop producing
+	new frames. It will send a ``V4L2_EVENT_EOS`` event when the last frame
+	has been encoded and all frames are ready to be dequeued and will set
+	the ``V4L2_BUF_FLAG_LAST`` buffer flag on the last buffer of the
+	capture queue to indicate there will be no new buffers produced to
+	dequeue. This buffer may be empty, indicated by the driver setting the
+	``bytesused`` field to 0. Once the buffer with the
+	``V4L2_BUF_FLAG_LAST`` flag set was dequeued, the :ref:`VIDIOC_DQBUF
+	<VIDIOC_QBUF>` ioctl will not block anymore, but return an ``EPIPE``
+	error code. No flags or other arguments are accepted in case of mem2mem
+	encoders.  See :ref:`encoder` for more details.
     * - ``V4L2_ENC_CMD_PAUSE``
       - 2
       - Pause the encoder. When the encoder has not been started yet, the