From: Hans Verkuil <hverkuil@xs4all.nl>
To: "Jernej Škrabec" <jernej.skrabec@gmail.com>,
mchehab@kernel.org, ezequiel@vanguardiasur.com.ar,
p.zabel@pengutronix.de, gregkh@linuxfoundation.org,
mripard@kernel.org, paul.kocialkowski@bootlin.com, wens@csie.org,
samuel@sholland.org, nicolas.dufresne@collabora.com,
andrzej.p@collabora.com,
"Benjamin Gaignard" <benjamin.gaignard@collabora.com>
Cc: linux-media@vger.kernel.org, linux-kernel@vger.kernel.org,
linux-rockchip@lists.infradead.org,
linux-staging@lists.linux.dev,
linux-arm-kernel@lists.infradead.org,
linux-sunxi@lists.linux.dev, kernel@collabora.com
Subject: Re: [PATCH v6 11/17] media: uapi: Add V4L2_CID_STATELESS_HEVC_ENTRY_POINT_OFFSETS control
Date: Mon, 30 May 2022 15:49:57 +0200 [thread overview]
Message-ID: <c47b3218-069b-790e-8781-7b6d195ac8a7@xs4all.nl> (raw)
In-Reply-To: <c639efc2-2089-359d-44c6-e9d92e8fd22e@xs4all.nl>
On 30/05/2022 11:18, Hans Verkuil wrote:
> On 29/05/2022 08:40, Jernej Škrabec wrote:
>> Hi!
>>
>> This series looks very good and I plan to test it shortly on Cedrus, but I
>> have one major concern below.
>>
>> Dne petek, 27. maj 2022 ob 16:31:28 CEST je Benjamin Gaignard napisal(a):
>>> The number of 'entry point offset' can be very variable.
>>> Instead of using a large static array define a v4l2 dynamic array
>>> of U32 (V4L2_CTRL_TYPE_U32).
>>> The number of entry point offsets is reported by the elems field
>>> and in struct v4l2_ctrl_hevc_slice_params.num_entry_point_offsets
>>> field.
>>
>> Slice control by itself is variable length array, so you would actually need
>> 2D variable array for entry points which is not supported. However, easy
>> workaround for that is to flatten 2D array to 1D and either have another slice
>> control field which would tell first entry point index for convenience or let
>> driver calculate it by adding up all num_entry_point_offsets of previous
>> slices.
>>
>> Hans, what do you think?
>
> If I would support 2D variable array sizes, wouldn't that be more elegant?
>
> The current implementation doesn't support that, but as the commit log for
> patch 1/17 says:
>
> "Currently dynamically sized arrays are limited to one dimensional arrays,
> but that might change in the future if there is a need for it."
>
> Let me know if you agree, and I'll try to implement this. It's been a while
> since I last looked at this, so I'm not sure how much work it is, but it is
> probably worth a shot.
Digging more into this made me realize that this doesn't actually help for this
particular case.
I would lean towards your second suggestion of adding up all num_entry_point_offsets
of previous slices.
Regards,
Hans
>
> Regards,
>
> Hans
>
>>
>> Note, it seems that H265 decoding on Cedrus still works without entry points,
>> so this problem can be solved later. I'm not sure what we lose with that but
>> it was suggested that this could influence speed or error resilience or both.
>> However, I think we're close to solve it, so I'd like to do that now.
>>
>> Best regards,
>> Jernej
>>
>>>
>>> Signed-off-by: Benjamin Gaignard <benjamin.gaignard@collabora.com>
>>> ---
>>> .../userspace-api/media/v4l/ext-ctrls-codec.rst | 11 +++++++++++
>>> drivers/media/v4l2-core/v4l2-ctrls-defs.c | 5 +++++
>>> include/media/hevc-ctrls.h | 5 ++++-
>>> 3 files changed, 20 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/Documentation/userspace-api/media/v4l/ext-ctrls-codec.rst b/
>> Documentation/userspace-api/media/v4l/ext-ctrls-codec.rst
>>> index 0796b1563daa..05228e280f66 100644
>>> --- a/Documentation/userspace-api/media/v4l/ext-ctrls-codec.rst
>>> +++ b/Documentation/userspace-api/media/v4l/ext-ctrls-codec.rst
>>> @@ -3010,6 +3010,9 @@ enum v4l2_mpeg_video_hevc_size_of_length_field -
>>> * - __u32
>>> - ``data_bit_offset``
>>> - Offset (in bits) to the video data in the current slice data.
>>> + * - __u32
>>> + - ``num_entry_point_offsets``
>>> + - Specifies the number of entry point offset syntax elements in the
>> slice header.
>>> * - __u8
>>> - ``nal_unit_type``
>>> - Specifies the coding type of the slice (B, P or I).
>>> @@ -3150,6 +3153,14 @@ enum v4l2_mpeg_video_hevc_size_of_length_field -
>>>
>>> \normalsize
>>>
>>> +``V4L2_CID_STATELESS_HEVC_ENTRY_POINT_OFFSETS (integer)``
>>> + Specifies entry point offsets in bytes.
>>> + This control is a dynamically sized array. The number of entry point
>>> + offsets is reported by the ``elems`` field.
>>> + This bitstream parameter is defined according to :ref:`hevc`.
>>> + They are described in section 7.4.7.1 "General slice segment header
>>> + semantics" of the specification.
>>> +
>>> ``V4L2_CID_STATELESS_HEVC_SCALING_MATRIX (struct)``
>>> Specifies the HEVC scaling matrix parameters used for the scaling
>> process
>>> for transform coefficients.
>>> diff --git a/drivers/media/v4l2-core/v4l2-ctrls-defs.c b/drivers/media/v4l2-
>> core/v4l2-ctrls-defs.c
>>> index d594efbcbb93..e22921e7ea61 100644
>>> --- a/drivers/media/v4l2-core/v4l2-ctrls-defs.c
>>> +++ b/drivers/media/v4l2-core/v4l2-ctrls-defs.c
>>> @@ -1188,6 +1188,7 @@ const char *v4l2_ctrl_get_name(u32 id)
>>> case V4L2_CID_STATELESS_HEVC_DECODE_PARAMS: return
>> "HEVC Decode Parameters";
>>> case V4L2_CID_STATELESS_HEVC_DECODE_MODE: return
>> "HEVC Decode Mode";
>>> case V4L2_CID_STATELESS_HEVC_START_CODE: return
>> "HEVC Start Code";
>>> + case V4L2_CID_STATELESS_HEVC_ENTRY_POINT_OFFSETS: return
>> "HEVC Entry Point Offsets";
>>>
>>> /* Colorimetry controls */
>>> /* Keep the order of the 'case's the same as in v4l2-controls.h!
>> */
>>> @@ -1518,6 +1519,10 @@ void v4l2_ctrl_fill(u32 id, const char **name, enum
>> v4l2_ctrl_type *type,
>>> case V4L2_CID_STATELESS_HEVC_DECODE_PARAMS:
>>> *type = V4L2_CTRL_TYPE_HEVC_DECODE_PARAMS;
>>> break;
>>> + case V4L2_CID_STATELESS_HEVC_ENTRY_POINT_OFFSETS:
>>> + *type = V4L2_CTRL_TYPE_U32;
>>> + *flags |= V4L2_CTRL_FLAG_DYNAMIC_ARRAY;
>>> + break;
>>> case V4L2_CID_STATELESS_VP9_COMPRESSED_HDR:
>>> *type = V4L2_CTRL_TYPE_VP9_COMPRESSED_HDR;
>>> break;
>>> diff --git a/include/media/hevc-ctrls.h b/include/media/hevc-ctrls.h
>>> index a3c829ef531a..1319cb99ae3f 100644
>>> --- a/include/media/hevc-ctrls.h
>>> +++ b/include/media/hevc-ctrls.h
>>> @@ -20,6 +20,7 @@
>>> #define V4L2_CID_STATELESS_HEVC_DECODE_PARAMS (V4L2_CID_CODEC_BASE
>> + 1012)
>>> #define V4L2_CID_STATELESS_HEVC_DECODE_MODE (V4L2_CID_CODEC_BASE + 1015)
>>> #define V4L2_CID_STATELESS_HEVC_START_CODE (V4L2_CID_CODEC_BASE + 1016)
>>> +#define V4L2_CID_STATELESS_HEVC_ENTRY_POINT_OFFSETS (V4L2_CID_CODEC_BASE +
>> 1017)
>>>
>>> /* enum v4l2_ctrl_type type values */
>>> #define V4L2_CTRL_TYPE_HEVC_SPS 0x0120
>>> @@ -318,6 +319,8 @@ struct v4l2_hevc_pred_weight_table {
>>> *
>>> * @bit_size: size (in bits) of the current slice data
>>> * @data_bit_offset: offset (in bits) to the video data in the current slice
>> data
>>> + * @num_entry_point_offsets: specifies the number of entry point offset syntax
>>> + * elements in the slice header.
>>> * @nal_unit_type: specifies the coding type of the slice (B, P or I)
>>> * @nuh_temporal_id_plus1: minus 1 specifies a temporal identifier for the
>> NAL unit
>>> * @slice_type: see V4L2_HEVC_SLICE_TYPE_{}
>>> @@ -360,7 +363,7 @@ struct v4l2_hevc_pred_weight_table {
>>> struct v4l2_ctrl_hevc_slice_params {
>>> __u32 bit_size;
>>> __u32 data_bit_offset;
>>> -
>>> + __u32 num_entry_point_offsets;
>>> /* ISO/IEC 23008-2, ITU-T Rec. H.265: NAL unit header */
>>> __u8 nal_unit_type;
>>> __u8 nuh_temporal_id_plus1;
>>> --
>>> 2.32.0
>>>
>>>
>>
>>
>
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
next prev parent reply other threads:[~2022-05-30 14:21 UTC|newest]
Thread overview: 36+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-05-27 14:31 [PATCH v6 00/17] Move HEVC stateless controls out of staging Benjamin Gaignard
2022-05-27 14:31 ` [PATCH v6 01/17] videodev2.h: add V4L2_CTRL_FLAG_DYNAMIC_ARRAY Benjamin Gaignard
2022-05-27 14:31 ` [PATCH v6 02/17] v4l2-ctrls: add support for dynamically allocated arrays Benjamin Gaignard
2022-05-27 14:31 ` [PATCH v6 03/17] vivid: add dynamic array test control Benjamin Gaignard
2022-05-27 14:31 ` [PATCH v6 04/17] media: uapi: HEVC: Add missing fields in HEVC controls Benjamin Gaignard
2022-05-27 14:31 ` [PATCH v6 05/17] media: uapi: HEVC: Rename HEVC stateless controls with STATELESS prefix Benjamin Gaignard
2022-05-27 14:31 ` [PATCH v6 06/17] media: uapi: HEVC: Change pic_order_cnt definition in v4l2_hevc_dpb_entry Benjamin Gaignard
2022-05-27 14:31 ` [PATCH v6 07/17] media: uapi: HEVC: Add SEI pic struct flags Benjamin Gaignard
2022-05-27 14:31 ` [PATCH v6 08/17] media: uapi: HEVC: Add documentation to uAPI structure Benjamin Gaignard
2022-05-27 14:31 ` [PATCH v6 09/17] media: uapi: HEVC: Define V4L2_CID_STATELESS_HEVC_SLICE_PARAMS as a dynamic array Benjamin Gaignard
2022-05-29 9:19 ` Jernej Škrabec
2022-06-01 15:43 ` Jernej Škrabec
2022-06-01 16:02 ` Benjamin Gaignard
2022-05-27 14:31 ` [PATCH v6 10/17] media: uapi: Move parsed HEVC pixel format out of staging Benjamin Gaignard
2022-05-27 14:31 ` [PATCH v6 11/17] media: uapi: Add V4L2_CID_STATELESS_HEVC_ENTRY_POINT_OFFSETS control Benjamin Gaignard
2022-05-29 6:40 ` Jernej Škrabec
2022-05-30 9:18 ` Hans Verkuil
2022-05-30 13:49 ` Hans Verkuil [this message]
2022-05-30 21:24 ` Jernej Škrabec
2022-05-31 6:58 ` Benjamin Gaignard
2022-05-31 18:20 ` Jernej Škrabec
2022-06-01 16:20 ` Nicolas Dufresne
2022-06-01 16:35 ` Jernej Škrabec
2022-06-01 17:07 ` Nicolas Dufresne
2022-06-10 14:08 ` John Cox
2022-05-27 14:31 ` [PATCH v6 12/17] media: uapi: Move the HEVC stateless control type out of staging Benjamin Gaignard
2022-05-27 14:31 ` [PATCH v6 13/17] media: controls: Log HEVC stateless control in .std_log Benjamin Gaignard
2022-05-27 14:31 ` [PATCH v6 14/17] media: hantro: Stop using Hantro dedicated control Benjamin Gaignard
2022-05-27 14:31 ` [PATCH v6 15/17] media: uapi: HEVC: fix padding in v4l2 control structures Benjamin Gaignard
2022-05-27 14:31 ` [PATCH v6 16/17] media: uapi: Change data_bit_offset definition Benjamin Gaignard
2022-05-29 6:45 ` Jernej Škrabec
2022-06-01 16:17 ` Jernej Škrabec
2022-06-01 16:33 ` Benjamin Gaignard
2022-06-12 20:40 ` Jernej Škrabec
2022-06-13 18:17 ` Jernej Škrabec
2022-06-10 14:00 ` [PATCH v6 00/17] Move HEVC stateless controls out of staging John Cox
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=c47b3218-069b-790e-8781-7b6d195ac8a7@xs4all.nl \
--to=hverkuil@xs4all.nl \
--cc=andrzej.p@collabora.com \
--cc=benjamin.gaignard@collabora.com \
--cc=ezequiel@vanguardiasur.com.ar \
--cc=gregkh@linuxfoundation.org \
--cc=jernej.skrabec@gmail.com \
--cc=kernel@collabora.com \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-media@vger.kernel.org \
--cc=linux-rockchip@lists.infradead.org \
--cc=linux-staging@lists.linux.dev \
--cc=linux-sunxi@lists.linux.dev \
--cc=mchehab@kernel.org \
--cc=mripard@kernel.org \
--cc=nicolas.dufresne@collabora.com \
--cc=p.zabel@pengutronix.de \
--cc=paul.kocialkowski@bootlin.com \
--cc=samuel@sholland.org \
--cc=wens@csie.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).