media: hantro: Add support for H264 decoding
Subsystems (7)
v5 cover media: hantro: Add support for H264 decoding
A new version of H264 uAPI review and Hantro G1 H264 decoding support. Compared to previous version, this version takes care of the suggestions from Hans Verkuil clarifying the uAPI spec.
This series consolidates the two recent H264 series submitted by Boris [1] [2]. Some patches from [2] have been merged (namely, helpers for the Hantro driver), and so I'm adding the remanining bits required to support H264 on Hantro G1 VPU.
* Patch 1 adds support for the sort_r() variant and has
been posted separately by Rasmus. It would be good to merge this patch via the media tree, ideally as soon as possible, to avoid the synchronisation burden that might appear if we decide to delay it.
* Patch 2 to 4 extends the H264 uAPI, introducing frame-based vs slice-based decoding granularity, and also support for different NALU start codes. Currently, Annex B and no start codes are the supported options.
With the introduction of the start code control, the H264 parsed slices pixel format should be renamed, dropping the _RAW suffix, which is now meaningless.
* Patch 5 removes the P0/B0/B1 ref lists from the decode_params control. These lists are no longer needed since we build them on the kernel side based on the DPB.
* Patch 6 and 7 exposes the proper decoding mode and start code on the cedrus driver. The driver functionality is not changed, and only the Cedrus support is now being properly exposed to userspace.
* Patch 8 is needed to properly propagate the OUTPUT buffer timestamp to the CAPTURE buffer one, which is required for intra-frame references.
* Patches 9 to 11 adds H264 support for Hantro G1 and then enable H264 decoding on RK3288.
This is based on media master and tested on Rockchip RK3288 for Hantro and Allwinner H3 boards for Cedrus. Philipp Zabel tested on i.MX8MQ EVK using [3].
The Ffmpeg branch used to test is based on the great work of Jonas and Boris, and is available in [4]. Instructions to build and run are as follows:
./configure --enable-v4l2-request --enable-libdrm
make -j4
(test via framebuffer rendering):
./ffmpeg -loglevel debug -hwaccel drm -hwaccel_device /dev/dri/card0 -i $some_file.avi -pix_fmt bgra -f fbdev /dev/fb0
[1] https://www.mail-archive.com/[email protected]/msg148299.html
[2] https://lkml.org/lkml/2019/6/19/379
[3] git://git.pengutronix.de/git/pza/linux.git hantro/imx8m-wip
[4] https://gitlab.collabora.com/ezequiel/ffmpeg/tree/stateless-mpeg2-vp8-h264-v4
Boris Brezillon (3):
media: uapi: h264: Add the concept of decoding mode
media: uapi: h264: Get rid of the p0/b0/b1 ref-lists
media: hantro: Move copy_metadata() before doing a decode operation
Ezequiel Garcia (4):
media: uapi: h264: Rename pixel format
media: uapi: h264: Add the concept of start code
media: cedrus: Cleanup control initialization
media: cedrus: Specify H264 startcode and decoding mode
Hertz Wong (3):
media: hantro: Add core bits to support H264 decoding
media: hantro: Add support for H264 decoding on G1
media: hantro: Enable H264 decoding on rk3288
Rasmus Villemoes (1):
lib/sort.c: implement sort() variant taking context argument
.../media/uapi/v4l/ext-ctrls-codec.rst | 89 ++- .../media/uapi/v4l/pixfmt-compressed.rst | 8 +- drivers/media/v4l2-core/v4l2-ctrls.c | 18 + drivers/media/v4l2-core/v4l2-ioctl.c | 2 +- drivers/staging/media/hantro/Makefile | 2 + drivers/staging/media/hantro/hantro.h | 9 +- drivers/staging/media/hantro/hantro_drv.c | 50 +- .../staging/media/hantro/hantro_g1_h264_dec.c | 292 ++++++++ drivers/staging/media/hantro/hantro_h264.c | 651 ++++++++++++++++++ drivers/staging/media/hantro/hantro_hw.h | 56 ++ drivers/staging/media/hantro/hantro_v4l2.c | 10 + drivers/staging/media/hantro/rk3288_vpu_hw.c | 21 +- drivers/staging/media/sunxi/cedrus/cedrus.c | 65 +- drivers/staging/media/sunxi/cedrus/cedrus.h | 3 +- .../staging/media/sunxi/cedrus/cedrus_dec.c | 2 +- .../staging/media/sunxi/cedrus/cedrus_video.c | 6 +- include/linux/sort.h | 5 + include/media/h264-ctrls.h | 22 +- lib/sort.c | 34 +- 19 files changed, 1290 insertions(+), 55 deletions(-) create mode 100644 drivers/staging/media/hantro/hantro_g1_h264_dec.c create mode 100644 drivers/staging/media/hantro/hantro_h264.c
Signature · 1 line
2.22.0
PATCH v5 1/11 lib/sort.c: implement sort() variant taking context argument
From: Rasmus Villemoes <linux-qQsb+v5E8BnlAoU/[email protected]>
Our list_sort() utility has always supported a context argument that is passed through to the comparison routine. Now there's a use case for the similar thing for sort().
This implements sort_r by simply extending the existing sort function in the obvious way. To avoid code duplication, we want to implement sort() in terms of sort_r(). The naive way to do that is
static int cmp_wrapper(const void *a, const void *b, const void *ctx) {
int (*real_cmp)(const void*, const void*) = ctx;
return real_cmp(a, b);
}
sort(..., cmp) { sort_r(..., cmp_wrapper, cmp) }
but this would do two indirect calls for each comparison. Instead, do as is done for the default swap functions - that only adds a cost of a single easily predicted branch to each comparison call.
Aside from introducing support for the context argument, this also serves as preparation for patches that will eliminate the indirect comparison calls in common cases.
--- include/linux/sort.h | 5 +++++ lib/sort.c | 34 ++++++++++++++++++++++++++++------ 2 files changed, 33 insertions(+), 6 deletions(-)
include/linux/sort.h +5 -0
@@ -4,6 +4,11 @@#include <linux/types.h> +void sort_r (void *base,size_t num,size_t size, +int (*cmp)(const void *,const void *,const void *), +void (*swap)(void *,void *,int ), +const void *priv); +void sort (void *base,size_t num,size_t size,int (*cmp)(const void *,const void *),void (*swap)(void *,void *,int ));
lib/sort.c +28 -6
@@ -144,6 +144,18 @@ static void do_swap(void *a, void *b, size_t size, swap_func_t swap_func)swap_func (a, b, (int )size); } +typedef int (*cmp_func_t )(const void *,const void *); +typedef int (*cmp_r_func_t )(const void *,const void *,const void *); +#define _CMP_WRAPPER ((cmp_r_func_t)0L) + +static int do_cmp (const void *a,const void *b, +cmp_r_func_t cmp,const void *priv) +{ +if (cmp == _CMP_WRAPPER) +return ((cmp_func_t)(priv))(a, b); +return cmp (a, b, priv); +} +/** * parent - given the offset of the child, find the offset of the parent. * @i: the offset of the heap element whose parent is sought. Non-zero.
@@ -171,12 +183,13 @@ static size_t parent(size_t i, unsigned int lsbit, size_t size) }/** -* sort - sort an array of elements +* sort_r - sort an array of elements * @base: pointer to data to sort * @num: number of elements * @size: size of each element * @cmp_func: pointer to comparison function * @swap_func: pointer to swap function or NULL +* @priv: third argument passed to comparison function * * This function does a heapsort on the given array. You may provide * a swap_func function if you need to do something more than a memory
@@ -188,9 +201,10 @@ static size_t parent(size_t i, unsigned int lsbit, size_t size)* O(n*n) worst-case behavior and extra memory requirements that make * it less suitable for kernel use. */ -void sort (void *base,size_t num,size_t size, -int (*cmp_func)(const void *,const void *), -void (*swap_func)(void *,void *,int size)) +void sort_r (void *base,size_t num,size_t size, +int (*cmp_func)(const void *,const void *,const void *), +void (*swap_func)(void *,void *,int size), +const void *priv) {/* pre-scale counters for performance */ size_t n = num * size, a = (num/2 ) * size;
@@ -238,12 +252,12 @@ void sort(void *base, size_t num, size_t size,* average, 3/4 worst-case.) */ for (b = a; c =2 *b + size, (d = c + size) < n;) - b =cmp_func (base + c, base + d) >=0 ? c : d; + b =do_cmp (base + c, base + d, cmp_func, priv) >=0 ? c : d;if (d == n)/* Special case last leaf with no sibling */ b = c;/* Now backtrack from "b" to the correct location for "a" */ -while (b != a &&cmp_func (base + a, base + b) >=0 ) +while (b != a &&do_cmp (base + a, base + b, cmp_func, priv) >=0 ) b =parent (b, lsbit, size); c = b;/* Where "a" belongs */ while (b != a) {/* Shift it into place */
@@ -252,4 +266,12 @@ void sort(void *base, size_t num, size_t size, } } } +EXPORT_SYMBOL (sort_r); + +void sort (void *base,size_t num,size_t size, +int (*cmp_func)(const void *,const void *), +void (*swap_func)(void *,void *,int size)) +{ +return sort_r (base, num, size, _CMP_WRAPPER, swap_func, cmp_func); +}EXPORT_SYMBOL (sort);
Signature · 1 line
2.22.0
PATCH v5 2/11 media: uapi: h264: Rename pixel format
In reply to Ezequiel GarciaThe V4L2_PIX_FMT_H264_SLICE_RAW name was originally suggested because the pixel format would represent H264 slices without any start code.
However, as we will now introduce a start code menu control, give the pixel format a more meaningful name, while it's still early enough to do so.
---
Changes in v5:
* None.
Changes in v4:
* New patch.
--- Documentation/media/uapi/v4l/pixfmt-compressed.rst | 4 ++-- drivers/media/v4l2-core/v4l2-ioctl.c | 2 +- drivers/staging/media/sunxi/cedrus/cedrus_dec.c | 2 +- drivers/staging/media/sunxi/cedrus/cedrus_video.c | 6 +++--- include/media/h264-ctrls.h | 2 +- 5 files changed, 8 insertions(+), 8 deletions(-)
Documentation/media/uapi/v4l/pixfmt-compressed.rst +2 -2
@@ -52,9 +52,9 @@ Compressed Formats - ``V4L2_PIX_FMT_H264_MVC`` - 'M264' - H264 MVC video elementary stream. - * .. _V4L2-PIX-FMT-H264-SLICE-RAW: + * .. _V4L2-PIX-FMT-H264-SLICE: - - ``V4L2_PIX_FMT_H264_SLICE_RAW`` + - ``V4L2_PIX_FMT_H264_SLICE`` - 'S264' - H264 parsed slice data, without the start code and as extracted from the H264 bitstream. This format is adapted for
drivers/media/v4l2-core/v4l2-ioctl.c +1 -1
@@ -1343,7 +1343,7 @@ static void v4l_fill_fmtdesc(struct v4l2_fmtdesc *fmt)case V4L2_PIX_FMT_H264 : descr ="H.264" ;break ;case V4L2_PIX_FMT_H264_NO_SC : descr ="H.264 (No Start Codes)" ;break ;case V4L2_PIX_FMT_H264_MVC : descr ="H.264 MVC" ;break ; -case V4L2_PIX_FMT_H264_SLICE_RAW : descr ="H.264 Parsed Slice Data" ;break ; +case V4L2_PIX_FMT_H264_SLICE : descr ="H.264 Parsed Slice Data" ;break ;case V4L2_PIX_FMT_H263 : descr ="H.263" ;break ;case V4L2_PIX_FMT_MPEG1 : descr ="MPEG-1 ES" ;break ;case V4L2_PIX_FMT_MPEG2 : descr ="MPEG-2 ES" ;break ;
drivers/staging/media/sunxi/cedrus/cedrus_dec.c +1 -1
@@ -46,7 +46,7 @@ void cedrus_device_run(void *priv) V4L2_CID_MPEG_VIDEO_MPEG2_QUANTIZATION);break ; -case V4L2_PIX_FMT_H264_SLICE_RAW : +case V4L2_PIX_FMT_H264_SLICE : run.h264.decode_params =cedrus_find_control_data (ctx,V4L2_CID_MPEG_VIDEO_H264_DECODE_PARAMS ); run.h264.pps = cedrus_find_control_data(ctx,
drivers/staging/media/sunxi/cedrus/cedrus_video.c +3 -3
@@ -38,7 +38,7 @@ static struct cedrus_format cedrus_formats[] = { .directions = CEDRUS_DECODE_SRC, }, { - .pixelformat = V4L2_PIX_FMT_H264_SLICE_RAW, + .pixelformat = V4L2_PIX_FMT_H264_SLICE, .directions = CEDRUS_DECODE_SRC, }, {
@@ -104,7 +104,7 @@ static void cedrus_prepare_format(struct v4l2_pix_format *pix_fmt)switch (pix_fmt->pixelformat) {case V4L2_PIX_FMT_MPEG2_SLICE : -case V4L2_PIX_FMT_H264_SLICE_RAW : +case V4L2_PIX_FMT_H264_SLICE :/* Zero bytes per line for encoded source. */ bytesperline =0 ;
@@ -469,7 +469,7 @@ static int cedrus_start_streaming(struct vb2_queue *vq, unsigned int count) ctx->current_codec =CEDRUS_CODEC_MPEG2 ;break ; -case V4L2_PIX_FMT_H264_SLICE_RAW : +case V4L2_PIX_FMT_H264_SLICE : ctx->current_codec =CEDRUS_CODEC_H264 ;break ;
include/media/h264-ctrls.h +1 -1
@@ -14,7 +14,7 @@#include <linux/videodev2.h> /* Our pixel format isn't stable at the moment */ -#define V4L2_PIX_FMT_H264_SLICE_RAW v4l2_fourcc('S', '2', '6', '4')/* H264 parsed slices */ +#define V4L2_PIX_FMT_H264_SLICE v4l2_fourcc('S', '2', '6', '4')/* H264 parsed slices */ /* * This is put insanely high to avoid conflicting with controls that
Signature · 1 line
2.22.0
PATCH v5 3/11 media: uapi: h264: Add the concept of decoding mode
In reply to Ezequiel GarciaFrom: Boris Brezillon <[email protected]>
Some stateless decoders don't support per-slice decoding granularity (or at least not in a way that would make them efficient or easy to use).
Expose a menu to control the supported decoding modes. Drivers are allowed to support only one decoding but they can support both too.
---
Changes in v5:
* Improve specification as suggested by Hans. Changes in v4:
* Typos/rewording fixes
Changes in v3:
* s/per-{slice,frame} decoding/{slice,frame}-based decoding/
* Add Paul's R-b
Changes in v2:
* Allow decoding multiple slices in per-slice decoding mode
* Minor doc improvement/fixes
--- .../media/uapi/v4l/ext-ctrls-codec.rst | 47 ++++++++++++++++++- .../media/uapi/v4l/pixfmt-compressed.rst | 3 +- drivers/media/v4l2-core/v4l2-ctrls.c | 9 ++++ include/media/h264-ctrls.h | 11 +++++ 4 files changed, 68 insertions(+), 2 deletions(-)
Documentation/media/uapi/v4l/ext-ctrls-codec.rst +46 -1
@@ -1747,6 +1747,11 @@ enum v4l2_mpeg_video_h264_hierarchical_coding_type - * - __u32 - ``size`` - + * - __u32 + - ``start_byte_offset`` + - Where the slice payload starts in the output buffer. Useful when the + OUTPUT buffer contains more than one slice (some codecs need to know + where each slice starts in this buffer). * - __u32 - ``header_bit_size`` -
@@ -1930,7 +1935,10 @@ enum v4l2_mpeg_video_h264_hierarchical_coding_type - - * - __u16 - ``num_slices`` - - Number of slices needed to decode the current frame + - Number of slices needed to decode the current frame/field. When + operating in slice-based decoding mode (see + :c:type:`v4l2_mpeg_video_h264_decoding_mode`), this field + should always be set to one. * - __u16 - ``nal_ref_idc`` - NAL reference ID value coming from the NAL Unit header
@@ -2021,6 +2029,43 @@ enum v4l2_mpeg_video_h264_hierarchical_coding_type - - 0x00000004 - The DPB entry is a long term reference frame +``V4L2_CID_MPEG_VIDEO_H264_DECODING_MODE (enum)`` + Specifies the decoding mode to use. Currently exposes slice-based and + frame-based decoding but new modes might be added later on. + This control is used to complement V4L2_PIX_FMT_H264_SLICE + pixel format. Applications that support V4L2_PIX_FMT_H264_SLICE + are required to set this control in order to specify the decoding mode + that is expected for the buffer. + Drivers may expose a single or multiple decoding modes, depending + on what they can support. + + .. note:: + + This menu control is not yet part of the public kernel API and + it is expected to change. + +.. c:type:: v4l2_mpeg_video_h264_decoding_mode + +.. cssclass:: longtable + +.. flat-table:: + :header-rows: 0 + :stub-columns: 0 + :widths: 1 1 2 + + * - ``V4L2_MPEG_VIDEO_H264_SLICE_BASED_DECODING`` + - 0 + - The decoding is done at the slice granularity. + v4l2_ctrl_h264_decode_params->num_slices should be set to 1. + The OUTPUT buffer must contain a single slice. + * - ``V4L2_MPEG_VIDEO_H264_FRAME_BASED_DECODING`` + - 1 + - The decoding is done at the frame granularity. + v4l2_ctrl_h264_decode_params->num_slices should be set to the number of + slices forming a frame. + The OUTPUT buffer must contain all slices needed to decode the + frame. The OUTPUT buffer must also contain both fields. + .. _v4l2-mpeg-mpeg2: ``V4L2_CID_MPEG_VIDEO_MPEG2_SLICE_PARAMS (struct)``
Documentation/media/uapi/v4l/pixfmt-compressed.rst +2 -1
@@ -61,7 +61,8 @@ Compressed Formats stateless video decoders that implement an H264 pipeline (using the :ref:`mem2mem` and :ref:`media-request-api`). Metadata associated with the frame to decode are required to - be passed through the ``V4L2_CID_MPEG_VIDEO_H264_SPS``, + be passed through the ``V4L2_CID_MPEG_VIDEO_H264_DECODING_MODE``, + ``V4L2_CID_MPEG_VIDEO_H264_SPS``, ``V4L2_CID_MPEG_VIDEO_H264_PPS``, ``V4L2_CID_MPEG_VIDEO_H264_SCALING_MATRIX``, ``V4L2_CID_MPEG_VIDEO_H264_SLICE_PARAMS`` and
drivers/media/v4l2-core/v4l2-ctrls.c +9 -0
@@ -402,6 +402,11 @@ const char * const *v4l2_ctrl_get_menu(u32 id)"Explicit" , NULL, }; +static const char *const h264_decoding_mode[] = { +"Slice-Based" , +"Frame-Based" , +NULL , + };static const char *const mpeg_mpeg2_level[] = {"Low" , "Main",
@@ -633,6 +638,8 @@ const char * const *v4l2_ctrl_get_menu(u32 id)return h264_fp_arrangement_type;case V4L2_CID_MPEG_VIDEO_H264_FMO_MAP_TYPE :return h264_fmo_map_type; +case V4L2_CID_MPEG_VIDEO_H264_DECODING_MODE : +return h264_decoding_mode;case V4L2_CID_MPEG_VIDEO_MPEG2_LEVEL :return mpeg_mpeg2_level;case V4L2_CID_MPEG_VIDEO_MPEG2_PROFILE :
@@ -852,6 +859,7 @@ const char *v4l2_ctrl_get_name(u32 id)case V4L2_CID_MPEG_VIDEO_H264_SCALING_MATRIX :return "H264 Scaling Matrix" ;case V4L2_CID_MPEG_VIDEO_H264_SLICE_PARAMS :return "H264 Slice Parameters" ;case V4L2_CID_MPEG_VIDEO_H264_DECODE_PARAMS :return "H264 Decode Parameters" ; +case V4L2_CID_MPEG_VIDEO_H264_DECODING_MODE :return "H264 Decoding Mode" ;case V4L2_CID_MPEG_VIDEO_MPEG2_LEVEL :return "MPEG2 Level" ;case V4L2_CID_MPEG_VIDEO_MPEG2_PROFILE :return "MPEG2 Profile" ;case V4L2_CID_MPEG_VIDEO_MPEG4_I_FRAME_QP :return "MPEG4 I-Frame QP Value" ;
@@ -1220,6 +1228,7 @@ void v4l2_ctrl_fill(u32 id, const char **name, enum v4l2_ctrl_type *type,case V4L2_CID_MPEG_VIDEO_H264_VUI_SAR_IDC :case V4L2_CID_MPEG_VIDEO_H264_SEI_FP_ARRANGEMENT_TYPE :case V4L2_CID_MPEG_VIDEO_H264_FMO_MAP_TYPE : +case V4L2_CID_MPEG_VIDEO_H264_DECODING_MODE :case V4L2_CID_MPEG_VIDEO_MPEG2_LEVEL :case V4L2_CID_MPEG_VIDEO_MPEG2_PROFILE :case V4L2_CID_MPEG_VIDEO_MPEG4_LEVEL :
include/media/h264-ctrls.h +11 -0
@@ -26,6 +26,7 @@#define V4L2_CID_MPEG_VIDEO_H264_SCALING_MATRIX (V4L2_CID_MPEG_BASE+1002)#define V4L2_CID_MPEG_VIDEO_H264_SLICE_PARAMS (V4L2_CID_MPEG_BASE+1003)#define V4L2_CID_MPEG_VIDEO_H264_DECODE_PARAMS (V4L2_CID_MPEG_BASE+1004) +#define V4L2_CID_MPEG_VIDEO_H264_DECODING_MODE (V4L2_CID_MPEG_BASE+1005)/* enum v4l2_ctrl_type type values */ #define V4L2_CTRL_TYPE_H264_SPS 0x0110
@@ -33,6 +34,12 @@#define V4L2_CTRL_TYPE_H264_SCALING_MATRIX 0x0112#define V4L2_CTRL_TYPE_H264_SLICE_PARAMS 0x0113#define V4L2_CTRL_TYPE_H264_DECODE_PARAMS 0x0114 +#define V4L2_CTRL_TYPE_H264_DECODING_MODE 0x0115 + +enum v4l2_mpeg_video_h264_decoding_mode { +V4L2_MPEG_VIDEO_H264_SLICE_BASED_DECODING , +V4L2_MPEG_VIDEO_H264_FRAME_BASED_DECODING , +};#define V4L2_H264_SPS_CONSTRAINT_SET0_FLAG 0x01#define V4L2_H264_SPS_CONSTRAINT_SET1_FLAG 0x02
@@ -125,6 +132,10 @@ struct v4l2_h264_pred_weight_table {struct v4l2_ctrl_h264_slice_params {/* Size in bytes, including header */ __u32 size; + +/* Offset in bytes to the start of slice in the OUTPUT buffer. */ +__u32 start_byte_offset; +/* Offset in bits to slice_data() from the beginning of this slice. */ __u32 header_bit_size;
Signature · 1 line
2.22.0
Re: [PATCH v5 03/11] media: uapi: h264: Add the concept of decoding mode
Quoted 106 lines · Ezequiel Garcia
From: Boris Brezillon <[email protected]>
Some stateless decoders don't support per-slice decoding granularity (or at least not in a way that would make them efficient or easy to use).
Expose a menu to control the supported decoding modes. Drivers are allowed to support only one decoding but they can support both too.
Signed-off-by Boris Brezillon <[email protected]>Reviewed-by Paul Kocialkowski <[email protected]>Tested-by Philipp Zabel <[email protected]>---
Changes in v5:
* Improve specification as suggested by Hans. Changes in v4:
* Typos/rewording fixes
Changes in v3:
* s/per-{slice,frame} decoding/{slice,frame}-based decoding/
* Add Paul's R-b
Changes in v2:
* Allow decoding multiple slices in per-slice decoding mode
* Minor doc improvement/fixes--- .../media/uapi/v4l/ext-ctrls-codec.rst | 47 ++++++++++++++++++- .../media/uapi/v4l/pixfmt-compressed.rst | 3 +- drivers/media/v4l2-core/v4l2-ctrls.c | 9 ++++ include/media/h264-ctrls.h | 11 +++++ 4 files changed, 68 insertions(+), 2 deletions(-)Documentation/media/uapi/v4l/ext-ctrls-codec.rst +46
-1@@ -1747,6 +1747,11 @@ enum v4l2_mpeg_video_h264_hierarchical_coding_type - * - __u32 - ``size`` - + * - __u32 + - ``start_byte_offset`` + - Where the slice payload starts in the output buffer. Useful when the + OUTPUT buffer contains more than one slice (some codecs need to know + where each slice starts in this buffer). * - __u32 - ``header_bit_size`` -@@ -1930,7 +1935,10 @@ enum v4l2_mpeg_video_h264_hierarchical_coding_type - - * - __u16 - ``num_slices`` - - Number of slices needed to decode the current frame + - Number of slices needed to decode the current frame/field. When + operating in slice-based decoding mode (see + :c:type:`v4l2_mpeg_video_h264_decoding_mode`), this field + should always be set to one. * - __u16 - ``nal_ref_idc`` - NAL reference ID value coming from the NAL Unit header@@ -2021,6 +2029,43 @@ enum v4l2_mpeg_video_h264_hierarchical_coding_type - - 0x00000004 - The DPB entry is a long term reference frame +``V4L2_CID_MPEG_VIDEO_H264_DECODING_MODE (enum)`` + Specifies the decoding mode to use. Currently exposes slice-based and + frame-based decoding but new modes might be added later on. + This control is used to complement V4L2_PIX_FMT_H264_SLICE + pixel format. Applications that support V4L2_PIX_FMT_H264_SLICE + are required to set this control in order to specify the decoding mode + that is expected for the buffer. + Drivers may expose a single or multiple decoding modes, depending + on what they can support. + + .. note:: + + This menu control is not yet part of the public kernel API and + it is expected to change. + +.. c:type:: v4l2_mpeg_video_h264_decoding_mode + +.. cssclass:: longtable + +.. flat-table:: + :header-rows: 0 + :stub-columns: 0 + :widths: 1 1 2 + + * - ``V4L2_MPEG_VIDEO_H264_SLICE_BASED_DECODING`` + - 0 + - The decoding is done at the slice granularity. + v4l2_ctrl_h264_decode_params->num_slices should be set to 1. + The OUTPUT buffer must contain a single slice. + * - ``V4L2_MPEG_VIDEO_H264_FRAME_BASED_DECODING`` + - 1 + - The decoding is done at the frame granularity. + v4l2_ctrl_h264_decode_params->num_slices should be set to the number of + slices forming a frame. + The OUTPUT buffer must contain all slices needed to decode the + frame. The OUTPUT buffer must also contain both fields. + .. _v4l2-mpeg-mpeg2: ``V4L2_CID_MPEG_VIDEO_MPEG2_SLICE_PARAMS (struct)``Documentation/media/uapi/v4l/pixfmt-compressed.rst +0
-0
Documentation/media/uapi/v4l/pixfmt-compressed.rst Compressed Formats · line 64
Quoted 5 lines
Documentation/media/uapi/v4l/pixfmt-compressed.rst +1
-1@@ -61,7 +61,8 @@ Compressed Formats stateless video decoders that implement an H264 pipeline (using the :ref:`mem2mem` and :ref:`media-request-api`). Metadata associated with the frame to decode are required to - be passed through the ``V4L2_CID_MPEG_VIDEO_H264_SPS``, + be passed through the ``V4L2_CID_MPEG_VIDEO_H264_DECODING_MODE``,
No, that's not right. H264_DECODING_MODE isn't metadata that has to be passed via a request. It's a pixelformat modifier and has to be set only once to describe the format that's going to be passed to the HW.
That should be made clear here. Same goes for the next patch.
Regards,
Hans
Quoted 82 lines
Documentation/media/uapi/v4l/pixfmt-compressed.rst +1
-0+ ``V4L2_CID_MPEG_VIDEO_H264_SPS``, ``V4L2_CID_MPEG_VIDEO_H264_PPS``, ``V4L2_CID_MPEG_VIDEO_H264_SCALING_MATRIX``, ``V4L2_CID_MPEG_VIDEO_H264_SLICE_PARAMS`` anddrivers/media/v4l2-core/v4l2-ctrls.c +9
-0@@ -402,6 +402,11 @@ const char * const *v4l2_ctrl_get_menu(u32 id)"Explicit" , NULL, }; +static const char *const h264_decoding_mode[] = { +"Slice-Based" , +"Frame-Based" , +NULL , + };static const char *const mpeg_mpeg2_level[] = {"Low" , "Main",@@ -633,6 +638,8 @@ const char * const *v4l2_ctrl_get_menu(u32 id)return h264_fp_arrangement_type;case V4L2_CID_MPEG_VIDEO_H264_FMO_MAP_TYPE :return h264_fmo_map_type; +case V4L2_CID_MPEG_VIDEO_H264_DECODING_MODE : +return h264_decoding_mode;case V4L2_CID_MPEG_VIDEO_MPEG2_LEVEL :return mpeg_mpeg2_level;case V4L2_CID_MPEG_VIDEO_MPEG2_PROFILE :@@ -852,6 +859,7 @@ const char *v4l2_ctrl_get_name(u32 id)case V4L2_CID_MPEG_VIDEO_H264_SCALING_MATRIX :return "H264 Scaling Matrix" ;case V4L2_CID_MPEG_VIDEO_H264_SLICE_PARAMS :return "H264 Slice Parameters" ;case V4L2_CID_MPEG_VIDEO_H264_DECODE_PARAMS :return "H264 Decode Parameters" ; +case V4L2_CID_MPEG_VIDEO_H264_DECODING_MODE :return "H264 Decoding Mode" ;case V4L2_CID_MPEG_VIDEO_MPEG2_LEVEL :return "MPEG2 Level" ;case V4L2_CID_MPEG_VIDEO_MPEG2_PROFILE :return "MPEG2 Profile" ;case V4L2_CID_MPEG_VIDEO_MPEG4_I_FRAME_QP :return "MPEG4 I-Frame QP Value" ;@@ -1220,6 +1228,7 @@ void v4l2_ctrl_fill(u32 id, const char **name, enum v4l2_ctrl_type *type,case V4L2_CID_MPEG_VIDEO_H264_VUI_SAR_IDC :case V4L2_CID_MPEG_VIDEO_H264_SEI_FP_ARRANGEMENT_TYPE :case V4L2_CID_MPEG_VIDEO_H264_FMO_MAP_TYPE : +case V4L2_CID_MPEG_VIDEO_H264_DECODING_MODE :case V4L2_CID_MPEG_VIDEO_MPEG2_LEVEL :case V4L2_CID_MPEG_VIDEO_MPEG2_PROFILE :case V4L2_CID_MPEG_VIDEO_MPEG4_LEVEL :include/media/h264-ctrls.h +11
-0@@ -26,6 +26,7 @@#define V4L2_CID_MPEG_VIDEO_H264_SCALING_MATRIX (V4L2_CID_MPEG_BASE+1002)#define V4L2_CID_MPEG_VIDEO_H264_SLICE_PARAMS (V4L2_CID_MPEG_BASE+1003)#define V4L2_CID_MPEG_VIDEO_H264_DECODE_PARAMS (V4L2_CID_MPEG_BASE+1004) +#define V4L2_CID_MPEG_VIDEO_H264_DECODING_MODE (V4L2_CID_MPEG_BASE+1005)/* enum v4l2_ctrl_type type values */ #define V4L2_CTRL_TYPE_H264_SPS 0x0110@@ -33,6 +34,12 @@#define V4L2_CTRL_TYPE_H264_SCALING_MATRIX 0x0112#define V4L2_CTRL_TYPE_H264_SLICE_PARAMS 0x0113#define V4L2_CTRL_TYPE_H264_DECODE_PARAMS 0x0114 +#define V4L2_CTRL_TYPE_H264_DECODING_MODE 0x0115 + +enum v4l2_mpeg_video_h264_decoding_mode { +V4L2_MPEG_VIDEO_H264_SLICE_BASED_DECODING , +V4L2_MPEG_VIDEO_H264_FRAME_BASED_DECODING , +};#define V4L2_H264_SPS_CONSTRAINT_SET0_FLAG 0x01#define V4L2_H264_SPS_CONSTRAINT_SET1_FLAG 0x02@@ -125,6 +132,10 @@ struct v4l2_h264_pred_weight_table {struct v4l2_ctrl_h264_slice_params {/* Size in bytes, including header */ __u32 size; + +/* Offset in bytes to the start of slice in the OUTPUT buffer. */ +__u32 start_byte_offset; +/* Offset in bits to slice_data() from the beginning of this slice. */ __u32 header_bit_size;
Re: [PATCH v5 03/11] media: uapi: h264: Add the concept of decoding mode
In reply to Ezequiel GarciaHi,
Quoted 34 lines · Ezequiel Garcia
From: Boris Brezillon <[email protected]>
Some stateless decoders don't support per-slice decoding granularity (or at least not in a way that would make them efficient or easy to use).
Expose a menu to control the supported decoding modes. Drivers are allowed to support only one decoding but they can support both too.
Signed-off-by Boris Brezillon <[email protected]>Reviewed-by Paul Kocialkowski <[email protected]>Tested-by Philipp Zabel <[email protected]>---
Changes in v5:
* Improve specification as suggested by Hans. Changes in v4:
* Typos/rewording fixes
Changes in v3:
* s/per-{slice,frame} decoding/{slice,frame}-based decoding/
* Add Paul's R-b
Changes in v2:
* Allow decoding multiple slices in per-slice decoding mode
* Minor doc improvement/fixes--- .../media/uapi/v4l/ext-ctrls-codec.rst | 47 ++++++++++++++++++- .../media/uapi/v4l/pixfmt-compressed.rst | 3 +- drivers/media/v4l2-core/v4l2-ctrls.c | 9 ++++ include/media/h264-ctrls.h | 11 +++++ 4 files changed, 68 insertions(+), 2 deletions(-)Documentation/media/uapi/v4l/ext-ctrls-codec.rst +0
-0
Documentation/media/uapi/v4l/ext-ctrls-codec.rst enum v4l2_mpeg_video_h264_hierarchical_coding_type - · line 1754
Quoted 8 lines
Documentation/media/uapi/v4l/ext-ctrls-codec.rst +5
-0@@ -1747,6 +1747,11 @@ enum v4l2_mpeg_video_h264_hierarchical_coding_type - * - __u32 - ``size`` - + * - __u32 + - ``start_byte_offset`` + - Where the slice payload starts in the output buffer. Useful when the + OUTPUT buffer contains more than one slice (some codecs need to know + where each slice starts in this buffer).
I think there is a possibility for misunderstanding here: does the "slice payload" include the start code when it (annex-b) is selected?
I'd suspect that the hardware needs to know where the start code begings rather than where the NAL unit starts when there is a start code. So I'd suggest specifying it as the offset to the beginning of the start code (if present) or to the slice NAL unit start otherwise.
The clarification should probably also concern header_bit_size, to make it clear where it is counted from. I think it makes sense to count it starting from the start_byte_offset position.
Documentation/media/uapi/v4l/ext-ctrls-codec.rst enum v4l2_mpeg_video_h264_hierarchical_coding_type - · line 1941
Quoted 12 lines
Documentation/media/uapi/v4l/ext-ctrls-codec.rst +4
-1* - __u32 - ``header_bit_size`` -@@ -1930,7 +1935,10 @@ enum v4l2_mpeg_video_h264_hierarchical_coding_type - - * - __u16 - ``num_slices`` - - Number of slices needed to decode the current frame + - Number of slices needed to decode the current frame/field. When + operating in slice-based decoding mode (see + :c:type:`v4l2_mpeg_video_h264_decoding_mode`), this field + should always be set to one.
I am fine with the current proposal for now, but I believe that we should keep discussing whether we want per-slice mode to strictly enforce having a single-slice, outside of this series.
One reason for that would be to gather multiple slices (up to what amounts to a frame) for efficient use of OUTPUT buffers and to avoid allocating a huge number of small ones (which is certainly quite inefficient due to CMA alignment).
Apparently, passing multiple slice_params controls is an issue so per-frame mode currently has to rely on the hardware being able to parse the slice header on its own. The issue exists in the V4L2 API, where we need to know in advance the maximum number of slices we want to use to make the control an array and copy all the controls with each request, even if only a single-one is used.
Maybe one way to solve that would be to use multiple requests with the same OUTPUT buffer, one per slice, so that we only have to pass a single slice_params control per-request. And the first slice of the frame would get the other controls too, while the others wouldn't (assuming we can trust that controls won't change in-between and I'm not sure how true that is). Worst case, all the controls have to be attached with each request, which is maybe still better than passing a big number of unused slice_params controls each time.
Again, I believe the current proposal is good to go for now as it allows the hantro driver to get-in staging, which is the priority.
But the issue I'm discussing should certainly be fixed before unstaging the API.
Quoted 34 lines
Documentation/media/uapi/v4l/ext-ctrls-codec.rst +27
-0* - __u16 - ``nal_ref_idc`` - NAL reference ID value coming from the NAL Unit header@@ -2021,6 +2029,43 @@ enum v4l2_mpeg_video_h264_hierarchical_coding_type - - 0x00000004 - The DPB entry is a long term reference frame +``V4L2_CID_MPEG_VIDEO_H264_DECODING_MODE (enum)`` + Specifies the decoding mode to use. Currently exposes slice-based and + frame-based decoding but new modes might be added later on. + This control is used to complement V4L2_PIX_FMT_H264_SLICE + pixel format. Applications that support V4L2_PIX_FMT_H264_SLICE + are required to set this control in order to specify the decoding mode + that is expected for the buffer. + Drivers may expose a single or multiple decoding modes, depending + on what they can support. + + .. note:: + + This menu control is not yet part of the public kernel API and + it is expected to change. + +.. c:type:: v4l2_mpeg_video_h264_decoding_mode + +.. cssclass:: longtable + +.. flat-table:: + :header-rows: 0 + :stub-columns: 0 + :widths: 1 1 2 + + * - ``V4L2_MPEG_VIDEO_H264_SLICE_BASED_DECODING`` + - 0 + - The decoding is done at the slice granularity.
Documentation/media/uapi/v4l/ext-ctrls-codec.rst enum v4l2_mpeg_video_h264_hierarchical_coding_type - · line 2068
Quoted 10 lines
Documentation/media/uapi/v4l/ext-ctrls-codec.rst +10
-0+ v4l2_ctrl_h264_decode_params->num_slices should be set to 1. + The OUTPUT buffer must contain a single slice. + * - ``V4L2_MPEG_VIDEO_H264_FRAME_BASED_DECODING`` + - 1 + - The decoding is done at the frame granularity. + v4l2_ctrl_h264_decode_params->num_slices should be set to the number of + slices forming a frame. + The OUTPUT buffer must contain all slices needed to decode the + frame. The OUTPUT buffer must also contain both fields. +
Nitpick/suggestion here: could we use the name of the enum as a prefix for the modes, like:
- V4L2_MPEG_VIDEO_H264_DECODING_MODE_PER_SLICE
- V4L2_MPEG_VIDEO_H264_DECODING_MODE_PER_FRAME
or
- V4L2_MPEG_VIDEO_H264_DECODING_MODE_SLICE_BASED
- V4L2_MPEG_VIDEO_H264_DECODING_MODE_FRAME_BASED
I personally find it more readable this way, since the same prefix is kept.
Cheers,
Paul
Quoted 97 lines
Documentation/media/uapi/v4l/ext-ctrls-codec.rst +0
-0.. _v4l2-mpeg-mpeg2: ``V4L2_CID_MPEG_VIDEO_MPEG2_SLICE_PARAMS (struct)``Documentation/media/uapi/v4l/pixfmt-compressed.rst +2
-1@@ -61,7 +61,8 @@ Compressed Formats stateless video decoders that implement an H264 pipeline (using the :ref:`mem2mem` and :ref:`media-request-api`). Metadata associated with the frame to decode are required to - be passed through the ``V4L2_CID_MPEG_VIDEO_H264_SPS``, + be passed through the ``V4L2_CID_MPEG_VIDEO_H264_DECODING_MODE``, + ``V4L2_CID_MPEG_VIDEO_H264_SPS``, ``V4L2_CID_MPEG_VIDEO_H264_PPS``, ``V4L2_CID_MPEG_VIDEO_H264_SCALING_MATRIX``, ``V4L2_CID_MPEG_VIDEO_H264_SLICE_PARAMS`` anddrivers/media/v4l2-core/v4l2-ctrls.c +9
-0@@ -402,6 +402,11 @@ const char * const *v4l2_ctrl_get_menu(u32 id)"Explicit" , NULL, }; +static const char *const h264_decoding_mode[] = { +"Slice-Based" , +"Frame-Based" , +NULL , + };static const char *const mpeg_mpeg2_level[] = {"Low" , "Main",@@ -633,6 +638,8 @@ const char * const *v4l2_ctrl_get_menu(u32 id)return h264_fp_arrangement_type;case V4L2_CID_MPEG_VIDEO_H264_FMO_MAP_TYPE :return h264_fmo_map_type; +case V4L2_CID_MPEG_VIDEO_H264_DECODING_MODE : +return h264_decoding_mode;case V4L2_CID_MPEG_VIDEO_MPEG2_LEVEL :return mpeg_mpeg2_level;case V4L2_CID_MPEG_VIDEO_MPEG2_PROFILE :@@ -852,6 +859,7 @@ const char *v4l2_ctrl_get_name(u32 id)case V4L2_CID_MPEG_VIDEO_H264_SCALING_MATRIX :return "H264 Scaling Matrix" ;case V4L2_CID_MPEG_VIDEO_H264_SLICE_PARAMS :return "H264 Slice Parameters" ;case V4L2_CID_MPEG_VIDEO_H264_DECODE_PARAMS :return "H264 Decode Parameters" ; +case V4L2_CID_MPEG_VIDEO_H264_DECODING_MODE :return "H264 Decoding Mode" ;case V4L2_CID_MPEG_VIDEO_MPEG2_LEVEL :return "MPEG2 Level" ;case V4L2_CID_MPEG_VIDEO_MPEG2_PROFILE :return "MPEG2 Profile" ;case V4L2_CID_MPEG_VIDEO_MPEG4_I_FRAME_QP :return "MPEG4 I-Frame QP Value" ;@@ -1220,6 +1228,7 @@ void v4l2_ctrl_fill(u32 id, const char **name, enum v4l2_ctrl_type *type,case V4L2_CID_MPEG_VIDEO_H264_VUI_SAR_IDC :case V4L2_CID_MPEG_VIDEO_H264_SEI_FP_ARRANGEMENT_TYPE :case V4L2_CID_MPEG_VIDEO_H264_FMO_MAP_TYPE : +case V4L2_CID_MPEG_VIDEO_H264_DECODING_MODE :case V4L2_CID_MPEG_VIDEO_MPEG2_LEVEL :case V4L2_CID_MPEG_VIDEO_MPEG2_PROFILE :case V4L2_CID_MPEG_VIDEO_MPEG4_LEVEL :include/media/h264-ctrls.h +11
-0@@ -26,6 +26,7 @@#define V4L2_CID_MPEG_VIDEO_H264_SCALING_MATRIX (V4L2_CID_MPEG_BASE+1002)#define V4L2_CID_MPEG_VIDEO_H264_SLICE_PARAMS (V4L2_CID_MPEG_BASE+1003)#define V4L2_CID_MPEG_VIDEO_H264_DECODE_PARAMS (V4L2_CID_MPEG_BASE+1004) +#define V4L2_CID_MPEG_VIDEO_H264_DECODING_MODE (V4L2_CID_MPEG_BASE+1005)/* enum v4l2_ctrl_type type values */ #define V4L2_CTRL_TYPE_H264_SPS 0x0110@@ -33,6 +34,12 @@#define V4L2_CTRL_TYPE_H264_SCALING_MATRIX 0x0112#define V4L2_CTRL_TYPE_H264_SLICE_PARAMS 0x0113#define V4L2_CTRL_TYPE_H264_DECODE_PARAMS 0x0114 +#define V4L2_CTRL_TYPE_H264_DECODING_MODE 0x0115 + +enum v4l2_mpeg_video_h264_decoding_mode { +V4L2_MPEG_VIDEO_H264_SLICE_BASED_DECODING , +V4L2_MPEG_VIDEO_H264_FRAME_BASED_DECODING , +};#define V4L2_H264_SPS_CONSTRAINT_SET0_FLAG 0x01#define V4L2_H264_SPS_CONSTRAINT_SET1_FLAG 0x02@@ -125,6 +132,10 @@ struct v4l2_h264_pred_weight_table {struct v4l2_ctrl_h264_slice_params {/* Size in bytes, including header */ __u32 size; + +/* Offset in bytes to the start of slice in the OUTPUT buffer. */ +__u32 start_byte_offset; +/* Offset in bits to slice_data() from the beginning of this slice. */ __u32 header_bit_size;Signature · 1 line
2.22.0
Signature · 3 lines
Paul Kocialkowski, Bootlin Embedded Linux and kernel engineering https://bootlin.com
Re: [PATCH v5 03/11] media: uapi: h264: Add the concept of decoding mode
Quoted 49 lines · Paul Kocialkowski
Hi,
Quoted 34 lines · Ezequiel Garcia
From: Boris Brezillon <[email protected]>
Some stateless decoders don't support per-slice decoding granularity (or at least not in a way that would make them efficient or easy to use).
Expose a menu to control the supported decoding modes. Drivers are allowed to support only one decoding but they can support both too.
Signed-off-by Boris Brezillon <[email protected]>Reviewed-by Paul Kocialkowski <[email protected]>Tested-by Philipp Zabel <[email protected]>---
Changes in v5:
* Improve specification as suggested by Hans. Changes in v4:
* Typos/rewording fixes
Changes in v3:
* s/per-{slice,frame} decoding/{slice,frame}-based decoding/
* Add Paul's R-b
Changes in v2:
* Allow decoding multiple slices in per-slice decoding mode
* Minor doc improvement/fixes--- .../media/uapi/v4l/ext-ctrls-codec.rst | 47 ++++++++++++++++++- .../media/uapi/v4l/pixfmt-compressed.rst | 3 +- drivers/media/v4l2-core/v4l2-ctrls.c | 9 ++++ include/media/h264-ctrls.h | 11 +++++ 4 files changed, 68 insertions(+), 2 deletions(-)Documentation/media/uapi/v4l/ext-ctrls-codec.rst +0
-0Documentation/media/uapi/v4l/ext-ctrls-codec.rst enum v4l2_mpeg_video_h264_hierarchical_coding_type - · line 1754
Quoted 8 lines
Documentation/media/uapi/v4l/ext-ctrls-codec.rst +5
-0@@ -1747,6 +1747,11 @@ enum v4l2_mpeg_video_h264_hierarchical_coding_type - * - __u32 - ``size`` - + * - __u32 + - ``start_byte_offset`` + - Where the slice payload starts in the output buffer. Useful when the + OUTPUT buffer contains more than one slice (some codecs need to know + where each slice starts in this buffer).I think there is a possibility for misunderstanding here: does the "slice payload" include the start code when it (annex-b) is selected?
Quoted 5 lines
I'd suspect that the hardware needs to know where the start code begings rather than where the NAL unit starts when there is a start code. So I'd suggest specifying it as the offset to the beginning of the start code (if present) or to the slice NAL unit start otherwise.
Yes, start_byte_offset is the offset in bytes from the beginning of the OUTPUT buffer to the start of this slice (which may start with a start code as per the start code control).
I guess the field could be named simply "offset", but at the same time, this is bikeshedding.
Quoted 4 lines
The clarification should probably also concern
header_bit_size, to make it clear where it is counted from. I think it makes sense to count it starting from thestart_byte_offsetposition.
The header_bit_size is currently counted from the start of slice. It seems this needs to be better documented, but not necessarily as part of this patch.
Quoted 8 lines
Quoted 8 lines
Documentation/media/uapi/v4l/ext-ctrls-codec.rst +0
-1* - __u32 - ``header_bit_size`` -@@ -1930,7 +1935,10 @@ enum v4l2_mpeg_video_h264_hierarchical_coding_type - - * - __u16 - ``num_slices`` - - Number of slices needed to decode the current frame
Quoted 9 lines
Documentation/media/uapi/v4l/ext-ctrls-codec.rst enum v4l2_mpeg_video_h264_hierarchical_coding_type - · line 1941
Quoted 4 lines
Documentation/media/uapi/v4l/ext-ctrls-codec.rst +4
-0+ - Number of slices needed to decode the current frame/field. When + operating in slice-based decoding mode (see + :c:type:`v4l2_mpeg_video_h264_decoding_mode`), this field + should always be set to one.I am fine with the current proposal for now, but I believe that we should keep discussing whether we want per-slice mode to strictly enforce having a single-slice, outside of this series.
Sounds good.
Quoted 19 lines
One reason for that would be to gather multiple slices (up to what amounts to a frame) for efficient use of OUTPUT buffers and to avoid allocating a huge number of small ones (which is certainly quite inefficient due to CMA alignment).
Apparently, passing multiple
slice_paramscontrols is an issue so per-frame mode currently has to rely on the hardware being able to parse the slice header on its own. The issue exists in the V4L2 API, where we need to know in advance the maximum number of slices we want to use to make the control an array and copy all the controls with each request, even if only a single-one is used.Maybe one way to solve that would be to use multiple requests with the same OUTPUT buffer, one per slice, so that we only have to pass a single
slice_paramscontrol per-request. And the first slice of the frame would get the other controls too, while the others wouldn't (assuming we can trust that controls won't change in-between and I'm not sure how true that is). Worst case, all the controls have to be attached with each request, which is maybe still better than passing a big number of unusedslice_paramscontrols each time.
Quoted 5 lines
Again, I believe the current proposal is good to go for now as it allows the hantro driver to get-in staging, which is the priority.
But the issue I'm discussing should certainly be fixed before unstaging the API.
Totally, I think the next step is to add multi-slice support to the cedrus driver. It will help consolidate the API.
Quoted 50 lines
Quoted 34 lines
Documentation/media/uapi/v4l/ext-ctrls-codec.rst +27
-0* - __u16 - ``nal_ref_idc`` - NAL reference ID value coming from the NAL Unit header@@ -2021,6 +2029,43 @@ enum v4l2_mpeg_video_h264_hierarchical_coding_type - - 0x00000004 - The DPB entry is a long term reference frame +``V4L2_CID_MPEG_VIDEO_H264_DECODING_MODE (enum)`` + Specifies the decoding mode to use. Currently exposes slice-based and + frame-based decoding but new modes might be added later on. + This control is used to complement V4L2_PIX_FMT_H264_SLICE + pixel format. Applications that support V4L2_PIX_FMT_H264_SLICE + are required to set this control in order to specify the decoding mode + that is expected for the buffer. + Drivers may expose a single or multiple decoding modes, depending + on what they can support. + + .. note:: + + This menu control is not yet part of the public kernel API and + it is expected to change. + +.. c:type:: v4l2_mpeg_video_h264_decoding_mode + +.. cssclass:: longtable + +.. flat-table:: + :header-rows: 0 + :stub-columns: 0 + :widths: 1 1 2 + + * - ``V4L2_MPEG_VIDEO_H264_SLICE_BASED_DECODING`` + - 0 + - The decoding is done at the slice granularity.Documentation/media/uapi/v4l/ext-ctrls-codec.rst enum v4l2_mpeg_video_h264_hierarchical_coding_type - · line 2068
Quoted 10 lines
Documentation/media/uapi/v4l/ext-ctrls-codec.rst +10
-0+ v4l2_ctrl_h264_decode_params->num_slices should be set to 1. + The OUTPUT buffer must contain a single slice. + * - ``V4L2_MPEG_VIDEO_H264_FRAME_BASED_DECODING`` + - 1 + - The decoding is done at the frame granularity. + v4l2_ctrl_h264_decode_params->num_slices should be set to the number of + slices forming a frame. + The OUTPUT buffer must contain all slices needed to decode the + frame. The OUTPUT buffer must also contain both fields. +Nitpick/suggestion here: could we use the name of the enum as a prefix for the modes, like:
-V4L2_MPEG_VIDEO_H264_DECODING_MODE_PER_SLICE
-V4L2_MPEG_VIDEO_H264_DECODING_MODE_PER_FRAME
Quoted 6 lines
or
-V4L2_MPEG_VIDEO_H264_DECODING_MODE_SLICE_BASED
-V4L2_MPEG_VIDEO_H264_DECODING_MODE_FRAME_BASEDI personally find it more readable this way, since the same prefix is kept.
Yes, that looks better.
Thanks for reviewing,
Ezequiel
PATCH v5 4/11 media: uapi: h264: Add the concept of start code
In reply to Ezequiel GarciaStateless decoders have different expectations about the start code that is prepended on H264 slices. Add a menu control to express the supported start code types (including no start code).
Drivers are allowed to support only one start code type, but they can support both too.
Note that this is independent of the H264 decoding mode, which specifies the granularity of the decoding operations. Either in frame-based or slice-based mode, this new control will allow to define the start code expected on H264 slices.
---
Changes in v5:
* Improve specification as suggested by Hans.
Changes in v4:
* New patch.
--- .../media/uapi/v4l/ext-ctrls-codec.rst | 33 +++++++++++++++++++ .../media/uapi/v4l/pixfmt-compressed.rst | 3 +- drivers/media/v4l2-core/v4l2-ctrls.c | 9 +++++ include/media/h264-ctrls.h | 6 ++++ 4 files changed, 50 insertions(+), 1 deletion(-)
Documentation/media/uapi/v4l/ext-ctrls-codec.rst +33 -0
@@ -2066,6 +2066,39 @@ enum v4l2_mpeg_video_h264_hierarchical_coding_type - The OUTPUT buffer must contain all slices needed to decode the frame. The OUTPUT buffer must also contain both fields. +``V4L2_CID_MPEG_VIDEO_H264_STARTCODE (enum)`` + Specifies the H264 slice start code expected for each slice. + This control is used to complement V4L2_PIX_FMT_H264_SLICE + pixel format. Applications that support V4L2_PIX_FMT_H264_SLICE + are required to set this control in order to specify the start code + that is expected for the buffer. + Drivers may expose a single or multiple start codes, depending + on what they can support. + + .. note:: + + This menu control is not yet part of the public kernel API and + it is expected to change. + +.. c:type:: v4l2_mpeg_video_h264_startcode + +.. cssclass:: longtable + +.. flat-table:: + :header-rows: 0 + :stub-columns: 0 + :widths: 1 1 2 + + * - ``V4L2_MPEG_VIDEO_H264_NO_STARTCODE`` + - 0 + - Selecting this value specifies that H264 slices are passed + to the driver without any start code. + * - ``V4L2_MPEG_VIDEO_H264_ANNEX_B_STARTCODE`` + - 1 + - Selecting this value specifies that H264 slices are expected + to be prefixed by Annex B start codes. According to :ref:`h264` + valid start codes can be 3-bytes 0x000001, or 4-bytes 0x00000001. + .. _v4l2-mpeg-mpeg2: ``V4L2_CID_MPEG_VIDEO_MPEG2_SLICE_PARAMS (struct)``
Documentation/media/uapi/v4l/pixfmt-compressed.rst +2 -1
@@ -61,7 +61,8 @@ Compressed Formats stateless video decoders that implement an H264 pipeline (using the :ref:`mem2mem` and :ref:`media-request-api`). Metadata associated with the frame to decode are required to - be passed through the ``V4L2_CID_MPEG_VIDEO_H264_DECODING_MODE``, + be passed through the ``V4L2_CID_MPEG_VIDEO_H264_STARTCODE``, + ``V4L2_CID_MPEG_VIDEO_H264_DECODING_MODE``, ``V4L2_CID_MPEG_VIDEO_H264_SPS``, ``V4L2_CID_MPEG_VIDEO_H264_PPS``, ``V4L2_CID_MPEG_VIDEO_H264_SCALING_MATRIX``,
drivers/media/v4l2-core/v4l2-ctrls.c +9 -0
@@ -407,6 +407,11 @@ const char * const *v4l2_ctrl_get_menu(u32 id)"Frame-Based" , NULL, }; +static const char *const h264_start_code[] = { +"No Start Code" , +"Annex B Start Code" , +NULL , + };static const char *const mpeg_mpeg2_level[] = {"Low" , "Main",
@@ -640,6 +645,8 @@ const char * const *v4l2_ctrl_get_menu(u32 id)return h264_fmo_map_type;case V4L2_CID_MPEG_VIDEO_H264_DECODING_MODE :return h264_decoding_mode; +case V4L2_CID_MPEG_VIDEO_H264_STARTCODE : +return h264_start_code;case V4L2_CID_MPEG_VIDEO_MPEG2_LEVEL :return mpeg_mpeg2_level;case V4L2_CID_MPEG_VIDEO_MPEG2_PROFILE :
@@ -860,6 +867,7 @@ const char *v4l2_ctrl_get_name(u32 id)case V4L2_CID_MPEG_VIDEO_H264_SLICE_PARAMS :return "H264 Slice Parameters" ;case V4L2_CID_MPEG_VIDEO_H264_DECODE_PARAMS :return "H264 Decode Parameters" ;case V4L2_CID_MPEG_VIDEO_H264_DECODING_MODE :return "H264 Decoding Mode" ; +case V4L2_CID_MPEG_VIDEO_H264_STARTCODE :return "H264 Start Code" ;case V4L2_CID_MPEG_VIDEO_MPEG2_LEVEL :return "MPEG2 Level" ;case V4L2_CID_MPEG_VIDEO_MPEG2_PROFILE :return "MPEG2 Profile" ;case V4L2_CID_MPEG_VIDEO_MPEG4_I_FRAME_QP :return "MPEG4 I-Frame QP Value" ;
@@ -1229,6 +1237,7 @@ void v4l2_ctrl_fill(u32 id, const char **name, enum v4l2_ctrl_type *type,case V4L2_CID_MPEG_VIDEO_H264_SEI_FP_ARRANGEMENT_TYPE :case V4L2_CID_MPEG_VIDEO_H264_FMO_MAP_TYPE :case V4L2_CID_MPEG_VIDEO_H264_DECODING_MODE : +case V4L2_CID_MPEG_VIDEO_H264_STARTCODE :case V4L2_CID_MPEG_VIDEO_MPEG2_LEVEL :case V4L2_CID_MPEG_VIDEO_MPEG2_PROFILE :case V4L2_CID_MPEG_VIDEO_MPEG4_LEVEL :
include/media/h264-ctrls.h +6 -0
@@ -27,6 +27,7 @@#define V4L2_CID_MPEG_VIDEO_H264_SLICE_PARAMS (V4L2_CID_MPEG_BASE+1003)#define V4L2_CID_MPEG_VIDEO_H264_DECODE_PARAMS (V4L2_CID_MPEG_BASE+1004)#define V4L2_CID_MPEG_VIDEO_H264_DECODING_MODE (V4L2_CID_MPEG_BASE+1005) +#define V4L2_CID_MPEG_VIDEO_H264_STARTCODE (V4L2_CID_MPEG_BASE+1006)/* enum v4l2_ctrl_type type values */ #define V4L2_CTRL_TYPE_H264_SPS 0x0110
@@ -41,6 +42,11 @@ enum v4l2_mpeg_video_h264_decoding_mode { V4L2_MPEG_VIDEO_H264_FRAME_BASED_DECODING, }; +enum v4l2_mpeg_video_h264_start_code { +V4L2_MPEG_VIDEO_H264_NO_STARTCODE , +V4L2_MPEG_VIDEO_H264_ANNEX_B_STARTCODE , +}; +#define V4L2_H264_SPS_CONSTRAINT_SET0_FLAG 0x01#define V4L2_H264_SPS_CONSTRAINT_SET1_FLAG 0x02#define V4L2_H264_SPS_CONSTRAINT_SET2_FLAG 0x04
Signature · 1 line
2.22.0
Re: [PATCH v5 04/11] media: uapi: h264: Add the concept of start code
Quoted 27 lines · Ezequiel Garcia
Stateless decoders have different expectations about the start code that is prepended on H264 slices. Add a menu control to express the supported start code types (including no start code).
Drivers are allowed to support only one start code type, but they can support both too.
Note that this is independent of the H264 decoding mode, which specifies the granularity of the decoding operations. Either in frame-based or slice-based mode, this new control will allow to define the start code expected on H264 slices.
Signed-off-by Ezequiel Garcia <[email protected]>Tested-by Philipp Zabel <[email protected]>---
Changes in v5:
* Improve specification as suggested by Hans.
Changes in v4:
* New patch.--- .../media/uapi/v4l/ext-ctrls-codec.rst | 33 +++++++++++++++++++ .../media/uapi/v4l/pixfmt-compressed.rst | 3 +- drivers/media/v4l2-core/v4l2-ctrls.c | 9 +++++ include/media/h264-ctrls.h | 6 ++++ 4 files changed, 50 insertions(+), 1 deletion(-)
<snip>
include/media/h264-ctrls.h line 30
Quoted 9 lines
include/media/h264-ctrls.h +1
-0@@ -27,6 +27,7 @@#define V4L2_CID_MPEG_VIDEO_H264_SLICE_PARAMS (V4L2_CID_MPEG_BASE+1003)#define V4L2_CID_MPEG_VIDEO_H264_DECODE_PARAMS (V4L2_CID_MPEG_BASE+1004)#define V4L2_CID_MPEG_VIDEO_H264_DECODING_MODE (V4L2_CID_MPEG_BASE+1005) +#define V4L2_CID_MPEG_VIDEO_H264_STARTCODE (V4L2_CID_MPEG_BASE+1006)
I almost forgot: can you change this to _START_CODE? Since it is two words?
Thanks!
Hans
Quoted 16 lines
include/media/h264-ctrls.h +5
-0/* enum v4l2_ctrl_type type values */ #define V4L2_CTRL_TYPE_H264_SPS 0x0110@@ -41,6 +42,11 @@ enum v4l2_mpeg_video_h264_decoding_mode { V4L2_MPEG_VIDEO_H264_FRAME_BASED_DECODING, }; +enum v4l2_mpeg_video_h264_start_code { +V4L2_MPEG_VIDEO_H264_NO_STARTCODE , +V4L2_MPEG_VIDEO_H264_ANNEX_B_STARTCODE , +}; +#define V4L2_H264_SPS_CONSTRAINT_SET0_FLAG 0x01#define V4L2_H264_SPS_CONSTRAINT_SET1_FLAG 0x02#define V4L2_H264_SPS_CONSTRAINT_SET2_FLAG 0x04
Re: [PATCH v5 04/11] media: uapi: h264: Add the concept of start code
Hi,
Quoted 31 lines · Hans Verkuil
Quoted 27 lines · Ezequiel Garcia
Stateless decoders have different expectations about the start code that is prepended on H264 slices. Add a menu control to express the supported start code types (including no start code).
Drivers are allowed to support only one start code type, but they can support both too.
Note that this is independent of the H264 decoding mode, which specifies the granularity of the decoding operations. Either in frame-based or slice-based mode, this new control will allow to define the start code expected on H264 slices.
Signed-off-by Ezequiel Garcia <[email protected]>Tested-by Philipp Zabel <[email protected]>---
Changes in v5:
* Improve specification as suggested by Hans.
Changes in v4:
* New patch.--- .../media/uapi/v4l/ext-ctrls-codec.rst | 33 +++++++++++++++++++ .../media/uapi/v4l/pixfmt-compressed.rst | 3 +- drivers/media/v4l2-core/v4l2-ctrls.c | 9 +++++ include/media/h264-ctrls.h | 6 ++++ 4 files changed, 50 insertions(+), 1 deletion(-)<snip>
Quoted 11 lines
include/media/h264-ctrls.h line 30
Quoted 9 lines
include/media/h264-ctrls.h +1
-0@@ -27,6 +27,7 @@#define V4L2_CID_MPEG_VIDEO_H264_SLICE_PARAMS (V4L2_CID_MPEG_BASE+1003)#define V4L2_CID_MPEG_VIDEO_H264_DECODE_PARAMS (V4L2_CID_MPEG_BASE+1004)#define V4L2_CID_MPEG_VIDEO_H264_DECODING_MODE (V4L2_CID_MPEG_BASE+1005) +#define V4L2_CID_MPEG_VIDEO_H264_STARTCODE (V4L2_CID_MPEG_BASE+1006)I almost forgot: can you change this to
_START_CODE? Since it is two words?
Agreed, I like it better this way too.
Quoted 11 lines
Thanks!
Hans
Quoted 7 lines
include/media/h264-ctrls.h +0
-0/* enum v4l2_ctrl_type type values */ #define V4L2_CTRL_TYPE_H264_SPS 0x0110@@ -41,6 +42,11 @@ enum v4l2_mpeg_video_h264_decoding_mode { V4L2_MPEG_VIDEO_H264_FRAME_BASED_DECODING, };
Quoted 3 lines
include/media/h264-ctrls.h enum v4l2_mpeg_video_h264_decoding_mode { · line 47
Quoted 3 lines
include/media/h264-ctrls.h +3
-0+enum v4l2_mpeg_video_h264_start_code { +V4L2_MPEG_VIDEO_H264_NO_STARTCODE , +V4L2_MPEG_VIDEO_H264_ANNEX_B_STARTCODE ,
Could we apply the same START_CODE renaming here too?
I was also thinking that it would be slightly more readable put like this, with START_CODE as a prefix since it's common to both options and the name of the enum:
- V4L2_MPEG_VIDEO_H264_START_CODE_NONE
- V4L2_MPEG_VIDEO_H264_START_CODE_ANNEX_B
What do you think?
Cheers,
Paul
Quoted 7 lines
Quoted 6 lines
include/media/h264-ctrls.h +2
-0+}; +#define V4L2_H264_SPS_CONSTRAINT_SET0_FLAG 0x01#define V4L2_H264_SPS_CONSTRAINT_SET1_FLAG 0x02#define V4L2_H264_SPS_CONSTRAINT_SET2_FLAG 0x04
Signature · 3 lines
Paul Kocialkowski, Bootlin Embedded Linux and kernel engineering https://bootlin.com
Re: [PATCH v5 04/11] media: uapi: h264: Add the concept of start code
Quoted 46 lines · Paul Kocialkowski
Hi,
Quoted 31 lines · Hans Verkuil
Quoted 27 lines · Ezequiel Garcia
Stateless decoders have different expectations about the start code that is prepended on H264 slices. Add a menu control to express the supported start code types (including no start code).
Drivers are allowed to support only one start code type, but they can support both too.
Note that this is independent of the H264 decoding mode, which specifies the granularity of the decoding operations. Either in frame-based or slice-based mode, this new control will allow to define the start code expected on H264 slices.
Signed-off-by Ezequiel Garcia <[email protected]>Tested-by Philipp Zabel <[email protected]>---
Changes in v5:
* Improve specification as suggested by Hans.
Changes in v4:
* New patch.--- .../media/uapi/v4l/ext-ctrls-codec.rst | 33 +++++++++++++++++++ .../media/uapi/v4l/pixfmt-compressed.rst | 3 +- drivers/media/v4l2-core/v4l2-ctrls.c | 9 +++++ include/media/h264-ctrls.h | 6 ++++ 4 files changed, 50 insertions(+), 1 deletion(-)<snip>
Quoted 11 lines
include/media/h264-ctrls.h line 30
Quoted 9 lines
include/media/h264-ctrls.h +1
-0@@ -27,6 +27,7 @@#define V4L2_CID_MPEG_VIDEO_H264_SLICE_PARAMS (V4L2_CID_MPEG_BASE+1003)#define V4L2_CID_MPEG_VIDEO_H264_DECODE_PARAMS (V4L2_CID_MPEG_BASE+1004)#define V4L2_CID_MPEG_VIDEO_H264_DECODING_MODE (V4L2_CID_MPEG_BASE+1005) +#define V4L2_CID_MPEG_VIDEO_H264_STARTCODE (V4L2_CID_MPEG_BASE+1006)I almost forgot: can you change this to
_START_CODE? Since it is two words?
Quoted 2 lines
Agreed, I like it better this way too.
The reason to have STARTCODE instead of START_CODE was to have some consistency with other controls (namely, V4L2_CID_MPEG_VIDEO_HEVC_WITHOUT_STARTCODE).
Consistency is really important in a API,
but at the same time, I agree START_CODE looks better.
Quoted 17 lines
Quoted 11 lines
Thanks!
Hans
Quoted 7 lines
include/media/h264-ctrls.h +0
-0/* enum v4l2_ctrl_type type values */ #define V4L2_CTRL_TYPE_H264_SPS 0x0110@@ -41,6 +42,11 @@ enum v4l2_mpeg_video_h264_decoding_mode { V4L2_MPEG_VIDEO_H264_FRAME_BASED_DECODING, };Quoted 3 lines
include/media/h264-ctrls.h enum v4l2_mpeg_video_h264_decoding_mode { · line 47
Quoted 3 lines
include/media/h264-ctrls.h +3
-0+enum v4l2_mpeg_video_h264_start_code { +V4L2_MPEG_VIDEO_H264_NO_STARTCODE , +V4L2_MPEG_VIDEO_H264_ANNEX_B_STARTCODE ,Could we apply the same
START_CODErenaming here too?
Quoted 7 lines
I was also thinking that it would be slightly more readable put like this, with
START_CODEas a prefix since it's common to both options and the name of the enum:-
V4L2_MPEG_VIDEO_H264_START_CODE_NONE
-V4L2_MPEG_VIDEO_H264_START_CODE_ANNEX_B
Yes, that looks much better.
Thanks,
Ezequiel
PATCH v5 5/11 media: uapi: h264: Get rid of the p0/b0/b1 ref-lists
In reply to Ezequiel GarciaFrom: Boris Brezillon <[email protected]>
Those lists can be extracted from the dpb, let's simplify userspace life and build that list kernel-side (generic helpers will be provided for drivers that need this list).
---
Changes in v5:
* New patch.
Changes in v4:
* Add R-b
* Remove the reflist from the struct
Changes in v3:
* None
Changes in v2:
* None
--- Documentation/media/uapi/v4l/ext-ctrls-codec.rst | 9 --------- include/media/h264-ctrls.h | 3 --- 2 files changed, 12 deletions(-)
Documentation/media/uapi/v4l/ext-ctrls-codec.rst +0 -9
@@ -1942,15 +1942,6 @@ enum v4l2_mpeg_video_h264_hierarchical_coding_type - * - __u16 - ``nal_ref_idc`` - NAL reference ID value coming from the NAL Unit header - * - __u8 - - ``ref_pic_list_p0[32]`` - - Backward reference list used by P-frames in the original bitstream order - * - __u8 - - ``ref_pic_list_b0[32]`` - - Backward reference list used by B-frames in the original bitstream order - * - __u8 - - ``ref_pic_list_b1[32]`` - - Forward reference list used by B-frames in the original bitstream order * - __s32 - ``top_field_order_cnt`` - Picture Order Count for the coded top field
include/media/h264-ctrls.h +0 -3
@@ -203,9 +203,6 @@ struct v4l2_ctrl_h264_decode_params {struct v4l2_h264_dpb_entry dpb[16 ];__u16 num_slices;__u16 nal_ref_idc; -__u8 ref_pic_list_p0[32 ]; -__u8 ref_pic_list_b0[32 ]; -__u8 ref_pic_list_b1[32 ];__s32 top_field_order_cnt;__s32 bottom_field_order_cnt;__u32 flags;/* V4L2_H264_DECODE_PARAM_FLAG_* */
Signature · 1 line
2.22.0
PATCH v5 6/11 media: cedrus: Cleanup control initialization
In reply to Ezequiel GarciaIn order to introduce other controls, the control initialization needs to support an initial struct v4l2_ctrl_control.
While here, let's cleanup the control initialization, removing unneeded fields.
---
Changes in v5:
* None.
Changes in v4:
* New patch.
--- drivers/staging/media/sunxi/cedrus/cedrus.c | 45 +++++++++++---------- drivers/staging/media/sunxi/cedrus/cedrus.h | 3 +- 2 files changed, 25 insertions(+), 23 deletions(-)
drivers/staging/media/sunxi/cedrus/cedrus.c +24 -21
@@ -29,44 +29,51 @@static const struct cedrus_control cedrus_controls[] = { { - .id =V4L2_CID_MPEG_VIDEO_MPEG2_SLICE_PARAMS , - .elem_size =sizeof (struct v4l2_ctrl_mpeg2_slice_params ), + .cfg = { + .id =V4L2_CID_MPEG_VIDEO_MPEG2_SLICE_PARAMS , + }, .codec =CEDRUS_CODEC_MPEG2 , .required = true, }, { - .id =V4L2_CID_MPEG_VIDEO_MPEG2_QUANTIZATION , - .elem_size =sizeof (struct v4l2_ctrl_mpeg2_quantization ), + .cfg = { + .id =V4L2_CID_MPEG_VIDEO_MPEG2_QUANTIZATION , + }, .codec =CEDRUS_CODEC_MPEG2 , .required = false, }, { - .id =V4L2_CID_MPEG_VIDEO_H264_DECODE_PARAMS , - .elem_size =sizeof (struct v4l2_ctrl_h264_decode_params ), + .cfg = { + .id =V4L2_CID_MPEG_VIDEO_H264_DECODE_PARAMS , + }, .codec =CEDRUS_CODEC_H264 , .required = true, }, { - .id =V4L2_CID_MPEG_VIDEO_H264_SLICE_PARAMS , - .elem_size =sizeof (struct v4l2_ctrl_h264_slice_params ), + .cfg = { + .id =V4L2_CID_MPEG_VIDEO_H264_SLICE_PARAMS , + }, .codec =CEDRUS_CODEC_H264 , .required = true, }, { - .id =V4L2_CID_MPEG_VIDEO_H264_SPS , - .elem_size =sizeof (struct v4l2_ctrl_h264_sps ), + .cfg = { + .id =V4L2_CID_MPEG_VIDEO_H264_SPS , + }, .codec =CEDRUS_CODEC_H264 , .required = true, }, { - .id =V4L2_CID_MPEG_VIDEO_H264_PPS , - .elem_size =sizeof (struct v4l2_ctrl_h264_pps ), + .cfg = { + .id =V4L2_CID_MPEG_VIDEO_H264_PPS , + }, .codec =CEDRUS_CODEC_H264 , .required = true, }, { - .id =V4L2_CID_MPEG_VIDEO_H264_SCALING_MATRIX , - .elem_size =sizeof (struct v4l2_ctrl_h264_scaling_matrix ), + .cfg = { + .id =V4L2_CID_MPEG_VIDEO_H264_SCALING_MATRIX , + }, .codec =CEDRUS_CODEC_H264 , .required = true, },
@@ -106,12 +113,8 @@ static int cedrus_init_ctrls(struct cedrus_dev *dev, struct cedrus_ctx *ctx)return -ENOMEM ;for (i =0 ; i <CEDRUS_CONTROLS_COUNT ; i++) { -struct v4l2_ctrl_config cfg = {}; - - cfg.elem_size = cedrus_controls[i].elem_size; - cfg.id = cedrus_controls[i].id; - - ctrl =v4l2_ctrl_new_custom (hdl, &cfg,NULL ); + ctrl =v4l2_ctrl_new_custom (hdl, &cedrus_controls[i].cfg, +NULL );if (hdl->error) {v4l2_err (&dev->v4l2_dev,"Failed to create new custom control\n" );
@@ -178,7 +181,7 @@ static int cedrus_request_validate(struct media_request *req)continue ; ctrl_test =v4l2_ctrl_request_hdl_ctrl_find (hdl, - cedrus_controls[i].id); + cedrus_controls[i].cfg.id);if (!ctrl_test) {v4l2_info (&ctx->dev->v4l2_dev,"Missing required codec control\n" );
drivers/staging/media/sunxi/cedrus/cedrus.h +1 -2
@@ -49,8 +49,7 @@ enum cedrus_h264_pic_type { };struct cedrus_control { -u32 id; -u32 elem_size; +struct v4l2_ctrl_config cfg;enum cedrus_codec codec;unsigned char required:1 ; };
Signature · 1 line
2.22.0
PATCH v5 7/11 media: cedrus: Specify H264 startcode and decoding mode
In reply to Ezequiel GarciaThe cedrus VPU is slice-based and expects V4L2_PIX_FMT_H264_SLICE buffers to contain H264 slices with no start code.
Expose this to userspace with the newly added menu control.
These two controls are specified as mandatory for applications, but we mark them as non-required on the driver side for backwards compatibility.
---
Changes in v5:
* Clarify commit log.
Changes in v4:
* New patch.
--- drivers/staging/media/sunxi/cedrus/cedrus.c | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+)
drivers/staging/media/sunxi/cedrus/cedrus.c +20 -0
@@ -77,6 +77,26 @@ static const struct cedrus_control cedrus_controls[] = { .codec = CEDRUS_CODEC_H264, .required = true, }, + { + .cfg = { + .id = V4L2_CID_MPEG_VIDEO_H264_DECODING_MODE, + .max = V4L2_MPEG_VIDEO_H264_SLICE_BASED_DECODING, + .def = V4L2_MPEG_VIDEO_H264_SLICE_BASED_DECODING, + .menu_skip_mask = BIT(V4L2_MPEG_VIDEO_H264_FRAME_BASED_DECODING), + }, + .codec = CEDRUS_CODEC_H264, + .required = false, + }, + { + .cfg = { + .id = V4L2_CID_MPEG_VIDEO_H264_STARTCODE, + .max = V4L2_MPEG_VIDEO_H264_NO_STARTCODE, + .def = V4L2_MPEG_VIDEO_H264_NO_STARTCODE, + .menu_skip_mask = BIT(V4L2_MPEG_VIDEO_H264_ANNEX_B_STARTCODE), + }, + .codec = CEDRUS_CODEC_H264, + .required = false, + }, };#define CEDRUS_CONTROLS_COUNT ARRAY_SIZE(cedrus_controls)
Signature · 1 line
2.22.0
PATCH v5 8/11 media: hantro: Move copy_metadata() before doing a decode operation
In reply to Ezequiel GarciaFrom: Boris Brezillon <boris.brezillon-ZGY8ohtN/[email protected]>
Some decoders use intra slice/frame references. The capture buffer pointed by these references might be new and thus have invalid timestamp which prevents the decoder logic from retrieving the vb2_buffer object based on the output buf timestamp.
Copy all metadata (including the timestamp) before starting the decode operation.
---
Changes in v5:
* None.
Changes in v4:
* None.
--- drivers/staging/media/hantro/hantro_drv.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-)
drivers/staging/media/hantro/hantro_drv.c +6 -2
@@ -111,8 +111,6 @@ static void hantro_job_finish(struct hantro_dev *vpu, src->sequence = ctx->sequence_out++; dst->sequence = ctx->sequence_cap++; -v4l2_m2m_buf_copy_metadata (src, dst, true); - ret = ctx->buf_finish (ctx, &dst->vb2_buf, bytesused);if (ret) result =VB2_BUF_STATE_ERROR ;
@@ -178,8 +176,12 @@ void hantro_finish_run(struct hantro_ctx *ctx)static void device_run (void *priv) {struct hantro_ctx *ctx = priv; +struct vb2_v4l2_buffer *src, *dst;int ret; + src =hantro_get_src_buf (ctx); + dst =hantro_get_dst_buf (ctx); + ret =clk_bulk_enable (ctx->dev->variant->num_clocks, ctx->dev->clocks);if (ret) goto err_cancel_job;
@@ -187,6 +189,8 @@ static void device_run(void *priv)if (ret <0 ) goto err_cancel_job; +v4l2_m2m_buf_copy_metadata (src, dst, true); + ctx->codec_ops->run (ctx);return ;
Signature · 1 line
2.22.0
PATCH v5 9/11 media: hantro: Add core bits to support H264 decoding
In reply to Ezequiel GarciaFrom: Hertz Wong <[email protected]>
Add helpers and patch hantro_{drv,v4l2}.c to prepare addition of H264 decoding support.
---
Changes in v5:
* None.
Changes in v4:
* Rework extra_size0, exposing the size via TRY_FMT/S_FMT
to allow buffer importation, as suggested by Tomasz.
* Drop max slice limit.
* Use a ternary operator instead of substracting POCs,
to avoid an overflow as pointed out by Rasmus.
* Specify annex B slice start code.
* Add missing extra_size0 to sizeimage
* Swap the first 2 entries of list B1 when B0 and B1 match (mandated by the spec)
* Move the update_dpb() call before the prepare_table() one to make the POCs stored in the private table match the content of the new DPB
--- drivers/staging/media/hantro/Makefile | 1 + drivers/staging/media/hantro/hantro.h | 9 +- drivers/staging/media/hantro/hantro_drv.c | 42 ++ drivers/staging/media/hantro/hantro_h264.c | 651 +++++++++++++++++++++ drivers/staging/media/hantro/hantro_hw.h | 55 ++ drivers/staging/media/hantro/hantro_v4l2.c | 10 + 6 files changed, 767 insertions(+), 1 deletion(-) create mode 100644 drivers/staging/media/hantro/hantro_h264.c
drivers/staging/media/hantro/Makefile +1 -0
@@ -10,6 +10,7 @@ hantro-vpu-y += \ rk3399_vpu_hw_mpeg2_dec.o \ rk3399_vpu_hw_vp8_dec.o \ hantro_jpeg.o \ + hantro_h264.o \ hantro_mpeg2.o \ hantro_vp8.o
drivers/staging/media/hantro/hantro.h +8 -1
@@ -30,6 +30,10 @@#define VP8_MB_WIDTH (w) DIV_ROUND_UP(w, VP8_MB_DIM)#define VP8_MB_HEIGHT (h) DIV_ROUND_UP(h, VP8_MB_DIM) +#define H264_MB_DIM 16 +#define H264_MB_WIDTH (w) DIV_ROUND_UP(w, H264_MB_DIM) +#define H264_MB_HEIGHT (h) DIV_ROUND_UP(h, H264_MB_DIM) +#define MPEG2_MB_DIM 16#define MPEG2_MB_WIDTH (w) DIV_ROUND_UP(w, MPEG2_MB_DIM)#define MPEG2_MB_HEIGHT (h) DIV_ROUND_UP(h, MPEG2_MB_DIM)
@@ -43,9 +47,9 @@ struct hantro_codec_ops;#define HANTRO_JPEG_ENCODER BIT(0)#define HANTRO_ENCODERS 0x0000ffff -#define HANTRO_MPEG2_DECODER BIT(16)#define HANTRO_VP8_DECODER BIT(17) +#define HANTRO_H264_DECODER BIT(18)#define HANTRO_DECODERS 0xffff0000/**
@@ -102,12 +106,14 @@ struct hantro_variant {* enum hantro_codec_mode - codec operating mode. * @HANTRO_MODE_NONE: No operating mode. Used for RAW video formats. * @HANTRO_MODE_JPEG_ENC: JPEG encoder. +* @HANTRO_MODE_H264_DEC: H264 decoder. * @HANTRO_MODE_MPEG2_DEC: MPEG-2 decoder. * @HANTRO_MODE_VP8_DEC: VP8 decoder. */ enum hantro_codec_mode {HANTRO_MODE_NONE =-1 ,HANTRO_MODE_JPEG_ENC , +HANTRO_MODE_H264_DEC ,HANTRO_MODE_MPEG2_DEC ,HANTRO_MODE_VP8_DEC , };
@@ -246,6 +252,7 @@ struct hantro_ctx {/* Specific for particular codec modes. */ union { +struct hantro_h264_dec_hw_ctx h264_dec;struct hantro_jpeg_enc_hw_ctx jpeg_enc;struct hantro_mpeg2_dec_hw_ctx mpeg2_dec;struct hantro_vp8_dec_hw_ctx vp8_dec;
drivers/staging/media/hantro/hantro_drv.c +42 -0
48 lines not on this page · raw message
drivers/staging/media/hantro/hantro_h264.c new file +651 -0
651 lines not on this page · raw message
drivers/staging/media/hantro/hantro_hw.h +55 -0
73 lines not on this page · raw message
drivers/staging/media/hantro/hantro_v4l2.c +10 -0
@@ -239,6 +239,15 @@ static int vidioc_try_fmt(struct file *file, void *priv, struct v4l2_format *f,/* Fill remaining fields */ v4l2_fill_pixfmt_mp (pix_mp, fmt->fourcc, pix_mp->width, pix_mp->height); +/* +* The H264 decoder needs extra space on the output buffers +* to store motion vectors. This is needed for reference +* frames. +*/ +if (ctx->vpu_src_fmt->fourcc ==V4L2_PIX_FMT_H264_SLICE ) + pix_mp->plane_fmt[0 ].sizeimage += +128 *DIV_ROUND_UP (pix_mp->width,16 ) * +DIV_ROUND_UP (pix_mp->height,16 ); }else if (!pix_mp->plane_fmt[0 ].sizeimage) {/* * For coded formats the application can specify
@@ -345,6 +354,7 @@ hantro_update_requires_request(struct hantro_ctx *ctx, u32 fourcc)break ;case V4L2_PIX_FMT_MPEG2_SLICE :case V4L2_PIX_FMT_VP8_FRAME : +case V4L2_PIX_FMT_H264_SLICE : ctx->fh.m2m_ctx->out_q_ctx.q.requires_requests = true;break ;default :
Signature · 1 line
2.22.0
PATCH v5 10/11 media: hantro: Add support for H264 decoding on G1
In reply to Ezequiel GarciaFrom: Hertz Wong <[email protected]>
Add the G1 specific bits to support H264 decoding.
---
Changes from v5:
* None.
Changes from v4:
* Update to use pic_size from the H264 private context struct.
* Remove unused variable.
--- drivers/staging/media/hantro/Makefile | 1 + .../staging/media/hantro/hantro_g1_h264_dec.c | 292 ++++++++++++++++++ drivers/staging/media/hantro/hantro_hw.h | 1 + 3 files changed, 294 insertions(+) create mode 100644 drivers/staging/media/hantro/hantro_g1_h264_dec.c
drivers/staging/media/hantro/Makefile +1 -0
@@ -4,6 +4,7 @@ hantro-vpu-y += \ hantro_drv.o \ hantro_v4l2.o \ hantro_h1_jpeg_enc.o \ + hantro_g1_h264_dec.o \ hantro_g1_mpeg2_dec.o \ hantro_g1_vp8_dec.o \ rk3399_vpu_hw_jpeg_enc.o \
drivers/staging/media/hantro/hantro_g1_h264_dec.c new file +292 -0
292 lines not on this page · raw message
drivers/staging/media/hantro/hantro_hw.h +1 -0
@@ -161,6 +161,7 @@ void hantro_jpeg_enc_exit(struct hantro_ctx *ctx);struct vb2_buffer *hantro_h264_get_ref_buf (struct hantro_ctx *ctx,unsigned int dpb_idx);int hantro_h264_dec_prepare_run (struct hantro_ctx *ctx); +void hantro_g1_h264_dec_run (struct hantro_ctx *ctx);int hantro_h264_dec_init (struct hantro_ctx *ctx);void hantro_h264_dec_exit (struct hantro_ctx *ctx);
Signature · 1 line
2.22.0
PATCH v5 11/11 media: hantro: Enable H264 decoding on rk3288
In reply to Ezequiel GarciaFrom: Hertz Wong <[email protected]>
Now that the generic bits have been added, we can activate H264 decoding on rk3288.
---
Changes in v5:
* None.
Changes in v4:
* None.
--- drivers/staging/media/hantro/rk3288_vpu_hw.c | 21 +++++++++++++++++++- 1 file changed, 20 insertions(+), 1 deletion(-)
drivers/staging/media/hantro/rk3288_vpu_hw.c +20 -1
@@ -61,6 +61,19 @@ static const struct hantro_fmt rk3288_vpu_dec_fmts[] = { .fourcc = V4L2_PIX_FMT_NV12, .codec_mode = HANTRO_MODE_NONE, }, + { + .fourcc = V4L2_PIX_FMT_H264_SLICE, + .codec_mode = HANTRO_MODE_H264_DEC, + .max_depth = 2, + .frmsize = { + .min_width = 48, + .max_width = 3840, + .step_width = H264_MB_DIM, + .min_height = 48, + .max_height = 2160, + .step_height = H264_MB_DIM, + }, + }, { .fourcc = V4L2_PIX_FMT_MPEG2_SLICE, .codec_mode = HANTRO_MODE_MPEG2_DEC,
@@ -162,6 +175,12 @@ static const struct hantro_codec_ops rk3288_vpu_codec_ops[] = { .init = hantro_jpeg_enc_init, .exit = hantro_jpeg_enc_exit, }, + [HANTRO_MODE_H264_DEC] = { + .run = hantro_g1_h264_dec_run, + .reset = rk3288_vpu_dec_reset, + .init = hantro_h264_dec_init, + .exit = hantro_h264_dec_exit, + }, [HANTRO_MODE_MPEG2_DEC] = { .run = hantro_g1_mpeg2_dec_run, .reset = rk3288_vpu_dec_reset,
@@ -197,7 +216,7 @@ const struct hantro_variant rk3288_vpu_variant = { .dec_fmts = rk3288_vpu_dec_fmts, .num_dec_fmts = ARRAY_SIZE(rk3288_vpu_dec_fmts), .codec =HANTRO_JPEG_ENCODER |HANTRO_MPEG2_DECODER | - HANTRO_VP8_DECODER, + HANTRO_VP8_DECODER | HANTRO_H264_DECODER, .codec_ops = rk3288_vpu_codec_ops, .irqs = rk3288_irqs, .num_irqs = ARRAY_SIZE(rk3288_irqs),
Signature · 1 line
2.22.0