[PATCH v5 1/7] avcodec/v4l2_context: don't reinit output queue when dynamic resolution change
in the v4l2 stateful video document, we can see the following description: During the resolution change sequence, the OUTPUT queue must remain streaming. Calling VIDIOC_STREAMOFF() on the OUTPUT queue would abort the sequence and initiate a seek. In principle, the OUTPUT queue operates separately from the CAPTURE queue and this remains true for the duration of the entire resolution change sequence as well. so don't reinit the output queue when handling the resolution change event Signed-off-by: Ming Qian <ming.qian@nxp.com> --- libavcodec/v4l2_context.c | 27 ++------------------------- 1 file changed, 2 insertions(+), 25 deletions(-) diff --git a/libavcodec/v4l2_context.c b/libavcodec/v4l2_context.c index ff1ea8e57b08..dda5157698c3 100644 --- a/libavcodec/v4l2_context.c +++ b/libavcodec/v4l2_context.c @@ -162,9 +162,8 @@ static int v4l2_handle_event(V4L2Context *ctx) { V4L2m2mContext *s = ctx_to_m2mctx(ctx); struct v4l2_format cap_fmt = s->capture.format; - struct v4l2_format out_fmt = s->output.format; struct v4l2_event evt = { 0 }; - int full_reinit, reinit, ret; + int reinit, ret; ret = ioctl(s->fd, VIDIOC_DQEVENT, &evt); if (ret < 0) { @@ -180,25 +179,12 @@ static int v4l2_handle_event(V4L2Context *ctx) if (evt.type != V4L2_EVENT_SOURCE_CHANGE) return 0; - ret = ioctl(s->fd, VIDIOC_G_FMT, &out_fmt); - if (ret) { - av_log(logger(ctx), AV_LOG_ERROR, "%s VIDIOC_G_FMT\n", s->output.name); - return 0; - } - ret = ioctl(s->fd, VIDIOC_G_FMT, &cap_fmt); if (ret) { av_log(logger(ctx), AV_LOG_ERROR, "%s VIDIOC_G_FMT\n", s->capture.name); return 0; } - full_reinit = v4l2_resolution_changed(&s->output, &out_fmt); - if (full_reinit) { - s->output.height = v4l2_get_height(&out_fmt); - s->output.width = v4l2_get_width(&out_fmt); - s->output.sample_aspect_ratio = v4l2_get_sar(&s->output); - } - reinit = v4l2_resolution_changed(&s->capture, &cap_fmt); if (reinit) { s->capture.height = v4l2_get_height(&cap_fmt); @@ -206,18 +192,9 @@ static int v4l2_handle_event(V4L2Context *ctx) s->capture.sample_aspect_ratio = v4l2_get_sar(&s->capture); } - if (full_reinit || reinit) + if (reinit) s->reinit = 1; - if (full_reinit) { - ret = ff_v4l2_m2m_codec_full_reinit(s); - if (ret) { - av_log(logger(ctx), AV_LOG_ERROR, "v4l2_m2m_codec_full_reinit\n"); - return AVERROR(EINVAL); - } - goto reinit_run; - } - if (reinit) { if (s->avctx) ret = ff_set_dimensions(s->avctx, s->capture.width, s->capture.height); -- 2.33.0
ff_v4l2_m2m_codec_full_reinit is not used any more, so remove it. Signed-off-by: Ming Qian <ming.qian@nxp.com> --- libavcodec/v4l2_m2m.c | 76 ------------------------------------------- libavcodec/v4l2_m2m.h | 10 ------ 2 files changed, 86 deletions(-) diff --git a/libavcodec/v4l2_m2m.c b/libavcodec/v4l2_m2m.c index cdfd579810f2..a0ee5a201360 100644 --- a/libavcodec/v4l2_m2m.c +++ b/libavcodec/v4l2_m2m.c @@ -245,82 +245,6 @@ int ff_v4l2_m2m_codec_reinit(V4L2m2mContext *s) return 0; } -int ff_v4l2_m2m_codec_full_reinit(V4L2m2mContext *s) -{ - void *log_ctx = s->avctx; - int ret; - - av_log(log_ctx, AV_LOG_DEBUG, "%s full reinit\n", s->devname); - - /* wait for pending buffer references */ - if (atomic_load(&s->refcount)) - while(sem_wait(&s->refsync) == -1 && errno == EINTR); - - ret = ff_v4l2_context_set_status(&s->output, VIDIOC_STREAMOFF); - if (ret) { - av_log(log_ctx, AV_LOG_ERROR, "output VIDIOC_STREAMOFF\n"); - goto error; - } - - ret = ff_v4l2_context_set_status(&s->capture, VIDIOC_STREAMOFF); - if (ret) { - av_log(log_ctx, AV_LOG_ERROR, "capture VIDIOC_STREAMOFF\n"); - goto error; - } - - /* release and unmmap the buffers */ - ff_v4l2_context_release(&s->output); - ff_v4l2_context_release(&s->capture); - - /* start again now that we know the stream dimensions */ - s->draining = 0; - s->reinit = 0; - - ret = ff_v4l2_context_get_format(&s->output, 0); - if (ret) { - av_log(log_ctx, AV_LOG_DEBUG, "v4l2 output format not supported\n"); - goto error; - } - - ret = ff_v4l2_context_get_format(&s->capture, 0); - if (ret) { - av_log(log_ctx, AV_LOG_DEBUG, "v4l2 capture format not supported\n"); - goto error; - } - - ret = ff_v4l2_context_set_format(&s->output); - if (ret) { - av_log(log_ctx, AV_LOG_ERROR, "can't set v4l2 output format\n"); - goto error; - } - - ret = ff_v4l2_context_set_format(&s->capture); - if (ret) { - av_log(log_ctx, AV_LOG_ERROR, "can't to set v4l2 capture format\n"); - goto error; - } - - ret = ff_v4l2_context_init(&s->output); - if (ret) { - av_log(log_ctx, AV_LOG_ERROR, "no v4l2 output context's buffers\n"); - goto error; - } - - /* decoder's buffers need to be updated at a later stage */ - if (s->avctx && !av_codec_is_decoder(s->avctx->codec)) { - ret = ff_v4l2_context_init(&s->capture); - if (ret) { - av_log(log_ctx, AV_LOG_ERROR, "no v4l2 capture context's buffers\n"); - goto error; - } - } - - return 0; - -error: - return ret; -} - static void v4l2_m2m_destroy_context(void *opaque, uint8_t *context) { V4L2m2mContext *s = (V4L2m2mContext*)context; diff --git a/libavcodec/v4l2_m2m.h b/libavcodec/v4l2_m2m.h index b67b21633109..16e0a6d6b90f 100644 --- a/libavcodec/v4l2_m2m.h +++ b/libavcodec/v4l2_m2m.h @@ -119,14 +119,4 @@ int ff_v4l2_m2m_codec_end(V4L2m2mPriv *priv); */ int ff_v4l2_m2m_codec_reinit(V4L2m2mContext *ctx); -/** - * Reinitializes the V4L2m2mContext when the driver cannot continue processing - * with the any of the current V4L2Contexts (ie, changes in output and capture). - * - * @param[in] ctx The V4L2m2mContext instantiated by the encoder/decoder. - * - * @returns 0 in case of success, negative number otherwise - */ -int ff_v4l2_m2m_codec_full_reinit(V4L2m2mContext *ctx); - #endif /* AVCODEC_V4L2_M2M_H */ -- 2.33.0
on dynamic resolution change, The decoding process must be resumed with either a pair of calls to VIDIOC_STREAMOFF() and VIDIOC_STREAMON() on the CAPTURE queue, or a call to VIDIOC_DECODER_CMD() with the V4L2_DEC_CMD_START command. Signed-off-by: Ming Qian <ming.qian@nxp.com> --- libavcodec/v4l2_context.c | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+) diff --git a/libavcodec/v4l2_context.c b/libavcodec/v4l2_context.c index dda5157698c3..a181f884d2a6 100644 --- a/libavcodec/v4l2_context.c +++ b/libavcodec/v4l2_context.c @@ -153,6 +153,21 @@ static inline void v4l2_save_to_context(V4L2Context* ctx, struct v4l2_format_upd } } +static int v4l2_start_decode(V4L2Context *ctx) +{ + struct v4l2_decoder_cmd cmd = { + .cmd = V4L2_DEC_CMD_START, + .flags = 0, + }; + int ret; + + ret = ioctl(ctx_to_m2mctx(ctx)->fd, VIDIOC_DECODER_CMD, &cmd); + if (ret) + return AVERROR(errno); + + return 0; +} + /** * handle resolution change event and end of stream event * returns 1 if reinit was successful, negative if it failed @@ -190,6 +205,9 @@ static int v4l2_handle_event(V4L2Context *ctx) s->capture.height = v4l2_get_height(&cap_fmt); s->capture.width = v4l2_get_width(&cap_fmt); s->capture.sample_aspect_ratio = v4l2_get_sar(&s->capture); + } else { + v4l2_start_decode(ctx); + return 0; } if (reinit) -- 2.33.0
On Tue, 04. Jan 17:08, Ming Qian wrote:
on dynamic resolution change, The decoding process must be resumed with either a pair of calls to VIDIOC_STREAMOFF() and VIDIOC_STREAMON() on the CAPTURE queue, or a call to VIDIOC_DECODER_CMD() with the V4L2_DEC_CMD_START command.
For the patch title, I think something like the following would be better: avcodec/v4l2_context: send start decode command after dynamic resolution change event I'll add that this fixes decoding of https://streams.videolan.org/ffmpeg/incoming/720p60.mp4 on RPi4.
Signed-off-by: Ming Qian <ming.qian@nxp.com> --- libavcodec/v4l2_context.c | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+)
diff --git a/libavcodec/v4l2_context.c b/libavcodec/v4l2_context.c index dda5157698c3..a181f884d2a6 100644 --- a/libavcodec/v4l2_context.c +++ b/libavcodec/v4l2_context.c @@ -153,6 +153,21 @@ static inline void v4l2_save_to_context(V4L2Context* ctx, struct v4l2_format_upd } }
+static int v4l2_start_decode(V4L2Context *ctx) +{ + struct v4l2_decoder_cmd cmd = { + .cmd = V4L2_DEC_CMD_START, + .flags = 0, + }; + int ret; + + ret = ioctl(ctx_to_m2mctx(ctx)->fd, VIDIOC_DECODER_CMD, &cmd); + if (ret) + return AVERROR(errno); + + return 0; +} + /** * handle resolution change event and end of stream event * returns 1 if reinit was successful, negative if it failed @@ -190,6 +205,9 @@ static int v4l2_handle_event(V4L2Context *ctx) s->capture.height = v4l2_get_height(&cap_fmt); s->capture.width = v4l2_get_width(&cap_fmt); s->capture.sample_aspect_ratio = v4l2_get_sar(&s->capture); + } else { + v4l2_start_decode(ctx); + return 0; }
if (reinit)
lgtm otherwise. Thanks -- Andriy
The dynamic resoltuion change occurs when he decoder detects a coded frame with one or more of the following parameters different from those previously established (and reflected by corresponding queries): 1. coded resolution (OUTPUT width and height) 2. visible resolution (selection rectangles) 3. the minimum number of buffers needed for decoding 4. bit-depth of the bitstream has been changed when the bit-depth is changed, the pixel format will be changed, so the format is changed after a source change event, we should handle it and reinit the capture queue. Signed-off-by: Ming Qian <ming.qian@nxp.com> --- libavcodec/v4l2_context.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/libavcodec/v4l2_context.c b/libavcodec/v4l2_context.c index a181f884d2a6..0a0f5b226960 100644 --- a/libavcodec/v4l2_context.c +++ b/libavcodec/v4l2_context.c @@ -85,9 +85,11 @@ static inline unsigned int v4l2_resolution_changed(V4L2Context *ctx, struct v4l2 { struct v4l2_format *fmt1 = &ctx->format; int ret = V4L2_TYPE_IS_MULTIPLANAR(ctx->type) ? + fmt1->fmt.pix_mp.pixelformat != fmt2->fmt.pix_mp.pixelformat || fmt1->fmt.pix_mp.width != fmt2->fmt.pix_mp.width || fmt1->fmt.pix_mp.height != fmt2->fmt.pix_mp.height : + fmt1->fmt.pix.pixelformat != fmt2->fmt.pix.pixelformat || fmt1->fmt.pix.width != fmt2->fmt.pix.width || fmt1->fmt.pix.height != fmt2->fmt.pix.height; -- 2.33.0
client need to resume the decoding process after it dequeues the source change event. no matter what's the return value of v4l2_resolution_changed(). if the client doesn't resume the decoding process, the decoder may keep waiting in documentation of v4l2 stateful decoder, we can see the following description: The client must continue the sequence as described below to continue the decoding process. 1. Dequeue the source change event. Important A source change triggers an implicit decoder drain, similar to the explicit Drain sequence. The decoder is stopped after it completes. The decoding process must be resumed with either a pair of calls to VIDIOC_STREAMOFF() and VIDIOC_STREAMON() on the CAPTURE queue, or a call to VIDIOC_DECODER_CMD() with the V4L2_DEC_CMD_START command. 2. Continue with the Capture Setup sequence. Signed-off-by: Ming Qian <ming.qian@nxp.com> --- libavcodec/v4l2_context.c | 31 +++++++++++-------------------- 1 file changed, 11 insertions(+), 20 deletions(-) diff --git a/libavcodec/v4l2_context.c b/libavcodec/v4l2_context.c index 0a0f5b226960..b7b584dfbef6 100644 --- a/libavcodec/v4l2_context.c +++ b/libavcodec/v4l2_context.c @@ -180,7 +180,7 @@ static int v4l2_handle_event(V4L2Context *ctx) V4L2m2mContext *s = ctx_to_m2mctx(ctx); struct v4l2_format cap_fmt = s->capture.format; struct v4l2_event evt = { 0 }; - int reinit, ret; + int ret; ret = ioctl(s->fd, VIDIOC_DQEVENT, &evt); if (ret < 0) { @@ -202,8 +202,7 @@ static int v4l2_handle_event(V4L2Context *ctx) return 0; } - reinit = v4l2_resolution_changed(&s->capture, &cap_fmt); - if (reinit) { + if (v4l2_resolution_changed(&s->capture, &cap_fmt)) { s->capture.height = v4l2_get_height(&cap_fmt); s->capture.width = v4l2_get_width(&cap_fmt); s->capture.sample_aspect_ratio = v4l2_get_sar(&s->capture); @@ -212,28 +211,20 @@ static int v4l2_handle_event(V4L2Context *ctx) return 0; } - if (reinit) - s->reinit = 1; + s->reinit = 1; - if (reinit) { - if (s->avctx) - ret = ff_set_dimensions(s->avctx, s->capture.width, s->capture.height); - if (ret < 0) - av_log(logger(ctx), AV_LOG_WARNING, "update avcodec height and width\n"); + if (s->avctx) + ret = ff_set_dimensions(s->avctx, s->capture.width, s->capture.height); + if (ret < 0) + av_log(logger(ctx), AV_LOG_WARNING, "update avcodec height and width\n"); - ret = ff_v4l2_m2m_codec_reinit(s); - if (ret) { - av_log(logger(ctx), AV_LOG_ERROR, "v4l2_m2m_codec_reinit\n"); - return AVERROR(EINVAL); - } - goto reinit_run; + ret = ff_v4l2_m2m_codec_reinit(s); + if (ret) { + av_log(logger(ctx), AV_LOG_ERROR, "v4l2_m2m_codec_reinit\n"); + return AVERROR(EINVAL); } - /* dummy event received */ - return 0; - /* reinit executed */ -reinit_run: return 1; } -- 2.33.0
On Tue, 04. Jan 17:08, Ming Qian wrote:
client need to resume the decoding process after it dequeues the source change event. no matter what's the return value of v4l2_resolution_changed(). if the client doesn't resume the decoding process, the decoder may keep waiting
in documentation of v4l2 stateful decoder, we can see the following description: The client must continue the sequence as described below to continue the decoding process. 1. Dequeue the source change event. Important A source change triggers an implicit decoder drain, similar to the explicit Drain sequence. The decoder is stopped after it completes. The decoding process must be resumed with either a pair of calls to VIDIOC_STREAMOFF() and VIDIOC_STREAMON() on the CAPTURE queue, or a call to VIDIOC_DECODER_CMD() with the V4L2_DEC_CMD_START command. 2. Continue with the Capture Setup sequence.
This is just a clean up to remove the reinit variable so the commit title and description are wrong. Thanks, -- Andriy
On Tue, 04. Jan 17:41, Andriy Gelman wrote:
On Tue, 04. Jan 17:08, Ming Qian wrote:
client need to resume the decoding process after it dequeues the source change event. no matter what's the return value of v4l2_resolution_changed(). if the client doesn't resume the decoding process, the decoder may keep waiting
in documentation of v4l2 stateful decoder, we can see the following description: The client must continue the sequence as described below to continue the decoding process. 1. Dequeue the source change event. Important A source change triggers an implicit decoder drain, similar to the explicit Drain sequence. The decoder is stopped after it completes. The decoding process must be resumed with either a pair of calls to VIDIOC_STREAMOFF() and VIDIOC_STREAMON() on the CAPTURE queue, or a call to VIDIOC_DECODER_CMD() with the V4L2_DEC_CMD_START command. 2. Continue with the Capture Setup sequence.
This is just a clean up to remove the reinit variable so the commit title and description are wrong.
Updated the commit message and pushed. Thanks, -- Andriy
Check v4l2 context before call VIDIOC_STREAMON() or VIDIOC_STREAMOFF(). Signed-off-by: Ming Qian <ming.qian@nxp.com> --- libavcodec/v4l2_context.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/libavcodec/v4l2_context.c b/libavcodec/v4l2_context.c index b7b584dfbef6..8910ae08d3a5 100644 --- a/libavcodec/v4l2_context.c +++ b/libavcodec/v4l2_context.c @@ -562,6 +562,9 @@ int ff_v4l2_context_set_status(V4L2Context* ctx, uint32_t cmd) int type = ctx->type; int ret; + if (ctx->streamon == (cmd == VIDIOC_STREAMON)) + return 0; + ret = ioctl(ctx_to_m2mctx(ctx)->fd, cmd, &type); if (ret < 0) return AVERROR(errno); -- 2.33.0
there are two proper ways to setup capture queue. 1. client wait the source change event, then setup the capture queue and streamon 2. client setup the capture queue in advance, but to avoid time issues, client should start the capture queue before it enqueue the sequence header to decoder driver through output queue. and the sequence header is always in the first frame, so client should start capture before enqueue the first frame. ffmpeg use the method 2 to setup capture queue, but currently ffmpeg enqueue the first frame before starting the capture queue. so in driver side, there are time issues. when driver has parsed the resolution from sequence header, but the client may not finished setup the capture. so driver can't decide whether to notify a source change event to client. and the following flow may be chaotic. And it's OK that client setup capture queue first, then enqueue the first frame. Signed-off-by: Ming Qian <ming.qian@nxp.com> --- libavcodec/v4l2_m2m_dec.c | 18 ++++++++---------- 1 file changed, 8 insertions(+), 10 deletions(-) diff --git a/libavcodec/v4l2_m2m_dec.c b/libavcodec/v4l2_m2m_dec.c index b0c3d30ac8ae..e67758531ace 100644 --- a/libavcodec/v4l2_m2m_dec.c +++ b/libavcodec/v4l2_m2m_dec.c @@ -153,6 +153,14 @@ static int v4l2_receive_frame(AVCodecContext *avctx, AVFrame *frame) if (s->draining) goto dequeue; + ret = v4l2_try_start(avctx); + if (ret) { + /* can't recover */ + if (ret != AVERROR(ENOMEM)) + ret = 0; + goto fail; + } + ret = ff_v4l2_context_enqueue_packet(output, &s->buf_pkt); if (ret < 0 && ret != AVERROR(EAGAIN)) goto fail; @@ -161,16 +169,6 @@ static int v4l2_receive_frame(AVCodecContext *avctx, AVFrame *frame) if (ret != AVERROR(EAGAIN)) av_packet_unref(&s->buf_pkt); - if (!s->draining) { - ret = v4l2_try_start(avctx); - if (ret) { - /* cant recover */ - if (ret != AVERROR(ENOMEM)) - ret = 0; - goto fail; - } - } - dequeue: return ff_v4l2_context_dequeue_frame(capture, frame, -1); fail: -- 2.33.0
Hi Andriy, What do you think of this patch? The Initialization flow defined in linux/Documentation/userspace-api/media/v4l/dev-decoder.rst 1. Set the coded format on OUTPUT via VIDIOC_S_FMT(). 2. Allocate source (bytestream) buffers via VIDIOC_REQBUFS() on OUTPUT. 3. Start streaming on the OUTPUT queue via VIDIOC_STREAMON(). 4. This step only applies to coded formats that contain resolution information in the stream. Continue queuing/dequeuing bytestream buffers to/from the OUTPUT queue via VIDIOC_QBUF() and VIDIOC_DQBUF(). The buffers will be processed and returned to the client in order, until required metadata to configure the CAPTURE queue are found. This is indicated by the decoder sending a V4L2_EVENT_SOURCE_CHANGE event with changes set to V4L2_EVENT_SRC_CH_RESOLUTION. Note: A client capable of acquiring stream parameters from the bytestream on its own may attempt to set the width and height of the OUTPUT format to non-zero values matching the coded size of the stream, skip this step and continue with the Capture Setup sequence. 5. Continue with the Capture Setup sequence. In ffmpeg's implementation, ffmpeg will set non-zero width and height on output queue, so the step 4 should be skipped, and setup the capture queue directly. So the flow should be: 1. Set the coded format and valid resolution on OUTPUT via VIDIOC_S_FMT(). 2. Allocate source (bytestream) buffers via VIDIOC_REQBUFS() on OUTPUT. 3. Start streaming on the OUTPUT queue via VIDIOC_STREAMON(). 5. Continue with the Capture Setup sequence. And this patch is just following the above flow. Ming
From: ffmpeg-devel [mailto:ffmpeg-devel-bounces@ffmpeg.org] On Behalf Of Ming Qian Sent: Tuesday, January 4, 2022 5:09 PM To: ffmpeg-devel@ffmpeg.org Subject: [EXT] [FFmpeg-devel] [PATCH v5 7/7] avcodec/v4l2_m2m_dec: setup capture queue before enqueue the first frame
Caution: EXT Email
there are two proper ways to setup capture queue. 1. client wait the source change event, then setup the capture queue and streamon 2. client setup the capture queue in advance, but to avoid time issues, client should start the capture queue before it enqueue the sequence header to decoder driver through output queue. and the sequence header is always in the first frame, so client should start capture before enqueue the first frame.
ffmpeg use the method 2 to setup capture queue, but currently ffmpeg enqueue the first frame before starting the capture queue. so in driver side, there are time issues. when driver has parsed the resolution from sequence header, but the client may not finished setup the capture. so driver can't decide whether to notify a source change event to client. and the following flow may be chaotic.
And it's OK that client setup capture queue first, then enqueue the first frame.
Signed-off-by: Ming Qian <ming.qian@nxp.com> --- libavcodec/v4l2_m2m_dec.c | 18 ++++++++---------- 1 file changed, 8 insertions(+), 10 deletions(-)
diff --git a/libavcodec/v4l2_m2m_dec.c b/libavcodec/v4l2_m2m_dec.c index b0c3d30ac8ae..e67758531ace 100644 --- a/libavcodec/v4l2_m2m_dec.c +++ b/libavcodec/v4l2_m2m_dec.c @@ -153,6 +153,14 @@ static int v4l2_receive_frame(AVCodecContext *avctx, AVFrame *frame) if (s->draining) goto dequeue;
+ ret = v4l2_try_start(avctx); + if (ret) { + /* can't recover */ + if (ret != AVERROR(ENOMEM)) + ret = 0; + goto fail; + } + ret = ff_v4l2_context_enqueue_packet(output, &s->buf_pkt); if (ret < 0 && ret != AVERROR(EAGAIN)) goto fail; @@ -161,16 +169,6 @@ static int v4l2_receive_frame(AVCodecContext *avctx, AVFrame *frame) if (ret != AVERROR(EAGAIN)) av_packet_unref(&s->buf_pkt);
- if (!s->draining) { - ret = v4l2_try_start(avctx); - if (ret) { - /* cant recover */ - if (ret != AVERROR(ENOMEM)) - ret = 0; - goto fail; - } - } - dequeue: return ff_v4l2_context_dequeue_frame(capture, frame, -1); fail: -- 2.33.0
_______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fffmpeg. org%2Fmailman%2Flistinfo%2Fffmpeg-devel&data=04%7C01%7Cming.qi an%40nxp.com%7C6b9621a0804c45f9da9108d9cf61fb87%7C686ea1d3bc2b 4c6fa92cd99c5c301635%7C0%7C1%7C637768841997715387%7CUnknown% 7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwi LCJXVCI6Mn0%3D%7C3000&sdata=E7V4kwlJtxbh6jnisMV8FWAwwDzksS qLp82kapSW4FA%3D&reserved=0
To unsubscribe, visit link above, or email ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
Hi Ming, On Mon, 21. Mar 07:27, Ming Qian wrote:
Hi Andriy,
What do you think of this patch?
The Initialization flow defined in linux/Documentation/userspace-api/media/v4l/dev-decoder.rst 1. Set the coded format on OUTPUT via VIDIOC_S_FMT(). 2. Allocate source (bytestream) buffers via VIDIOC_REQBUFS() on OUTPUT. 3. Start streaming on the OUTPUT queue via VIDIOC_STREAMON(). 4. This step only applies to coded formats that contain resolution information in the stream. Continue queuing/dequeuing bytestream buffers to/from the OUTPUT queue via VIDIOC_QBUF() and VIDIOC_DQBUF(). The buffers will be processed and returned to the client in order, until required metadata to configure the CAPTURE queue are found. This is indicated by the decoder sending a V4L2_EVENT_SOURCE_CHANGE event with changes set to V4L2_EVENT_SRC_CH_RESOLUTION. Note: A client capable of acquiring stream parameters from the bytestream on its own may attempt to set the width and height of the OUTPUT format to non-zero values matching the coded size of the stream, skip this step and continue with the Capture Setup sequence.
5. Continue with the Capture Setup sequence.
In ffmpeg's implementation, ffmpeg will set non-zero width and height on output queue, so the step 4 should be skipped, and setup the capture queue directly. So the flow should be: 1. Set the coded format and valid resolution on OUTPUT via VIDIOC_S_FMT(). 2. Allocate source (bytestream) buffers via VIDIOC_REQBUFS() on OUTPUT. 3. Start streaming on the OUTPUT queue via VIDIOC_STREAMON(). 5. Continue with the Capture Setup sequence.
And this patch is just following the above flow.
Ming
The v4 version didn't work for me on odroid xu4 http://ffmpeg.org/pipermail/ffmpeg-devel/2022-January/290679.html I haven't had time retest but don't think anything has changed. -- Andriy
From: Andriy Gelman [mailto:andriy.gelman@gmail.com] Sent: Saturday, March 26, 2022 11:43 PM To: FFmpeg development discussions and patches <ffmpeg-devel@ffmpeg.org> Cc: Ming Qian <ming.qian@nxp.com> Subject: Re: [FFmpeg-devel] [EXT] [PATCH v5 7/7] avcodec/v4l2_m2m_dec: setup capture queue before enqueue the first frame
Caution: EXT Email
Hi Ming,
Hi Andriy,
What do you think of this patch?
The Initialization flow defined in
1. Set the coded format on OUTPUT via VIDIOC_S_FMT(). 2. Allocate source (bytestream) buffers via VIDIOC_REQBUFS() on OUTPUT. 3. Start streaming on the OUTPUT queue via VIDIOC_STREAMON(). 4. This step only applies to coded formats that contain resolution information in the stream. Continue queuing/dequeuing bytestream buffers to/from the OUTPUT queue via VIDIOC_QBUF() and VIDIOC_DQBUF(). The buffers will be processed and returned to the client in order, until required
On Mon, 21. Mar 07:27, Ming Qian wrote: linux/Documentation/userspace-api/media/v4l/dev-decoder.rst metadata to configure the CAPTURE queue are found. This is indicated by the decoder sending a V4L2_EVENT_SOURCE_CHANGE event with changes set to V4L2_EVENT_SRC_CH_RESOLUTION.
Note: A client capable of acquiring stream parameters from the
bytestream on its own may attempt to set the width and height of the OUTPUT format to non-zero values matching the coded size of the stream, skip this step and continue with the Capture Setup sequence.
5. Continue with the Capture Setup sequence.
In ffmpeg's implementation, ffmpeg will set non-zero width and height
on output queue, so the step 4 should be skipped, and setup the capture queue directly.
So the flow should be: 1. Set the coded format and valid resolution on OUTPUT via VIDIOC_S_FMT(). 2. Allocate source (bytestream) buffers via VIDIOC_REQBUFS() on OUTPUT. 3. Start streaming on the OUTPUT queue via VIDIOC_STREAMON(). 5. Continue with the Capture Setup sequence.
And this patch is just following the above flow.
Ming
The v4 version didn't work for me on odroid xu4 https://eur01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fffmpeg.o rg%2Fpipermail%2Fffmpeg-devel%2F2022-January%2F290679.html&dat a=04%7C01%7Cming.qian%40nxp.com%7Cca988d592253473cbac408da0f3f4 801%7C686ea1d3bc2b4c6fa92cd99c5c301635%7C0%7C0%7C637839061704 093567%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2l uMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=I80bpE70S6JI 8A%2BaXQGq4JyxL0eBbtLN58p5SpqlA6s%3D&reserved=0
I haven't had time retest but don't think anything has changed.
-- Andriy
Hi Andriy, I check the S5P MFC Video Codec driver, and indeed it requires the client enqueue the output buffer which contains the sequence header. And in vidioc_g_fmt(), driver will block and wait until the header is parsed. And indeed my patch will not work on it. But it seems don't meet the specification. For S5P MFC, the step 4 can't be skipped, and the client should wait the V4L2_EVENT_SOURCE_CHANGE event before the step 5 setting up capture queue. The ffmpeg v4l2 decoder don't skip step 4, but also doesn't wait the V4L2_EVENT_SOURCE_CHANGE event. So in current, we should make it work on more devices instead of more spec? The Initialization flow defined in linux/Documentation/userspace-api/media/v4l/dev-decoder.rst 1. Set the coded format on OUTPUT via VIDIOC_S_FMT(). 2. Allocate source (bytestream) buffers via VIDIOC_REQBUFS() on OUTPUT. 3. Start streaming on the OUTPUT queue via VIDIOC_STREAMON(). 4. This step only applies to coded formats that contain resolution information in the stream. Continue queuing/dequeuing bytestream buffers to/from the OUTPUT queue via VIDIOC_QBUF() and VIDIOC_DQBUF(). The buffers will be processed and returned to the client in order, until required metadata to configure the CAPTURE queue are found. This is indicated by the decoder sending a V4L2_EVENT_SOURCE_CHANGE event with changes set to V4L2_EVENT_SRC_CH_RESOLUTION. Note: A client capable of acquiring stream parameters from the bytestream on its own may attempt to set the width and height of the OUTPUT format to non-zero values matching the coded size of the stream, skip this step and continue with the Capture Setup sequence. 5. Continue with the Capture Setup sequence. Ming
Hi Ming, Sorry for the late reply. On Tue, 29. Mar 06:51, Ming Qian wrote:
From: Andriy Gelman [mailto:andriy.gelman@gmail.com] Sent: Saturday, March 26, 2022 11:43 PM To: FFmpeg development discussions and patches <ffmpeg-devel@ffmpeg.org> Cc: Ming Qian <ming.qian@nxp.com> Subject: Re: [FFmpeg-devel] [EXT] [PATCH v5 7/7] avcodec/v4l2_m2m_dec: setup capture queue before enqueue the first frame
Caution: EXT Email
Hi Ming,
Hi Andriy,
What do you think of this patch?
The Initialization flow defined in
1. Set the coded format on OUTPUT via VIDIOC_S_FMT(). 2. Allocate source (bytestream) buffers via VIDIOC_REQBUFS() on OUTPUT. 3. Start streaming on the OUTPUT queue via VIDIOC_STREAMON(). 4. This step only applies to coded formats that contain resolution information in the stream. Continue queuing/dequeuing bytestream buffers to/from the OUTPUT queue via VIDIOC_QBUF() and VIDIOC_DQBUF(). The buffers will be processed and returned to the client in order, until required
On Mon, 21. Mar 07:27, Ming Qian wrote: linux/Documentation/userspace-api/media/v4l/dev-decoder.rst metadata to configure the CAPTURE queue are found. This is indicated by the decoder sending a V4L2_EVENT_SOURCE_CHANGE event with changes set to V4L2_EVENT_SRC_CH_RESOLUTION.
Note: A client capable of acquiring stream parameters from the
bytestream on its own may attempt to set the width and height of the OUTPUT format to non-zero values matching the coded size of the stream, skip this step and continue with the Capture Setup sequence.
5. Continue with the Capture Setup sequence.
In ffmpeg's implementation, ffmpeg will set non-zero width and height
on output queue, so the step 4 should be skipped, and setup the capture queue directly.
So the flow should be: 1. Set the coded format and valid resolution on OUTPUT via VIDIOC_S_FMT(). 2. Allocate source (bytestream) buffers via VIDIOC_REQBUFS() on OUTPUT. 3. Start streaming on the OUTPUT queue via VIDIOC_STREAMON(). 5. Continue with the Capture Setup sequence.
And this patch is just following the above flow.
Ming
The v4 version didn't work for me on odroid xu4 https://eur01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fffmpeg.o rg%2Fpipermail%2Fffmpeg-devel%2F2022-January%2F290679.html&dat a=04%7C01%7Cming.qian%40nxp.com%7Cca988d592253473cbac408da0f3f4 801%7C686ea1d3bc2b4c6fa92cd99c5c301635%7C0%7C0%7C637839061704 093567%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2l uMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=I80bpE70S6JI 8A%2BaXQGq4JyxL0eBbtLN58p5SpqlA6s%3D&reserved=0
I haven't had time retest but don't think anything has changed.
-- Andriy
Hi Andriy, I check the S5P MFC Video Codec driver, and indeed it requires the client enqueue the output buffer which contains the sequence header. And in vidioc_g_fmt(), driver will block and wait until the header is parsed. And indeed my patch will not work on it. But it seems don't meet the specification. For S5P MFC, the step 4 can't be skipped, and the client should wait the V4L2_EVENT_SOURCE_CHANGE event before the step 5 setting up capture queue. The ffmpeg v4l2 decoder don't skip step 4, but also doesn't wait the V4L2_EVENT_SOURCE_CHANGE event.
So in current, we should make it work on more devices instead of more spec?
Personally I'd prefer not to break users. Odroid xu4 is a popular board afaik. But I'm also not the maintainer of v4l2m2m. IMO it would be better the fix the problem in the driver first.
The Initialization flow defined in linux/Documentation/userspace-api/media/v4l/dev-decoder.rst 1. Set the coded format on OUTPUT via VIDIOC_S_FMT(). 2. Allocate source (bytestream) buffers via VIDIOC_REQBUFS() on OUTPUT. 3. Start streaming on the OUTPUT queue via VIDIOC_STREAMON(). 4. This step only applies to coded formats that contain resolution information in the stream. Continue queuing/dequeuing bytestream buffers to/from the OUTPUT queue via VIDIOC_QBUF() and VIDIOC_DQBUF(). The buffers will be processed and returned to the client in order, until required metadata to configure the CAPTURE queue are found. This is indicated by the decoder sending a V4L2_EVENT_SOURCE_CHANGE event with changes set to V4L2_EVENT_SRC_CH_RESOLUTION. Note: A client capable of acquiring stream parameters from the bytestream on its own may attempt to set the width and height of the OUTPUT format to non-zero values matching the coded size of the stream, skip this step and continue with the Capture Setup sequence. 5. Continue with the Capture Setup sequence.
-- Andriy
From: Andriy Gelman <andriy.gelman@gmail.com> Sent: Monday, April 4, 2022 6:29 AM To: Ming Qian <ming.qian@nxp.com> Cc: FFmpeg development discussions and patches <ffmpeg- devel@ffmpeg.org> Subject: Re: [FFmpeg-devel] [EXT] [PATCH v5 7/7] avcodec/v4l2_m2m_dec: setup capture queue before enqueue the first frame
Caution: EXT Email
Hi Ming,
Sorry for the late reply.
From: Andriy Gelman [mailto:andriy.gelman@gmail.com] Sent: Saturday, March 26, 2022 11:43 PM To: FFmpeg development discussions and patches <ffmpeg-devel@ffmpeg.org> Cc: Ming Qian <ming.qian@nxp.com> Subject: Re: [FFmpeg-devel] [EXT] [PATCH v5 7/7] avcodec/v4l2_m2m_dec: setup capture queue before enqueue the first frame
Caution: EXT Email
Hi Ming,
Hi Andriy,
What do you think of this patch?
The Initialization flow defined in
On Mon, 21. Mar 07:27, Ming Qian wrote: linux/Documentation/userspace-api/media/v4l/dev-decoder.rst
1. Set the coded format on OUTPUT via VIDIOC_S_FMT(). 2. Allocate source (bytestream) buffers via VIDIOC_REQBUFS() on OUTPUT. 3. Start streaming on the OUTPUT queue via VIDIOC_STREAMON(). 4. This step only applies to coded formats that contain resolution information in the stream. Continue queuing/dequeuing bytestream buffers to/from the OUTPUT queue via VIDIOC_QBUF() and VIDIOC_DQBUF(). The buffers will be processed and returned to the client in order, until required metadata to configure the CAPTURE queue are found. This is indicated by the decoder sending a V4L2_EVENT_SOURCE_CHANGE event with changes set to V4L2_EVENT_SRC_CH_RESOLUTION. Note: A client capable of acquiring stream parameters from the bytestream on its own may attempt to set the width and height of the OUTPUT format to non-zero values matching the coded size of the stream, skip this step and continue with the Capture Setup sequence.
5. Continue with the Capture Setup sequence.
In ffmpeg's implementation, ffmpeg will set non-zero width and height on output queue, so the step 4 should be skipped, and setup the capture queue directly. So the flow should be: 1. Set the coded format and valid resolution on OUTPUT via VIDIOC_S_FMT(). 2. Allocate source (bytestream) buffers via VIDIOC_REQBUFS() on OUTPUT. 3. Start streaming on the OUTPUT queue via VIDIOC_STREAMON(). 5. Continue with the Capture Setup sequence.
And this patch is just following the above flow.
Ming
The v4 version didn't work for me on odroid xu4
https://eur01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fffm
On Tue, 29. Mar 06:51, Ming Qian wrote: peg.o%2F&data=04%7C01%7Cming.qian%40nxp.com%7Cfa858e4a13434 ee727
3508da15c1695b%7C686ea1d3bc2b4c6fa92cd99c5c301635%7C0%7C1%7C6378 4621
7684591813%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIj oiV2luM
zIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=wl%2BZaEBNIPZ TOyv
7dgmnHhSQyAs37PjZCkvhwPL%2B6xg%3D&reserved=0 rg%2Fpipermail%2Fffmpeg-devel%2F2022- January%2F290679.html&dat
a=04%7C01%7Cming.qian%40nxp.com%7Cca988d592253473cbac408da0f3f4
801%7C686ea1d3bc2b4c6fa92cd99c5c301635%7C0%7C0%7C637839061704
093567%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2 l
uMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=I80bpE70S6JI
8A%2BaXQGq4JyxL0eBbtLN58p5SpqlA6s%3D&reserved=0
I haven't had time retest but don't think anything has changed.
-- Andriy
Hi Andriy, I check the S5P MFC Video Codec driver, and indeed it requires the client enqueue the output buffer which contains the sequence header. And in vidioc_g_fmt(), driver will block and wait until the header is parsed. And indeed my patch will not work on it. But it seems don't meet the specification. For S5P MFC, the step 4 can't be skipped, and the client should wait the V4L2_EVENT_SOURCE_CHANGE event before the step 5 setting up capture queue. The ffmpeg v4l2 decoder don't skip step 4, but also doesn't wait the V4L2_EVENT_SOURCE_CHANGE event.
So in current, we should make it work on more devices instead of more
spec?
Personally I'd prefer not to break users. Odroid xu4 is a popular board afaik. But I'm also not the maintainer of v4l2m2m.
IMO it would be better the fix the problem in the driver first.
Hi Andriy, Thanks for your reply. Maybe I can try to work on the driver first. Ming.
The Initialization flow defined in linux/Documentation/userspace-
api/media/v4l/dev-decoder.rst
1. Set the coded format on OUTPUT via VIDIOC_S_FMT(). 2. Allocate source (bytestream) buffers via VIDIOC_REQBUFS() on
OUTPUT.
3. Start streaming on the OUTPUT queue via VIDIOC_STREAMON(). 4. This step only applies to coded formats that contain resolution
information in the stream. Continue queuing/dequeuing bytestream buffers to/from the OUTPUT queue via VIDIOC_QBUF() and VIDIOC_DQBUF(). The buffers will be processed and returned to the client in order, until required metadata to configure the CAPTURE queue are found. This is indicated by the decoder sending a V4L2_EVENT_SOURCE_CHANGE event with changes set to V4L2_EVENT_SRC_CH_RESOLUTION.
Note: A client capable of acquiring stream parameters from the
bytestream on its own may attempt to set the width and height of the OUTPUT format to non-zero values matching the coded size of the stream, skip this step and continue with the Capture Setup sequence.
5. Continue with the Capture Setup sequence.
-- Andriy
On Tue, 04. Jan 17:08, Ming Qian wrote:
in the v4l2 stateful video document, we can see the following description: During the resolution change sequence, the OUTPUT queue must remain streaming. Calling VIDIOC_STREAMOFF() on the OUTPUT queue would abort the sequence and initiate a seek.
In principle, the OUTPUT queue operates separately from the CAPTURE queue and this remains true for the duration of the entire resolution change sequence as well.
so don't reinit the output queue when handling the resolution change event
Signed-off-by: Ming Qian <ming.qian@nxp.com> --- libavcodec/v4l2_context.c | 27 ++------------------------- 1 file changed, 2 insertions(+), 25 deletions(-)
diff --git a/libavcodec/v4l2_context.c b/libavcodec/v4l2_context.c index ff1ea8e57b08..dda5157698c3 100644 --- a/libavcodec/v4l2_context.c +++ b/libavcodec/v4l2_context.c @@ -162,9 +162,8 @@ static int v4l2_handle_event(V4L2Context *ctx) { V4L2m2mContext *s = ctx_to_m2mctx(ctx); struct v4l2_format cap_fmt = s->capture.format; - struct v4l2_format out_fmt = s->output.format; struct v4l2_event evt = { 0 }; - int full_reinit, reinit, ret; + int reinit, ret;
ret = ioctl(s->fd, VIDIOC_DQEVENT, &evt); if (ret < 0) { @@ -180,25 +179,12 @@ static int v4l2_handle_event(V4L2Context *ctx) if (evt.type != V4L2_EVENT_SOURCE_CHANGE) return 0;
- ret = ioctl(s->fd, VIDIOC_G_FMT, &out_fmt); - if (ret) { - av_log(logger(ctx), AV_LOG_ERROR, "%s VIDIOC_G_FMT\n", s->output.name); - return 0; - } - ret = ioctl(s->fd, VIDIOC_G_FMT, &cap_fmt); if (ret) { av_log(logger(ctx), AV_LOG_ERROR, "%s VIDIOC_G_FMT\n", s->capture.name); return 0; }
- full_reinit = v4l2_resolution_changed(&s->output, &out_fmt); - if (full_reinit) { - s->output.height = v4l2_get_height(&out_fmt); - s->output.width = v4l2_get_width(&out_fmt); - s->output.sample_aspect_ratio = v4l2_get_sar(&s->output); - } - reinit = v4l2_resolution_changed(&s->capture, &cap_fmt); if (reinit) { s->capture.height = v4l2_get_height(&cap_fmt); @@ -206,18 +192,9 @@ static int v4l2_handle_event(V4L2Context *ctx) s->capture.sample_aspect_ratio = v4l2_get_sar(&s->capture); }
- if (full_reinit || reinit) + if (reinit) s->reinit = 1;
- if (full_reinit) { - ret = ff_v4l2_m2m_codec_full_reinit(s); - if (ret) { - av_log(logger(ctx), AV_LOG_ERROR, "v4l2_m2m_codec_full_reinit\n"); - return AVERROR(EINVAL); - } - goto reinit_run; - } - if (reinit) { if (s->avctx) ret = ff_set_dimensions(s->avctx, s->capture.width, s->capture.height); -- 2.33.0
lgtm Thanks, -- Andriy
participants (3)
-
Andriy Gelman -
Ming Qian -
Ming Qian