#include <stdio.h>
#include <unistd.h>

#include <libavcodec/avcodec.h>
#include <libavformat/avformat.h>
#include <libavutil/hwcontext.h>
#include <libavutil/hwcontext_drm.h>
#include <libavutil/pixdesc.h>

static enum AVPixelFormat hw_pix_fmt = AV_PIX_FMT_NONE;

static int check_hw_device_type(enum AVHWDeviceType type) {
	enum AVHWDeviceType t = AV_HWDEVICE_TYPE_NONE;
	while (1) {
		t = av_hwdevice_iterate_types(t);
		if (t == AV_HWDEVICE_TYPE_NONE) {
			break;
		}
		if (t == type) {
			return 0;
		}
	}
	return -1;
}

static enum AVPixelFormat get_hw_format(AVCodecContext *ctx,
		const enum AVPixelFormat *pix_fmts) {
	for (size_t i = 0; pix_fmts[i] != AV_PIX_FMT_NONE; i++) {
		if (pix_fmts[i] == hw_pix_fmt) {
			return hw_pix_fmt;
		}
	}

	fprintf(stderr, "Failed to find HW pixel format\n");
	return AV_PIX_FMT_NONE;
}

static int decode_frame(AVFormatContext *input_ctx, int video_stream,
		AVCodecContext *decoder_ctx, AVFrame **frame_ptr) {
	int ret;
	while (1) {
		AVPacket packet;
		ret = av_read_frame(input_ctx, &packet);
		if (ret < 0) {
			fprintf(stderr, "Failed to read frames into packet\n");
			return ret;
		}

		if (video_stream != packet.stream_index) {
			av_packet_unref(&packet);
			continue;
		}

		ret = avcodec_send_packet(decoder_ctx, &packet);
		if (ret < 0) {
			fprintf(stderr, "Failed to send packet to decoder\n");
			return ret;
		}
		av_packet_unref(&packet);

		AVFrame *frame = av_frame_alloc();
		if (!frame) {
			fprintf(stderr, "Failed to allocate frame\n");
			return AVERROR(ENOMEM);
		}

		while (1) {
			ret = avcodec_receive_frame(decoder_ctx, frame);
			if (ret == AVERROR(EAGAIN) || ret == AVERROR_EOF) {
				break;
			} else if (ret < 0) {
				fprintf(stderr, "Failed to receive frame from decoder\n");
				return ret;
			}

			if (frame->format != hw_pix_fmt) {
				continue;
			}

			*frame_ptr = frame;
			return 0;
		}

		av_frame_free(&frame);
	}

	return AVERROR_EOF;
}

int main(int argc, char *argv[]) {
	if (argc < 2) {
		fprintf(stderr, "usage: %s <file>\n", argv[0]);
		return 1;
	}

	enum AVHWDeviceType type = AV_HWDEVICE_TYPE_VAAPI;
	int ret = check_hw_device_type(type);
	if (ret != 0) {
		fprintf(stderr, "VA-API not supported\n");
		return 1;
	}

	AVFormatContext *input_ctx = NULL;
	if (avformat_open_input(&input_ctx, argv[1], NULL, NULL) != 0) {
		fprintf(stderr, "Failed to open input file\n");
		return 1;
	}

	if (avformat_find_stream_info(input_ctx, NULL) < 0) {
		fprintf(stderr, "Failed to find input stream info\n");
		return 1;
	}

	AVCodec *decoder = NULL;
	int video_stream = av_find_best_stream(input_ctx, AVMEDIA_TYPE_VIDEO,
		-1, -1, &decoder, 0);
	if (video_stream < 0) {
		fprintf(stderr, "Failed to find video stream in input file\n");
		return 1;
	}
	fprintf(stderr, "Selected decoder %s\n", decoder->name);

	int i = 0;
	while (1) {
		const AVCodecHWConfig *config = avcodec_get_hw_config(decoder, i);
		if (!config) {
			fprintf(stderr, "Decoder %s doesn't support device type %s\n",
				decoder->name, av_hwdevice_get_type_name(type));
			return 1;
		}

		if ((config->methods & AV_CODEC_HW_CONFIG_METHOD_HW_DEVICE_CTX) &&
				config->device_type == type) {
			hw_pix_fmt = config->pix_fmt;
			break;
		}

		i++;
	}
	fprintf(stderr, "Selected pixel format %s\n", av_get_pix_fmt_name(hw_pix_fmt));

	AVCodecContext *decoder_ctx = avcodec_alloc_context3(decoder);
	if (!decoder_ctx) {
		fprintf(stderr, "Failed to allocate decoder context\n");
		return 1;
	}

	AVStream *video = input_ctx->streams[video_stream];
	if (avcodec_parameters_to_context(decoder_ctx, video->codecpar) < 0) {
		fprintf(stderr, "Failed to fill decoder context from video parameters\n");
		return 1;
	}

	decoder_ctx->get_format = get_hw_format;

	AVBufferRef *hw_device_ctx = NULL;
	ret = av_hwdevice_ctx_create(&hw_device_ctx, type, NULL, NULL, 0);
	if (ret < 0) {
		fprintf(stderr, "Failed to create HW device context\n");
		return 1;
	}
	decoder_ctx->hw_device_ctx = av_buffer_ref(hw_device_ctx);

	ret = avcodec_open2(decoder_ctx, decoder, NULL);
	if (ret < 0) {
		fprintf(stderr, "Failed to open codec for stream #%u\n", video_stream);
		return 1;
	}

	AVFrame *frame = NULL;
	ret = decode_frame(input_ctx, video_stream, decoder_ctx, &frame);
	if (ret < 0) {
		return 1;
	}

	fprintf(stderr, "Got %dx%d frame!\n", frame->width, frame->height);

	AVFrame *drm_frame = av_frame_alloc();
	drm_frame->format = AV_PIX_FMT_DRM_PRIME;
	drm_frame->hw_frames_ctx = av_buffer_ref(frame->hw_frames_ctx);

	// Convert the VA-API frame into a DMA-BUF frame
	ret = av_hwframe_map(drm_frame, frame, 0);
	if (ret < 0) {
		fprintf(stderr, "Failed to map frame: %s\n", av_err2str(ret));
		return 1;
	}

	AVDRMFrameDescriptor *drm_frame_desc = (void *)drm_frame->data[0];
	fprintf(stderr, "DRM frame:\n");
	for (int i = 0; i < drm_frame_desc->nb_objects; i++) {
		AVDRMObjectDescriptor *drm_object = &drm_frame_desc->objects[i];
		fprintf(stderr, "  object #%i: fd = %d, size = %zu, "
			"format_modifier = 0x%lX\n", i, drm_object->fd,
			drm_object->size, drm_object->format_modifier);
	}
	for (int i = 0; i < drm_frame_desc->nb_layers; i++) {
		AVDRMLayerDescriptor *drm_layer = &drm_frame_desc->layers[i];
		fprintf(stderr, "  layer #%d: format = 0x%X\n",
			i, drm_layer->format);
		for (int j = 0; j < drm_layer->nb_planes; j++) {
			AVDRMPlaneDescriptor *drm_plane = &drm_layer->planes[j];
			fprintf(stderr, "    plane #%d: object_index = %d, "
				"offset = %ld, pitch = %ld\n", j, drm_plane->object_index,
				drm_plane->offset, drm_plane->pitch);
		}
	}

	for (int i = 0; i < drm_frame_desc->nb_objects; i++) {
		close(drm_frame_desc->objects[i].fd);
		drm_frame_desc->objects[i].fd = -1;
	}

	av_frame_free(&drm_frame);
	av_frame_free(&frame);

	avcodec_free_context(&decoder_ctx);
	avformat_close_input(&input_ctx);
	av_buffer_unref(&hw_device_ctx);

	return 0;
}