/* * This file is part of FFmpeg. * * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ #include "vp9dec.h" #include "vulkan_decode.h" const FFVulkanDecodeDescriptor ff_vk_dec_vp9_desc = { .codec_id = AV_CODEC_ID_VP9, .decode_extension = FF_VK_EXT_VIDEO_DECODE_VP9, .queue_flags = VK_QUEUE_VIDEO_DECODE_BIT_KHR, .decode_op = VK_VIDEO_CODEC_OPERATION_DECODE_VP9_BIT_KHR, .ext_props = { .extensionName = VK_STD_VULKAN_VIDEO_CODEC_VP9_DECODE_EXTENSION_NAME, .specVersion = VK_STD_VULKAN_VIDEO_CODEC_VP9_DECODE_SPEC_VERSION, }, }; typedef struct VP9VulkanDecodePicture { FFVulkanDecodePicture vp; /* TODO: investigate if this can be removed to make decoding completely * independent. */ FFVulkanDecodeContext *dec; /* Current picture */ StdVideoVP9ColorConfig color_config; StdVideoVP9Segmentation segmentation; StdVideoVP9LoopFilter loop_filter; StdVideoDecodeVP9PictureInfo std_pic_info; VkVideoDecodeVP9PictureInfoKHR vp9_pic_info; const VP9Frame *ref_src[8]; uint8_t frame_id_set; uint8_t frame_id; uint8_t ref_frame_sign_bias_mask; } VP9VulkanDecodePicture; static int vk_vp9_fill_pict(AVCodecContext *avctx, const VP9Frame **ref_src, VkVideoReferenceSlotInfoKHR *ref_slot, /* Main structure */ VkVideoPictureResourceInfoKHR *ref, /* Goes in ^ */ const VP9Frame *pic, int is_current) { FFVulkanDecodeContext *dec = avctx->internal->hwaccel_priv_data; FFVulkanDecodeShared *ctx = dec->shared_ctx; VP9VulkanDecodePicture *hp = pic->hwaccel_picture_private; FFVulkanDecodePicture *vkpic = &hp->vp; int err = ff_vk_decode_prepare_frame(dec, pic->tf.f, vkpic, is_current, dec->dedicated_dpb); if (err < 0) return err; *ref = (VkVideoPictureResourceInfoKHR) { .sType = VK_STRUCTURE_TYPE_VIDEO_PICTURE_RESOURCE_INFO_KHR, .codedOffset = (VkOffset2D){ 0, 0 }, .codedExtent = (VkExtent2D){ pic->tf.f->width, pic->tf.f->height }, .baseArrayLayer = (dec->dedicated_dpb && ctx->common.layered_dpb) ? hp->frame_id : 0, .imageViewBinding = vkpic->view.ref[0], }; *ref_slot = (VkVideoReferenceSlotInfoKHR) { .sType = VK_STRUCTURE_TYPE_VIDEO_REFERENCE_SLOT_INFO_KHR, .slotIndex = hp->frame_id, .pPictureResource = ref, }; if (ref_src) *ref_src = pic; return 0; } static enum StdVideoVP9InterpolationFilter remap_interp(uint8_t is_filter_switchable, uint8_t raw_interpolation_filter_type) { static const enum StdVideoVP9InterpolationFilter remap[] = { STD_VIDEO_VP9_INTERPOLATION_FILTER_EIGHTTAP_SMOOTH, STD_VIDEO_VP9_INTERPOLATION_FILTER_EIGHTTAP, STD_VIDEO_VP9_INTERPOLATION_FILTER_EIGHTTAP_SHARP, STD_VIDEO_VP9_INTERPOLATION_FILTER_BILINEAR, }; if (is_filter_switchable) return STD_VIDEO_VP9_INTERPOLATION_FILTER_SWITCHABLE; return remap[raw_interpolation_filter_type]; } static int vk_vp9_start_frame(AVCodecContext *avctx, av_unused const AVBufferRef *buffer_ref, av_unused const uint8_t *buffer, av_unused uint32_t size) { int err; int ref_count = 0; const VP9Context *priv = avctx->priv_data; const CodedBitstreamVP9Context *vp9 = priv->cbc->priv_data; const VP9SharedContext *s = &priv->s; uint32_t frame_id_alloc_mask = 0; const VP9Frame *pic = &s->frames[CUR_FRAME]; FFVulkanDecodeContext *dec = avctx->internal->hwaccel_priv_data; uint8_t profile = (pic->frame_header->profile_high_bit << 1) | pic->frame_header->profile_low_bit; VP9VulkanDecodePicture *ap = pic->hwaccel_picture_private; FFVulkanDecodePicture *vp = &ap->vp; /* Use the current frame_ids in ref_frames[] to decide occupied frame_ids */ for (int i = 0; i < STD_VIDEO_VP9_NUM_REF_FRAMES; i++) { const VP9VulkanDecodePicture* rp = s->ref_frames[i].hwaccel_picture_private; if (rp) frame_id_alloc_mask |= 1 << rp->frame_id; } if (!ap->frame_id_set) { unsigned slot_idx = 0; for (unsigned i = 0; i < 32; i++) { if (!(frame_id_alloc_mask & (1 << i))) { slot_idx = i; break; } } ap->frame_id = slot_idx; ap->frame_id_set = 1; frame_id_alloc_mask |= (1 << slot_idx); } for (int i = 0; i < STD_VIDEO_VP9_REFS_PER_FRAME; i++) { const int idx = pic->frame_header->ref_frame_idx[i]; const VP9Frame *ref_frame = &s->ref_frames[idx]; VP9VulkanDecodePicture *hp = ref_frame->hwaccel_picture_private; int found = 0; if (!ref_frame->tf.f) continue; for (int j = 0; j < ref_count; j++) { if (vp->ref_slots[j].slotIndex == hp->frame_id) { found = 1; break; } } if (found) continue; err = vk_vp9_fill_pict(avctx, &ap->ref_src[ref_count], &vp->ref_slots[ref_count], &vp->refs[ref_count], ref_frame, 0); if (err < 0) return err; ref_count++; } err = vk_vp9_fill_pict(avctx, NULL, &vp->ref_slot, &vp->ref, pic, 1); if (err < 0) return err; ap->loop_filter = (StdVideoVP9LoopFilter) { .flags = (StdVideoVP9LoopFilterFlags) { .loop_filter_delta_enabled = pic->frame_header->loop_filter_delta_enabled, .loop_filter_delta_update = pic->frame_header->loop_filter_delta_update, }, .loop_filter_level = pic->frame_header->loop_filter_level, .loop_filter_sharpness = pic->frame_header->loop_filter_sharpness, .update_ref_delta = 0x0, .update_mode_delta = 0x0, }; for (int i = 0; i < STD_VIDEO_VP9_MAX_REF_FRAMES; i++) { ap->loop_filter.loop_filter_ref_deltas[i] = vp9->loop_filter_ref_deltas[i]; ap->loop_filter.update_ref_delta |= pic->frame_header->update_ref_delta[i]; } for (int i = 0; i < STD_VIDEO_VP9_LOOP_FILTER_ADJUSTMENTS; i++) { ap->loop_filter.loop_filter_mode_deltas[i] = vp9->loop_filter_mode_deltas[i]; ap->loop_filter.update_mode_delta |= pic->frame_header->update_mode_delta[i]; } ap->segmentation = (StdVideoVP9Segmentation) { .flags = (StdVideoVP9SegmentationFlags) { .segmentation_update_map = pic->frame_header->segmentation_update_map, .segmentation_temporal_update = pic->frame_header->segmentation_temporal_update, .segmentation_update_data = pic->frame_header->segmentation_update_data, .segmentation_abs_or_delta_update = pic->frame_header->segmentation_abs_or_delta_update, }, }; for (int i = 0; i < STD_VIDEO_VP9_MAX_SEGMENTATION_TREE_PROBS; i++) ap->segmentation.segmentation_tree_probs[i] = vp9->segmentation_tree_probs[i]; for (int i = 0; i < STD_VIDEO_VP9_MAX_SEGMENTATION_PRED_PROB; i++) ap->segmentation.segmentation_pred_prob[i] = vp9->segmentation_pred_prob[i]; for (int i = 0; i < STD_VIDEO_VP9_MAX_SEGMENTS; i++) { ap->segmentation.FeatureEnabled[i] = 0x0; for (int j = 0; j < STD_VIDEO_VP9_SEG_LVL_MAX; j++) { ap->segmentation.FeatureEnabled[i] |= vp9->feature_enabled[i][j] << j; ap->segmentation.FeatureData[i][j] = vp9->feature_sign[i][j] ? -vp9->feature_value[i][j] : +vp9->feature_value[i][j]; } } ap->color_config = (StdVideoVP9ColorConfig) { .flags = (StdVideoVP9ColorConfigFlags) { .color_range = pic->frame_header->color_range, }, .BitDepth = profile < 2 ? 8 : pic->frame_header->ten_or_twelve_bit ? 12 : 10, .subsampling_x = pic->frame_header->subsampling_x, .subsampling_y = pic->frame_header->subsampling_y, .color_space = pic->frame_header->color_space, }; ap->std_pic_info = (StdVideoDecodeVP9PictureInfo) { .flags = (StdVideoDecodeVP9PictureInfoFlags) { .error_resilient_mode = pic->frame_header->error_resilient_mode, .intra_only = pic->frame_header->intra_only, .allow_high_precision_mv = pic->frame_header->allow_high_precision_mv, .refresh_frame_context = pic->frame_header->refresh_frame_context, .frame_parallel_decoding_mode = pic->frame_header->frame_parallel_decoding_mode, .segmentation_enabled = pic->frame_header->segmentation_enabled, .show_frame = pic->frame_header->segmentation_enabled, .UsePrevFrameMvs = s->h.use_last_frame_mvs, }, .profile = profile, .frame_type = pic->frame_header->frame_type, .frame_context_idx = pic->frame_header->frame_context_idx, .reset_frame_context = pic->frame_header->reset_frame_context, .refresh_frame_flags = pic->frame_header->refresh_frame_flags, .ref_frame_sign_bias_mask = 0x0, .interpolation_filter = remap_interp(pic->frame_header->is_filter_switchable, pic->frame_header->raw_interpolation_filter_type), .base_q_idx = pic->frame_header->base_q_idx, .delta_q_y_dc = pic->frame_header->delta_q_y_dc, .delta_q_uv_dc = pic->frame_header->delta_q_uv_dc, .delta_q_uv_ac = pic->frame_header->delta_q_uv_ac, .tile_cols_log2 = pic->frame_header->tile_cols_log2, .tile_rows_log2 = pic->frame_header->tile_rows_log2, /* Reserved */ .pColorConfig = &ap->color_config, .pLoopFilter = &ap->loop_filter, .pSegmentation = &ap->segmentation, }; for (int i = VP9_LAST_FRAME; i <= VP9_ALTREF_FRAME; i++) ap->std_pic_info.ref_frame_sign_bias_mask |= pic->frame_header->ref_frame_sign_bias[i] << i; ap->vp9_pic_info = (VkVideoDecodeVP9PictureInfoKHR) { .sType = VK_STRUCTURE_TYPE_VIDEO_DECODE_VP9_PICTURE_INFO_KHR, .pStdPictureInfo = &ap->std_pic_info, .uncompressedHeaderOffset = 0, .compressedHeaderOffset = s->h.uncompressed_header_size, .tilesOffset = s->h.uncompressed_header_size + s->h.compressed_header_size, }; for (int i = 0; i < STD_VIDEO_VP9_REFS_PER_FRAME; i++) { const int idx = pic->frame_header->ref_frame_idx[i]; const VP9Frame *ref_frame = &s->ref_frames[idx]; VP9VulkanDecodePicture *hp = ref_frame->hwaccel_picture_private; if (!ref_frame->tf.f) ap->vp9_pic_info.referenceNameSlotIndices[i] = -1; else ap->vp9_pic_info.referenceNameSlotIndices[i] = hp->frame_id; } vp->decode_info = (VkVideoDecodeInfoKHR) { .sType = VK_STRUCTURE_TYPE_VIDEO_DECODE_INFO_KHR, .pNext = &ap->vp9_pic_info, .flags = 0x0, .pSetupReferenceSlot = &vp->ref_slot, .referenceSlotCount = ref_count, .pReferenceSlots = vp->ref_slots, .dstPictureResource = (VkVideoPictureResourceInfoKHR) { .sType = VK_STRUCTURE_TYPE_VIDEO_PICTURE_RESOURCE_INFO_KHR, .codedOffset = (VkOffset2D){ 0, 0 }, .codedExtent = (VkExtent2D){ pic->tf.f->width, pic->tf.f->height }, .baseArrayLayer = 0, .imageViewBinding = vp->view.out[0], }, }; ap->dec = dec; return 0; } static int vk_vp9_decode_slice(AVCodecContext *avctx, const uint8_t *data, uint32_t size) { int err; const VP9SharedContext *s = avctx->priv_data; VP9VulkanDecodePicture *ap = s->frames[CUR_FRAME].hwaccel_picture_private; FFVulkanDecodePicture *vp = &ap->vp; err = ff_vk_decode_add_slice(avctx, vp, data, size, 0, NULL, NULL); if (err < 0) return err; return 0; } static int vk_vp9_end_frame(AVCodecContext *avctx) { const VP9SharedContext *s = avctx->priv_data; const VP9Frame *pic = &s->frames[CUR_FRAME]; VP9VulkanDecodePicture *ap = pic->hwaccel_picture_private; FFVulkanDecodePicture *vp = &ap->vp; FFVulkanDecodePicture *rvp[STD_VIDEO_VP9_REFS_PER_FRAME] = { 0 }; AVFrame *rav[STD_VIDEO_VP9_REFS_PER_FRAME] = { 0 }; for (int i = 0; i < vp->decode_info.referenceSlotCount; i++) { const VP9Frame *rp = ap->ref_src[i]; VP9VulkanDecodePicture *rhp = rp->hwaccel_picture_private; rvp[i] = &rhp->vp; rav[i] = ap->ref_src[i]->tf.f; } av_log(avctx, AV_LOG_VERBOSE, "Decoding frame, %"SIZE_SPECIFIER" bytes\n", vp->slices_size); return ff_vk_decode_frame(avctx, pic->tf.f, vp, rav, rvp); } static void vk_vp9_free_frame_priv(AVRefStructOpaque _hwctx, void *data) { AVHWDeviceContext *hwctx = _hwctx.nc; VP9VulkanDecodePicture *ap = data; /* Free frame resources, this also destroys the session parameters. */ ff_vk_decode_free_frame(hwctx, &ap->vp); } const FFHWAccel ff_vp9_vulkan_hwaccel = { .p.name = "vp9_vulkan", .p.type = AVMEDIA_TYPE_VIDEO, .p.id = AV_CODEC_ID_VP9, .p.pix_fmt = AV_PIX_FMT_VULKAN, .start_frame = &vk_vp9_start_frame, .decode_slice = &vk_vp9_decode_slice, .end_frame = &vk_vp9_end_frame, .free_frame_priv = &vk_vp9_free_frame_priv, .frame_priv_data_size = sizeof(VP9VulkanDecodePicture), .init = &ff_vk_decode_init, .update_thread_context = &ff_vk_update_thread_context, .flush = &ff_vk_decode_flush, .uninit = &ff_vk_decode_uninit, .frame_params = &ff_vk_frame_params, .priv_data_size = sizeof(FFVulkanDecodeContext), .caps_internal = HWACCEL_CAP_ASYNC_SAFE, };