Merge commit 'a0f2946068c62e18cb05ac25c0df3d86077251a6'
[ffmpeg.git] / libavcodec / vdpau.c
index 8606624..0c38935 100644 (file)
@@ -4,20 +4,20 @@
  *
  * Copyright (c) 2008 NVIDIA
  *
- * This file is part of Libav.
+ * This file is part of FFmpeg.
  *
- * Libav is free software; you can redistribute it and/or
+ * 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.
  *
- * Libav is distributed in the hope that it will be useful,
+ * 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 Libav; if not, write to the Free Software
+ * License along with FFmpeg; if not, write to the Free Software
  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  */
 
@@ -64,6 +64,13 @@ static int vdpau_error(VdpStatus status)
     }
 }
 
+AVVDPAUContext *av_alloc_vdpaucontext(void)
+{
+    return av_vdpau_alloc_context();
+}
+
+MAKE_ACCESSORS(AVVDPAUContext, vdpau_hwaccel, AVVDPAU_Render2, render2)
+
 int av_vdpau_get_surface_parameters(AVCodecContext *avctx,
                                     VdpChromaType *type,
                                     uint32_t *width, uint32_t *height)
@@ -122,7 +129,12 @@ int ff_vdpau_common_init(AVCodecContext *avctx, VdpDecoderProfile profile,
 
     vdctx->width            = UINT32_MAX;
     vdctx->height           = UINT32_MAX;
-    hwctx->reset            = 0;
+
+    if (!hwctx) {
+        vdctx->device  = VDP_INVALID_HANDLE;
+        av_log(avctx, AV_LOG_WARNING, "hwaccel_context has not been setup by the user application, cannot initialize\n");
+        return 0;
+    }
 
     if (hwctx->context.decoder != VDP_INVALID_HANDLE) {
         vdctx->decoder = hwctx->context.decoder;
@@ -130,6 +142,7 @@ int ff_vdpau_common_init(AVCodecContext *avctx, VdpDecoderProfile profile,
         vdctx->device  = VDP_INVALID_HANDLE;
         return 0; /* Decoder created by user */
     }
+    hwctx->reset            = 0;
 
     vdctx->device           = hwctx->device;
     vdctx->get_proc_address = hwctx->get_proc_address;
@@ -264,6 +277,7 @@ int ff_vdpau_common_end_frame(AVCodecContext *avctx, AVFrame *frame,
                               struct vdpau_picture_context *pic_ctx)
 {
     VDPAUContext *vdctx = avctx->internal->hwaccel_priv_data;
+    AVVDPAUContext *hwctx = avctx->hwaccel_context;
     VdpVideoSurface surf = ff_vdpau_get_surface_id(frame);
     VdpStatus status;
     int val;
@@ -272,11 +286,34 @@ int ff_vdpau_common_end_frame(AVCodecContext *avctx, AVFrame *frame,
     if (val < 0)
         return val;
 
+#if FF_API_BUFS_VDPAU
+FF_DISABLE_DEPRECATION_WARNINGS
+    av_assert0(sizeof(hwctx->info) <= sizeof(pic_ctx->info));
+    memcpy(&hwctx->info, &pic_ctx->info, sizeof(hwctx->info));
+    hwctx->bitstream_buffers = pic_ctx->bitstream_buffers;
+    hwctx->bitstream_buffers_used = pic_ctx->bitstream_buffers_used;
+    hwctx->bitstream_buffers_allocated = pic_ctx->bitstream_buffers_allocated;
+FF_ENABLE_DEPRECATION_WARNINGS
+#endif
+
+    if (!hwctx->render && hwctx->render2) {
+        status = hwctx->render2(avctx, frame, (void *)&pic_ctx->info,
+                                pic_ctx->bitstream_buffers_used, pic_ctx->bitstream_buffers);
+    } else
     status = vdctx->render(vdctx->decoder, surf, (void *)&pic_ctx->info,
                            pic_ctx->bitstream_buffers_used,
                            pic_ctx->bitstream_buffers);
 
     av_freep(&pic_ctx->bitstream_buffers);
+
+#if FF_API_BUFS_VDPAU
+FF_DISABLE_DEPRECATION_WARNINGS
+    hwctx->bitstream_buffers = NULL;
+    hwctx->bitstream_buffers_used = 0;
+    hwctx->bitstream_buffers_allocated = 0;
+FF_ENABLE_DEPRECATION_WARNINGS
+#endif
+
     return vdpau_error(status);
 }
 
@@ -318,6 +355,343 @@ int ff_vdpau_add_buffer(struct vdpau_picture_context *pic_ctx,
     return 0;
 }
 
+/* Obsolete non-hwaccel VDPAU support below... */
+
+void ff_vdpau_h264_set_reference_frames(H264Context *h)
+{
+    struct vdpau_render_state *render, *render_ref;
+    VdpReferenceFrameH264 *rf, *rf2;
+    H264Picture *pic;
+    int i, list, pic_frame_idx;
+
+    render = (struct vdpau_render_state *)h->cur_pic_ptr->f->data[0];
+    assert(render);
+
+    rf = &render->info.h264.referenceFrames[0];
+#define H264_RF_COUNT FF_ARRAY_ELEMS(render->info.h264.referenceFrames)
+
+    for (list = 0; list < 2; ++list) {
+        H264Picture **lp = list ? h->long_ref : h->short_ref;
+        int ls = list ? 16 : h->short_ref_count;
+
+        for (i = 0; i < ls; ++i) {
+            pic = lp[i];
+            if (!pic || !pic->reference)
+                continue;
+            pic_frame_idx = pic->long_ref ? pic->pic_id : pic->frame_num;
+
+            render_ref = (struct vdpau_render_state *)pic->f->data[0];
+            assert(render_ref);
+
+            rf2 = &render->info.h264.referenceFrames[0];
+            while (rf2 != rf) {
+                if (
+                    (rf2->surface == render_ref->surface)
+                    && (rf2->is_long_term == pic->long_ref)
+                    && (rf2->frame_idx == pic_frame_idx)
+                )
+                    break;
+                ++rf2;
+            }
+            if (rf2 != rf) {
+                rf2->top_is_reference    |= (pic->reference & PICT_TOP_FIELD)    ? VDP_TRUE : VDP_FALSE;
+                rf2->bottom_is_reference |= (pic->reference & PICT_BOTTOM_FIELD) ? VDP_TRUE : VDP_FALSE;
+                continue;
+            }
+
+            if (rf >= &render->info.h264.referenceFrames[H264_RF_COUNT])
+                continue;
+
+            rf->surface             = render_ref->surface;
+            rf->is_long_term        = pic->long_ref;
+            rf->top_is_reference    = (pic->reference & PICT_TOP_FIELD)    ? VDP_TRUE : VDP_FALSE;
+            rf->bottom_is_reference = (pic->reference & PICT_BOTTOM_FIELD) ? VDP_TRUE : VDP_FALSE;
+            rf->field_order_cnt[0]  = pic->field_poc[0];
+            rf->field_order_cnt[1]  = pic->field_poc[1];
+            rf->frame_idx           = pic_frame_idx;
+
+            ++rf;
+        }
+    }
+
+    for (; rf < &render->info.h264.referenceFrames[H264_RF_COUNT]; ++rf) {
+        rf->surface             = VDP_INVALID_HANDLE;
+        rf->is_long_term        = 0;
+        rf->top_is_reference    = 0;
+        rf->bottom_is_reference = 0;
+        rf->field_order_cnt[0]  = 0;
+        rf->field_order_cnt[1]  = 0;
+        rf->frame_idx           = 0;
+    }
+}
+
+void ff_vdpau_add_data_chunk(uint8_t *data, const uint8_t *buf, int buf_size)
+{
+    struct vdpau_render_state *render = (struct vdpau_render_state*)data;
+    assert(render);
+
+    render->bitstream_buffers= av_fast_realloc(
+        render->bitstream_buffers,
+        &render->bitstream_buffers_allocated,
+        sizeof(*render->bitstream_buffers)*(render->bitstream_buffers_used + 1)
+    );
+
+    render->bitstream_buffers[render->bitstream_buffers_used].struct_version  = VDP_BITSTREAM_BUFFER_VERSION;
+    render->bitstream_buffers[render->bitstream_buffers_used].bitstream       = buf;
+    render->bitstream_buffers[render->bitstream_buffers_used].bitstream_bytes = buf_size;
+    render->bitstream_buffers_used++;
+}
+
+#if CONFIG_H264_VDPAU_DECODER
+void ff_vdpau_h264_picture_start(H264Context *h)
+{
+    struct vdpau_render_state *render;
+    int i;
+
+    render = (struct vdpau_render_state *)h->cur_pic_ptr->f->data[0];
+    assert(render);
+
+    for (i = 0; i < 2; ++i) {
+        int foc = h->cur_pic_ptr->field_poc[i];
+        if (foc == INT_MAX)
+            foc = 0;
+        render->info.h264.field_order_cnt[i] = foc;
+    }
+
+    render->info.h264.frame_num = h->frame_num;
+}
+
+void ff_vdpau_h264_picture_complete(H264Context *h)
+{
+    struct vdpau_render_state *render;
+
+    render = (struct vdpau_render_state *)h->cur_pic_ptr->f->data[0];
+    assert(render);
+
+    render->info.h264.slice_count = h->current_slice;
+    if (render->info.h264.slice_count < 1)
+        return;
+
+    render->info.h264.is_reference                           = (h->cur_pic_ptr->reference & 3) ? VDP_TRUE : VDP_FALSE;
+    render->info.h264.field_pic_flag                         = h->picture_structure != PICT_FRAME;
+    render->info.h264.bottom_field_flag                      = h->picture_structure == PICT_BOTTOM_FIELD;
+    render->info.h264.num_ref_frames                         = h->sps.ref_frame_count;
+    render->info.h264.mb_adaptive_frame_field_flag           = h->sps.mb_aff && !render->info.h264.field_pic_flag;
+    render->info.h264.constrained_intra_pred_flag            = h->pps.constrained_intra_pred;
+    render->info.h264.weighted_pred_flag                     = h->pps.weighted_pred;
+    render->info.h264.weighted_bipred_idc                    = h->pps.weighted_bipred_idc;
+    render->info.h264.frame_mbs_only_flag                    = h->sps.frame_mbs_only_flag;
+    render->info.h264.transform_8x8_mode_flag                = h->pps.transform_8x8_mode;
+    render->info.h264.chroma_qp_index_offset                 = h->pps.chroma_qp_index_offset[0];
+    render->info.h264.second_chroma_qp_index_offset          = h->pps.chroma_qp_index_offset[1];
+    render->info.h264.pic_init_qp_minus26                    = h->pps.init_qp - 26;
+    render->info.h264.num_ref_idx_l0_active_minus1           = h->pps.ref_count[0] - 1;
+    render->info.h264.num_ref_idx_l1_active_minus1           = h->pps.ref_count[1] - 1;
+    render->info.h264.log2_max_frame_num_minus4              = h->sps.log2_max_frame_num - 4;
+    render->info.h264.pic_order_cnt_type                     = h->sps.poc_type;
+    render->info.h264.log2_max_pic_order_cnt_lsb_minus4      = h->sps.poc_type ? 0 : h->sps.log2_max_poc_lsb - 4;
+    render->info.h264.delta_pic_order_always_zero_flag       = h->sps.delta_pic_order_always_zero_flag;
+    render->info.h264.direct_8x8_inference_flag              = h->sps.direct_8x8_inference_flag;
+    render->info.h264.entropy_coding_mode_flag               = h->pps.cabac;
+    render->info.h264.pic_order_present_flag                 = h->pps.pic_order_present;
+    render->info.h264.deblocking_filter_control_present_flag = h->pps.deblocking_filter_parameters_present;
+    render->info.h264.redundant_pic_cnt_present_flag         = h->pps.redundant_pic_cnt_present;
+    memcpy(render->info.h264.scaling_lists_4x4, h->pps.scaling_matrix4, sizeof(render->info.h264.scaling_lists_4x4));
+    memcpy(render->info.h264.scaling_lists_8x8[0], h->pps.scaling_matrix8[0], sizeof(render->info.h264.scaling_lists_8x8[0]));
+    memcpy(render->info.h264.scaling_lists_8x8[1], h->pps.scaling_matrix8[3], sizeof(render->info.h264.scaling_lists_8x8[0]));
+
+    ff_h264_draw_horiz_band(h, &h->slice_ctx[0], 0, h->avctx->height);
+    render->bitstream_buffers_used = 0;
+}
+#endif /* CONFIG_H264_VDPAU_DECODER */
+
+#if CONFIG_MPEG_VDPAU_DECODER || CONFIG_MPEG1_VDPAU_DECODER
+void ff_vdpau_mpeg_picture_complete(MpegEncContext *s, const uint8_t *buf,
+                                    int buf_size, int slice_count)
+{
+    struct vdpau_render_state *render, *last, *next;
+    int i;
+
+    if (!s->current_picture_ptr) return;
+
+    render = (struct vdpau_render_state *)s->current_picture_ptr->f->data[0];
+    assert(render);
+
+    /* fill VdpPictureInfoMPEG1Or2 struct */
+    render->info.mpeg.picture_structure          = s->picture_structure;
+    render->info.mpeg.picture_coding_type        = s->pict_type;
+    render->info.mpeg.intra_dc_precision         = s->intra_dc_precision;
+    render->info.mpeg.frame_pred_frame_dct       = s->frame_pred_frame_dct;
+    render->info.mpeg.concealment_motion_vectors = s->concealment_motion_vectors;
+    render->info.mpeg.intra_vlc_format           = s->intra_vlc_format;
+    render->info.mpeg.alternate_scan             = s->alternate_scan;
+    render->info.mpeg.q_scale_type               = s->q_scale_type;
+    render->info.mpeg.top_field_first            = s->top_field_first;
+    render->info.mpeg.full_pel_forward_vector    = s->full_pel[0]; // MPEG-1 only.  Set 0 for MPEG-2
+    render->info.mpeg.full_pel_backward_vector   = s->full_pel[1]; // MPEG-1 only.  Set 0 for MPEG-2
+    render->info.mpeg.f_code[0][0]               = s->mpeg_f_code[0][0]; // For MPEG-1 fill both horiz. & vert.
+    render->info.mpeg.f_code[0][1]               = s->mpeg_f_code[0][1];
+    render->info.mpeg.f_code[1][0]               = s->mpeg_f_code[1][0];
+    render->info.mpeg.f_code[1][1]               = s->mpeg_f_code[1][1];
+    for (i = 0; i < 64; ++i) {
+        render->info.mpeg.intra_quantizer_matrix[i]     = s->intra_matrix[i];
+        render->info.mpeg.non_intra_quantizer_matrix[i] = s->inter_matrix[i];
+    }
+
+    render->info.mpeg.forward_reference          = VDP_INVALID_HANDLE;
+    render->info.mpeg.backward_reference         = VDP_INVALID_HANDLE;
+
+    switch(s->pict_type){
+    case  AV_PICTURE_TYPE_B:
+        next = (struct vdpau_render_state *)s->next_picture.f->data[0];
+        assert(next);
+        render->info.mpeg.backward_reference     = next->surface;
+        // no return here, going to set forward prediction
+    case  AV_PICTURE_TYPE_P:
+        last = (struct vdpau_render_state *)s->last_picture.f->data[0];
+        if (!last) // FIXME: Does this test make sense?
+            last = render; // predict second field from the first
+        render->info.mpeg.forward_reference      = last->surface;
+    }
+
+    ff_vdpau_add_data_chunk(s->current_picture_ptr->f->data[0], buf, buf_size);
+
+    render->info.mpeg.slice_count                = slice_count;
+
+    if (slice_count)
+        ff_mpeg_draw_horiz_band(s, 0, s->avctx->height);
+    render->bitstream_buffers_used               = 0;
+}
+#endif /* CONFIG_MPEG_VDPAU_DECODER || CONFIG_MPEG1_VDPAU_DECODER */
+
+#if CONFIG_VC1_VDPAU_DECODER
+void ff_vdpau_vc1_decode_picture(MpegEncContext *s, const uint8_t *buf,
+                                 int buf_size)
+{
+    VC1Context *v = s->avctx->priv_data;
+    struct vdpau_render_state *render, *last, *next;
+
+    render = (struct vdpau_render_state *)s->current_picture.f->data[0];
+    assert(render);
+
+    /*  fill LvPictureInfoVC1 struct */
+    render->info.vc1.frame_coding_mode  = v->fcm ? v->fcm + 1 : 0;
+    render->info.vc1.postprocflag       = v->postprocflag;
+    render->info.vc1.pulldown           = v->broadcast;
+    render->info.vc1.interlace          = v->interlace;
+    render->info.vc1.tfcntrflag         = v->tfcntrflag;
+    render->info.vc1.finterpflag        = v->finterpflag;
+    render->info.vc1.psf                = v->psf;
+    render->info.vc1.dquant             = v->dquant;
+    render->info.vc1.panscan_flag       = v->panscanflag;
+    render->info.vc1.refdist_flag       = v->refdist_flag;
+    render->info.vc1.quantizer          = v->quantizer_mode;
+    render->info.vc1.extended_mv        = v->extended_mv;
+    render->info.vc1.extended_dmv       = v->extended_dmv;
+    render->info.vc1.overlap            = v->overlap;
+    render->info.vc1.vstransform        = v->vstransform;
+    render->info.vc1.loopfilter         = v->s.loop_filter;
+    render->info.vc1.fastuvmc           = v->fastuvmc;
+    render->info.vc1.range_mapy_flag    = v->range_mapy_flag;
+    render->info.vc1.range_mapy         = v->range_mapy;
+    render->info.vc1.range_mapuv_flag   = v->range_mapuv_flag;
+    render->info.vc1.range_mapuv        = v->range_mapuv;
+    /* Specific to simple/main profile only */
+    render->info.vc1.multires           = v->multires;
+    render->info.vc1.syncmarker         = v->resync_marker;
+    render->info.vc1.rangered           = v->rangered | (v->rangeredfrm << 1);
+    render->info.vc1.maxbframes         = v->s.max_b_frames;
+
+    render->info.vc1.deblockEnable      = v->postprocflag & 1;
+    render->info.vc1.pquant             = v->pq;
+
+    render->info.vc1.forward_reference  = VDP_INVALID_HANDLE;
+    render->info.vc1.backward_reference = VDP_INVALID_HANDLE;
+
+    if (v->bi_type)
+        render->info.vc1.picture_type = 4;
+    else
+        render->info.vc1.picture_type = s->pict_type - 1 + s->pict_type / 3;
+
+    switch(s->pict_type){
+    case  AV_PICTURE_TYPE_B:
+        next = (struct vdpau_render_state *)s->next_picture.f->data[0];
+        assert(next);
+        render->info.vc1.backward_reference = next->surface;
+        // no break here, going to set forward prediction
+    case  AV_PICTURE_TYPE_P:
+        last = (struct vdpau_render_state *)s->last_picture.f->data[0];
+        if (!last) // FIXME: Does this test make sense?
+            last = render; // predict second field from the first
+        render->info.vc1.forward_reference = last->surface;
+    }
+
+    ff_vdpau_add_data_chunk(s->current_picture_ptr->f->data[0], buf, buf_size);
+
+    render->info.vc1.slice_count          = 1;
+
+    ff_mpeg_draw_horiz_band(s, 0, s->avctx->height);
+    render->bitstream_buffers_used        = 0;
+}
+#endif /* (CONFIG_VC1_VDPAU_DECODER */
+
+#if CONFIG_MPEG4_VDPAU_DECODER
+void ff_vdpau_mpeg4_decode_picture(Mpeg4DecContext *ctx, const uint8_t *buf,
+                                   int buf_size)
+{
+    MpegEncContext *s = &ctx->m;
+    struct vdpau_render_state *render, *last, *next;
+    int i;
+
+    if (!s->current_picture_ptr) return;
+
+    render = (struct vdpau_render_state *)s->current_picture_ptr->f->data[0];
+    assert(render);
+
+    /* fill VdpPictureInfoMPEG4Part2 struct */
+    render->info.mpeg4.trd[0]                            = s->pp_time;
+    render->info.mpeg4.trb[0]                            = s->pb_time;
+    render->info.mpeg4.trd[1]                            = s->pp_field_time >> 1;
+    render->info.mpeg4.trb[1]                            = s->pb_field_time >> 1;
+    render->info.mpeg4.vop_time_increment_resolution     = s->avctx->time_base.den;
+    render->info.mpeg4.vop_coding_type                   = 0;
+    render->info.mpeg4.vop_fcode_forward                 = s->f_code;
+    render->info.mpeg4.vop_fcode_backward                = s->b_code;
+    render->info.mpeg4.resync_marker_disable             = !ctx->resync_marker;
+    render->info.mpeg4.interlaced                        = !s->progressive_sequence;
+    render->info.mpeg4.quant_type                        = s->mpeg_quant;
+    render->info.mpeg4.quarter_sample                    = s->quarter_sample;
+    render->info.mpeg4.short_video_header                = s->avctx->codec->id == AV_CODEC_ID_H263;
+    render->info.mpeg4.rounding_control                  = s->no_rounding;
+    render->info.mpeg4.alternate_vertical_scan_flag      = s->alternate_scan;
+    render->info.mpeg4.top_field_first                   = s->top_field_first;
+    for (i = 0; i < 64; ++i) {
+        render->info.mpeg4.intra_quantizer_matrix[i]     = s->intra_matrix[i];
+        render->info.mpeg4.non_intra_quantizer_matrix[i] = s->inter_matrix[i];
+    }
+    render->info.mpeg4.forward_reference                 = VDP_INVALID_HANDLE;
+    render->info.mpeg4.backward_reference                = VDP_INVALID_HANDLE;
+
+    switch (s->pict_type) {
+    case AV_PICTURE_TYPE_B:
+        next = (struct vdpau_render_state *)s->next_picture.f->data[0];
+        assert(next);
+        render->info.mpeg4.backward_reference     = next->surface;
+        render->info.mpeg4.vop_coding_type        = 2;
+        // no break here, going to set forward prediction
+    case AV_PICTURE_TYPE_P:
+        last = (struct vdpau_render_state *)s->last_picture.f->data[0];
+        assert(last);
+        render->info.mpeg4.forward_reference      = last->surface;
+    }
+
+    ff_vdpau_add_data_chunk(s->current_picture_ptr->f->data[0], buf, buf_size);
+
+    ff_mpeg_draw_horiz_band(s, 0, s->avctx->height);
+    render->bitstream_buffers_used = 0;
+}
+#endif /* CONFIG_MPEG4_VDPAU_DECODER */
+
 int av_vdpau_get_profile(AVCodecContext *avctx, VdpDecoderProfile *profile)
 {
 #define PROFILE(prof)                      \