avcodec: add new Videotoolbox hwaccel.

21 files changed, 1113 insertions, 289 deletions

diff --git a/Changelog b/Changelog
index d39c1e5e9e..dc37343c5a 100644
--- a/Changelog
+++ b/Changelog
@@ -29,6 +29,7 @@ version <next>:
 - acrossfade audio filter
 - allyuv video source
 - atadenoise video filter
+- OS X VideoToolbox support
 
 
 version 2.7:
diff --git a/MAINTAINERS b/MAINTAINERS
index 89083bdd69..da02165137 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -305,6 +305,7 @@ Hardware acceleration:
   vaapi*                                Gwenole Beauchesne
   vda*                                  Sebastien Zwickert
   vdpau*                                Carl Eugen Hoyos
+  videotoolbox*                         Sebastien Zwickert
 
 
 libavdevice
diff --git a/Makefile b/Makefile
index 4e788cdae3..b0ab3ed0cc 100644
--- a/Makefile
+++ b/Makefile
@@ -31,7 +31,10 @@ $(foreach prog,$(AVBASENAMES),$(eval OBJS-$(prog)-$(CONFIG_OPENCL) += cmdutils_o
 OBJS-ffmpeg                   += ffmpeg_opt.o ffmpeg_filter.o
 OBJS-ffmpeg-$(HAVE_VDPAU_X11) += ffmpeg_vdpau.o
 OBJS-ffmpeg-$(HAVE_DXVA2_LIB) += ffmpeg_dxva2.o
-OBJS-ffmpeg-$(CONFIG_VDA)     += ffmpeg_vda.o
+ifndef CONFIG_VIDEOTOOLBOX
+OBJS-ffmpeg-$(CONFIG_VDA)     += ffmpeg_videotoolbox.o
+endif
+OBJS-ffmpeg-$(CONFIG_VIDEOTOOLBOX) += ffmpeg_videotoolbox.o
 OBJS-ffserver                 += ffserver_config.o
 
 TESTTOOLS   = audiogen videogen rotozoom tiny_psnr tiny_ssim base64
diff --git a/configure b/configure
index f3550a09c3..0f126bd7e0 100755
--- a/configure
+++ b/configure
@@ -155,6 +155,7 @@ Hardware accelerators:
   --disable-vaapi          disable VAAPI code [autodetect]
   --disable-vda            disable VDA code [autodetect]
   --disable-vdpau          disable VDPAU code [autodetect]
+  --enable-videotoolbox    enable VideoToolbox code [autodetect]
 
 Individual component options:
   --disable-everything     disable all components listed below
@@ -1470,6 +1471,7 @@ HWACCEL_LIST="
     vaapi
     vda
     vdpau
+    videotoolbox
     xvmc
 "
 
@@ -2380,14 +2382,18 @@ d3d11va_deps="d3d11_h dxva_h ID3D11VideoDecoder"
 dxva2_deps="dxva2api_h DXVA2_ConfigPictureDecode"
 vaapi_deps="va_va_h"
 vda_deps="VideoDecodeAcceleration_VDADecoder_h pthreads"
-vda_extralibs="-framework CoreFoundation -framework VideoDecodeAcceleration -framework QuartzCore"
+vda_extralibs="-framework CoreFoundation -framework VideoDecodeAcceleration -framework QuartzCore -framework CoreServices"
 vdpau_deps="vdpau_vdpau_h vdpau_vdpau_x11_h"
+videotoolbox_deps="VideoToolbox_VideoToolbox_h pthreads"
+videotoolbox_extralibs="-framework CoreFoundation -framework VideoToolbox -framework CoreMedia -framework QuartzCore -framework CoreServices"
 xvmc_deps="X11_extensions_XvMClib_h"
 
 h263_vaapi_hwaccel_deps="vaapi"
 h263_vaapi_hwaccel_select="h263_decoder"
 h263_vdpau_hwaccel_deps="vdpau"
 h263_vdpau_hwaccel_select="h263_decoder"
+h263_videotoolbox_hwaccel_deps="videotoolbox"
+h263_videotoolbox_hwaccel_select="h263_decoder"
 h264_crystalhd_decoder_select="crystalhd h264_mp4toannexb_bsf h264_parser"
 h264_d3d11va_hwaccel_deps="d3d11va"
 h264_d3d11va_hwaccel_select="h264_decoder"
@@ -2410,6 +2416,8 @@ h264_vdpau_decoder_deps="vdpau"
 h264_vdpau_decoder_select="h264_decoder"
 h264_vdpau_hwaccel_deps="vdpau"
 h264_vdpau_hwaccel_select="h264_decoder"
+h264_videotoolbox_hwaccel_deps="videotoolbox"
+h264_videotoolbox_hwaccel_select="h264_decoder"
 hevc_d3d11va_hwaccel_deps="d3d11va DXVA_PicParams_HEVC"
 hevc_d3d11va_hwaccel_select="hevc_decoder"
 hevc_dxva2_hwaccel_deps="dxva2 DXVA_PicParams_HEVC"
@@ -2425,6 +2433,8 @@ mpeg1_vdpau_decoder_deps="vdpau"
 mpeg1_vdpau_decoder_select="mpeg1video_decoder"
 mpeg1_vdpau_hwaccel_deps="vdpau"
 mpeg1_vdpau_hwaccel_select="mpeg1video_decoder"
+mpeg1_videotoolbox_hwaccel_deps="videotoolbox"
+mpeg1_videotoolbox_hwaccel_select="mpeg1video_decoder"
 mpeg1_xvmc_hwaccel_deps="xvmc"
 mpeg1_xvmc_hwaccel_select="mpeg1video_decoder"
 mpeg2_crystalhd_decoder_select="crystalhd"
@@ -2438,6 +2448,8 @@ mpeg2_vaapi_hwaccel_deps="vaapi"
 mpeg2_vaapi_hwaccel_select="mpeg2video_decoder"
 mpeg2_vdpau_hwaccel_deps="vdpau"
 mpeg2_vdpau_hwaccel_select="mpeg2video_decoder"
+mpeg2_videotoolbox_hwaccel_deps="videotoolbox"
+mpeg2_videotoolbox_hwaccel_select="mpeg2video_decoder"
 mpeg2_xvmc_hwaccel_deps="xvmc"
 mpeg2_xvmc_hwaccel_select="mpeg2video_decoder"
 mpeg4_crystalhd_decoder_select="crystalhd"
@@ -2447,6 +2459,8 @@ mpeg4_vdpau_decoder_deps="vdpau"
 mpeg4_vdpau_decoder_select="mpeg4_decoder"
 mpeg4_vdpau_hwaccel_deps="vdpau"
 mpeg4_vdpau_hwaccel_select="mpeg4_decoder"
+mpeg4_videotoolbox_hwaccel_deps="videotoolbox"
+mpeg4_videotoolbox_hwaccel_select="mpeg4_decoder"
 msmpeg4_crystalhd_decoder_select="crystalhd"
 vc1_crystalhd_decoder_select="crystalhd"
 vc1_d3d11va_hwaccel_deps="d3d11va"
@@ -2902,7 +2916,7 @@ sws_max_filter_size_default=256
 set_default sws_max_filter_size
 
 # Enable hwaccels by default.
-enable d3d11va dxva2 vaapi vda vdpau xvmc
+enable d3d11va dxva2 vaapi vda vdpau videotoolbox xvmc
 enable xlib
 
 # build settings
@@ -5108,6 +5122,7 @@ check_header valgrind/valgrind.h
 check_header vdpau/vdpau.h
 check_header vdpau/vdpau_x11.h
 check_header VideoDecodeAcceleration/VDADecoder.h
+check_header VideoToolbox/VideoToolbox.h
 check_header windows.h
 check_header X11/extensions/XvMClib.h
 check_header asm/types.h
diff --git a/ffmpeg.h b/ffmpeg.h
index 7467b165f8..03ba2d0d99 100644
--- a/ffmpeg.h
+++ b/ffmpeg.h
@@ -63,6 +63,7 @@ enum HWAccelID {
     HWACCEL_VDPAU,
     HWACCEL_DXVA2,
     HWACCEL_VDA,
+    HWACCEL_VIDEOTOOLBOX,
 };
 
 typedef struct HWAccel {
@@ -520,6 +521,7 @@ extern int frame_bits_per_raw_sample;
 extern AVIOContext *progress_avio;
 extern float max_error_rate;
 extern int vdpau_api_ver;
+extern char *videotoolbox_pixfmt;
 
 extern const AVIOInterruptCB int_cb;
 
@@ -554,5 +556,6 @@ int ffmpeg_parse_options(int argc, char **argv);
 int vdpau_init(AVCodecContext *s);
 int dxva2_init(AVCodecContext *s);
 int vda_init(AVCodecContext *s);
+int videotoolbox_init(AVCodecContext *s);
 
 #endif /* FFMPEG_H */
diff --git a/ffmpeg_opt.c b/ffmpeg_opt.c
index d912cb02b5..66a43b4093 100644
--- a/ffmpeg_opt.c
+++ b/ffmpeg_opt.c
@@ -74,7 +74,10 @@ const HWAccel hwaccels[] = {
     { "dxva2", dxva2_init, HWACCEL_DXVA2, AV_PIX_FMT_DXVA2_VLD },
 #endif
 #if CONFIG_VDA
-    { "vda",   vda_init,   HWACCEL_VDA,   AV_PIX_FMT_VDA },
+    { "vda",   videotoolbox_init,   HWACCEL_VDA,   AV_PIX_FMT_VDA },
+#endif
+#if CONFIG_VIDEOTOOLBOX
+    { "videotoolbox",   videotoolbox_init,   HWACCEL_VIDEOTOOLBOX,   AV_PIX_FMT_VIDEOTOOLBOX },
 #endif
     { 0 },
 };
@@ -3233,6 +3236,9 @@ const OptionDef options[] = {
 #if HAVE_VDPAU_X11
     { "vdpau_api_ver", HAS_ARG | OPT_INT | OPT_EXPERT, { &vdpau_api_ver }, "" },
 #endif
+#if CONFIG_VDA || CONFIG_VIDEOTOOLBOX
+    { "videotoolbox_pixfmt", HAS_ARG | OPT_STRING | OPT_EXPERT, { &videotoolbox_pixfmt}, "" },
+#endif
     { "autorotate",       HAS_ARG | OPT_BOOL | OPT_SPEC |
                           OPT_EXPERT | OPT_INPUT,                                { .off = OFFSET(autorotate) },
         "automatically insert correct rotate filters" },
diff --git a/ffmpeg_vda.c b/ffmpeg_vda.c
deleted file mode 100644
index 79a8503e7d..0000000000
--- a/ffmpeg_vda.c
+++ /dev/null
@@ -1,137 +0,0 @@
-/*
- * 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 "libavcodec/avcodec.h"
-#include "libavcodec/vda.h"
-#include "libavutil/imgutils.h"
-
-#include "ffmpeg.h"
-
-typedef struct VDAContext {
-    AVFrame *tmp_frame;
-} VDAContext;
-
-static int vda_retrieve_data(AVCodecContext *s, AVFrame *frame)
-{
-    InputStream *ist = s->opaque;
-    VDAContext  *vda = ist->hwaccel_ctx;
-    CVPixelBufferRef pixbuf = (CVPixelBufferRef)frame->data[3];
-    OSType pixel_format = CVPixelBufferGetPixelFormatType(pixbuf);
-    CVReturn err;
-    uint8_t *data[4] = { 0 };
-    int linesize[4] = { 0 };
-    int planes, ret, i;
-
-    av_frame_unref(vda->tmp_frame);
-
-    switch (pixel_format) {
-    case kCVPixelFormatType_420YpCbCr8Planar: vda->tmp_frame->format = AV_PIX_FMT_YUV420P; break;
-    case kCVPixelFormatType_422YpCbCr8:       vda->tmp_frame->format = AV_PIX_FMT_UYVY422; break;
-    default:
-        av_log(NULL, AV_LOG_ERROR,
-               "Unsupported pixel format: %u\n", pixel_format);
-        return AVERROR(ENOSYS);
-    }
-
-    vda->tmp_frame->width  = frame->width;
-    vda->tmp_frame->height = frame->height;
-    ret = av_frame_get_buffer(vda->tmp_frame, 32);
-    if (ret < 0)
-        return ret;
-
-    err = CVPixelBufferLockBaseAddress(pixbuf, kCVPixelBufferLock_ReadOnly);
-    if (err != kCVReturnSuccess) {
-        av_log(NULL, AV_LOG_ERROR, "Error locking the pixel buffer.\n");
-        return AVERROR_UNKNOWN;
-    }
-
-    if (CVPixelBufferIsPlanar(pixbuf)) {
-
-        planes = CVPixelBufferGetPlaneCount(pixbuf);
-        for (i = 0; i < planes; i++) {
-            data[i]     = CVPixelBufferGetBaseAddressOfPlane(pixbuf, i);
-            linesize[i] = CVPixelBufferGetBytesPerRowOfPlane(pixbuf, i);
-        }
-    } else {
-        data[0] = CVPixelBufferGetBaseAddress(pixbuf);
-        linesize[0] = CVPixelBufferGetBytesPerRow(pixbuf);
-    }
-
-    av_image_copy(vda->tmp_frame->data, vda->tmp_frame->linesize,
-                  (const uint8_t **)data, linesize, vda->tmp_frame->format,
-                  frame->width, frame->height);
-
-    CVPixelBufferUnlockBaseAddress(pixbuf, kCVPixelBufferLock_ReadOnly);
-
-    ret = av_frame_copy_props(vda->tmp_frame, frame);
-
-    if (ret < 0)
-        return ret;
-
-    av_frame_unref(frame);
-    av_frame_move_ref(frame, vda->tmp_frame);
-
-    return 0;
-}
-
-static void vda_uninit(AVCodecContext *s)
-{
-    InputStream *ist = s->opaque;
-    VDAContext  *vda = ist->hwaccel_ctx;
-
-    ist->hwaccel_uninit        = NULL;
-    ist->hwaccel_retrieve_data = NULL;
-
-    av_frame_free(&vda->tmp_frame);
-
-    av_vda_default_free(s);
-    av_freep(&ist->hwaccel_ctx);
-}
-
-int vda_init(AVCodecContext *s)
-{
-    InputStream *ist = s->opaque;
-    int loglevel = (ist->hwaccel_id == HWACCEL_AUTO) ? AV_LOG_VERBOSE : AV_LOG_ERROR;
-    VDAContext *vda;
-    int ret;
-
-    vda = av_mallocz(sizeof(*vda));
-    if (!vda)
-        return AVERROR(ENOMEM);
-
-    ist->hwaccel_ctx           = vda;
-    ist->hwaccel_uninit        = vda_uninit;
-    ist->hwaccel_retrieve_data = vda_retrieve_data;
-
-    vda->tmp_frame = av_frame_alloc();
-    if (!vda->tmp_frame) {
-        ret = AVERROR(ENOMEM);
-        goto fail;
-    }
-
-    ret = av_vda_default_init(s);
-    if (ret < 0) {
-        av_log(NULL, loglevel, "Error creating VDA decoder.\n");
-        goto fail;
-    }
-
-    return 0;
-fail:
-    vda_uninit(s);
-    return ret;
-}
diff --git a/ffmpeg_videotoolbox.c b/ffmpeg_videotoolbox.c
new file mode 100644
index 0000000000..6688452c04
--- /dev/null
+++ b/ffmpeg_videotoolbox.c
@@ -0,0 +1,187 @@
+/*
+ * 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 <CoreServices/CoreServices.h>
+
+#include "config.h"
+#include "libavcodec/avcodec.h"
+#if CONFIG_VDA
+#  include "libavcodec/vda.h"
+#endif
+#if CONFIG_VIDEOTOOLBOX
+#  include "libavcodec/videotoolbox.h"
+#endif
+#include "libavutil/imgutils.h"
+#include "ffmpeg.h"
+
+typedef struct VTContext {
+    AVFrame *tmp_frame;
+} VTContext;
+
+char *videotoolbox_pixfmt;
+
+static int videotoolbox_retrieve_data(AVCodecContext *s, AVFrame *frame)
+{
+    InputStream *ist = s->opaque;
+    VTContext  *vt = ist->hwaccel_ctx;
+    CVPixelBufferRef pixbuf = (CVPixelBufferRef)frame->data[3];
+    OSType pixel_format = CVPixelBufferGetPixelFormatType(pixbuf);
+    CVReturn err;
+    uint8_t *data[4] = { 0 };
+    int linesize[4] = { 0 };
+    int planes, ret, i;
+    char codec_str[32];
+
+    av_frame_unref(vt->tmp_frame);
+
+    switch (pixel_format) {
+    case kCVPixelFormatType_420YpCbCr8Planar: vt->tmp_frame->format = AV_PIX_FMT_YUV420P; break;
+    case kCVPixelFormatType_422YpCbCr8:       vt->tmp_frame->format = AV_PIX_FMT_UYVY422; break;
+    case kCVPixelFormatType_32BGRA:           vt->tmp_frame->format = AV_PIX_FMT_BGRA; break;
+#ifdef kCFCoreFoundationVersionNumber10_7
+    case kCVPixelFormatType_420YpCbCr8BiPlanarVideoRange: vt->tmp_frame->format = AV_PIX_FMT_NV12; break;
+#endif
+    default:
+        av_get_codec_tag_string(codec_str, sizeof(codec_str), s->codec_tag);
+        av_log(NULL, AV_LOG_ERROR,
+               "%s: Unsupported pixel format: %s\n", codec_str, videotoolbox_pixfmt);
+        return AVERROR(ENOSYS);
+    }
+
+    vt->tmp_frame->width  = frame->width;
+    vt->tmp_frame->height = frame->height;
+    ret = av_frame_get_buffer(vt->tmp_frame, 32);
+    if (ret < 0)
+        return ret;
+
+    err = CVPixelBufferLockBaseAddress(pixbuf, kCVPixelBufferLock_ReadOnly);
+    if (err != kCVReturnSuccess) {
+        av_log(NULL, AV_LOG_ERROR, "Error locking the pixel buffer.\n");
+        return AVERROR_UNKNOWN;
+    }
+
+    if (CVPixelBufferIsPlanar(pixbuf)) {
+
+        planes = CVPixelBufferGetPlaneCount(pixbuf);
+        for (i = 0; i < planes; i++) {
+            data[i]     = CVPixelBufferGetBaseAddressOfPlane(pixbuf, i);
+            linesize[i] = CVPixelBufferGetBytesPerRowOfPlane(pixbuf, i);
+        }
+    } else {
+        data[0] = CVPixelBufferGetBaseAddress(pixbuf);
+        linesize[0] = CVPixelBufferGetBytesPerRow(pixbuf);
+    }
+
+    av_image_copy(vt->tmp_frame->data, vt->tmp_frame->linesize,
+                  (const uint8_t **)data, linesize, vt->tmp_frame->format,
+                  frame->width, frame->height);
+
+    ret = av_frame_copy_props(vt->tmp_frame, frame);
+    CVPixelBufferUnlockBaseAddress(pixbuf, kCVPixelBufferLock_ReadOnly);
+    if (ret < 0)
+        return ret;
+
+    av_frame_unref(frame);
+    av_frame_move_ref(frame, vt->tmp_frame);
+
+    return 0;
+}
+
+static void videotoolbox_uninit(AVCodecContext *s)
+{
+    InputStream *ist = s->opaque;
+    VTContext  *vt = ist->hwaccel_ctx;
+
+    ist->hwaccel_uninit        = NULL;
+    ist->hwaccel_retrieve_data = NULL;
+
+    av_frame_free(&vt->tmp_frame);
+
+    if (ist->hwaccel_id == HWACCEL_VIDEOTOOLBOX) {
+#if CONFIG_VIDEOTOOLBOX
+        av_videotoolbox_default_free(s);
+#endif
+    } else {
+#if CONFIG_VDA
+        av_vda_default_free(s);
+#endif
+    }
+    av_freep(&ist->hwaccel_ctx);
+}
+
+int videotoolbox_init(AVCodecContext *s)
+{
+    InputStream *ist = s->opaque;
+    int loglevel = (ist->hwaccel_id == HWACCEL_AUTO) ? AV_LOG_VERBOSE : AV_LOG_ERROR;
+    int ret = 0;
+    VTContext *vt;
+
+    vt = av_mallocz(sizeof(*vt));
+    if (!vt)
+        return AVERROR(ENOMEM);
+
+    ist->hwaccel_ctx           = vt;
+    ist->hwaccel_uninit        = videotoolbox_uninit;
+    ist->hwaccel_retrieve_data = videotoolbox_retrieve_data;
+
+    vt->tmp_frame = av_frame_alloc();
+    if (!vt->tmp_frame) {
+        ret = AVERROR(ENOMEM);
+        goto fail;
+    }
+
+    if (ist->hwaccel_id == HWACCEL_VIDEOTOOLBOX) {
+#if CONFIG_VIDEOTOOLBOX
+        if (!videotoolbox_pixfmt) {
+            ret = av_videotoolbox_default_init(s);
+        } else {
+            AVVideotoolboxContext *vtctx = av_videotoolbox_alloc_context();
+            CFStringRef pixfmt_str = CFStringCreateWithCString(kCFAllocatorDefault,
+                                                               videotoolbox_pixfmt,
+                                                               kCFStringEncodingUTF8);
+            vtctx->cv_pix_fmt_type = UTGetOSTypeFromString(pixfmt_str);
+            ret = av_videotoolbox_default_init2(s, vtctx);
+            CFRelease(pixfmt_str);
+        }
+#endif
+    } else {
+#if CONFIG_VDA
+        if (!videotoolbox_pixfmt) {
+            ret = av_vda_default_init(s);
+        } else {
+            AVVDAContext *vdactx = av_vda_alloc_context();
+            CFStringRef pixfmt_str = CFStringCreateWithCString(kCFAllocatorDefault,
+                                                               videotoolbox_pixfmt,
+                                                               kCFStringEncodingUTF8);
+            vdactx->cv_pix_fmt_type = UTGetOSTypeFromString(pixfmt_str);
+            ret = av_vda_default_init2(s, vdactx);
+            CFRelease(pixfmt_str);
+        }
+#endif
+    }
+    if (ret < 0) {
+        av_log(NULL, loglevel,
+               "Error creating %s decoder.\n", ist->hwaccel_id == HWACCEL_VIDEOTOOLBOX ? "Videotoolbox" : "VDA");
+        goto fail;
+    }
+
+    return 0;
+fail:
+    videotoolbox_uninit(s);
+    return ret;
+}
diff --git a/libavcodec/Makefile b/libavcodec/Makefile
index 7cf24031d3..e61b9cd4eb 100644
--- a/libavcodec/Makefile
+++ b/libavcodec/Makefile
@@ -13,6 +13,7 @@ HEADERS = avcodec.h                                                     \
           vda.h                                                         \
           vdpau.h                                                       \
           version.h                                                     \
+          videotoolbox.h                                                \
           vorbis_parser.h                                               \
           xvmc.h                                                        \
 
@@ -696,28 +697,34 @@ OBJS-$(CONFIG_VIMA_DECODER)               += vima.o adpcm_data.o
 OBJS-$(CONFIG_D3D11VA)                    += dxva2.o
 OBJS-$(CONFIG_DXVA2)                      += dxva2.o
 OBJS-$(CONFIG_VAAPI)                      += vaapi.o
-OBJS-$(CONFIG_VDA)                        += vda.o
+OBJS-$(CONFIG_VDA)                        += vda.o videotoolbox.o
+OBJS-$(CONFIG_VIDEOTOOLBOX)               += videotoolbox.o
 OBJS-$(CONFIG_VDPAU)                      += vdpau.o
 
 OBJS-$(CONFIG_H263_VAAPI_HWACCEL)         += vaapi_mpeg4.o
 OBJS-$(CONFIG_H263_VDPAU_HWACCEL)         += vdpau_mpeg4.o
+OBJS-$(CONFIG_H263_VIDEOTOOLBOX_HWACCEL)  += videotoolbox.o
 OBJS-$(CONFIG_H264_D3D11VA_HWACCEL)       += dxva2_h264.o
 OBJS-$(CONFIG_H264_DXVA2_HWACCEL)         += dxva2_h264.o
 OBJS-$(CONFIG_H264_VAAPI_HWACCEL)         += vaapi_h264.o
 OBJS-$(CONFIG_H264_VDA_HWACCEL)           += vda_h264.o
 OBJS-$(CONFIG_H264_VDPAU_HWACCEL)         += vdpau_h264.o
+OBJS-$(CONFIG_H264_VIDEOTOOLBOX_HWACCEL)  += videotoolbox.o
 OBJS-$(CONFIG_HEVC_D3D11VA_HWACCEL)       += dxva2_hevc.o
 OBJS-$(CONFIG_HEVC_DXVA2_HWACCEL)         += dxva2_hevc.o
 OBJS-$(CONFIG_HEVC_VDPAU_HWACCEL)         += vdpau_hevc.o
 OBJS-$(CONFIG_MPEG1_VDPAU_HWACCEL)        += vdpau_mpeg12.o
+OBJS-$(CONFIG_MPEG1_VIDEOTOOLBOX_HWACCEL) += videotoolbox.o
 OBJS-$(CONFIG_MPEG1_XVMC_HWACCEL)         += mpegvideo_xvmc.o
 OBJS-$(CONFIG_MPEG2_D3D11VA_HWACCEL)      += dxva2_mpeg2.o
 OBJS-$(CONFIG_MPEG2_DXVA2_HWACCEL)        += dxva2_mpeg2.o
 OBJS-$(CONFIG_MPEG2_VAAPI_HWACCEL)        += vaapi_mpeg2.o
 OBJS-$(CONFIG_MPEG2_VDPAU_HWACCEL)        += vdpau_mpeg12.o
+OBJS-$(CONFIG_MPEG2_VIDEOTOOLBOX_HWACCEL) += videotoolbox.o
 OBJS-$(CONFIG_MPEG2_XVMC_HWACCEL)         += mpegvideo_xvmc.o
 OBJS-$(CONFIG_MPEG4_VAAPI_HWACCEL)        += vaapi_mpeg4.o
 OBJS-$(CONFIG_MPEG4_VDPAU_HWACCEL)        += vdpau_mpeg4.o
+OBJS-$(CONFIG_MPEG4_VIDEOTOOLBOX_HWACCEL) += videotoolbox.o
 OBJS-$(CONFIG_VC1_D3D11VA_HWACCEL)        += dxva2_vc1.o
 OBJS-$(CONFIG_VC1_DXVA2_HWACCEL)          += dxva2_vc1.o
 OBJS-$(CONFIG_VC1_VAAPI_HWACCEL)          += vaapi_vc1.o
@@ -912,8 +919,9 @@ SKIPHEADERS-$(CONFIG_QSVDEC)           += qsvdec.h
 SKIPHEADERS-$(CONFIG_QSVENC)           += qsvenc.h
 SKIPHEADERS-$(CONFIG_XVMC)             += xvmc.h
 SKIPHEADERS-$(CONFIG_VAAPI)            += vaapi_internal.h
-SKIPHEADERS-$(CONFIG_VDA)              += vda.h vda_internal.h
+SKIPHEADERS-$(CONFIG_VDA)              += vda.h vda_vt_internal.h
 SKIPHEADERS-$(CONFIG_VDPAU)            += vdpau.h vdpau_internal.h
+SKIPHEADERS-$(CONFIG_VIDEOTOOLBOX)     += videotoolbox.h vda_vt_internal.h
 
 TESTPROGS = imgconvert                                                  \
             jpeg2000dwt                                                 \
diff --git a/libavcodec/allcodecs.c b/libavcodec/allcodecs.c
index 417f824041..ed0975ef7c 100644
--- a/libavcodec/allcodecs.c
+++ b/libavcodec/allcodecs.c
@@ -76,6 +76,7 @@ void avcodec_register_all(void)
     /* hardware accelerators */
     REGISTER_HWACCEL(H263_VAAPI,        h263_vaapi);
     REGISTER_HWACCEL(H263_VDPAU,        h263_vdpau);
+    REGISTER_HWACCEL(H263_VIDEOTOOLBOX, h263_videotoolbox);
     REGISTER_HWACCEL(H264_D3D11VA,      h264_d3d11va);
     REGISTER_HWACCEL(H264_DXVA2,        h264_dxva2);
     REGISTER_HWACCEL(H264_MMAL,         h264_mmal);
@@ -84,20 +85,24 @@ void avcodec_register_all(void)
     REGISTER_HWACCEL(H264_VDA,          h264_vda);
     REGISTER_HWACCEL(H264_VDA_OLD,      h264_vda_old);
     REGISTER_HWACCEL(H264_VDPAU,        h264_vdpau);
+    REGISTER_HWACCEL(H264_VIDEOTOOLBOX, h264_videotoolbox);
     REGISTER_HWACCEL(HEVC_D3D11VA,      hevc_d3d11va);
     REGISTER_HWACCEL(HEVC_DXVA2,        hevc_dxva2);
     REGISTER_HWACCEL(HEVC_QSV,          hevc_qsv);
     REGISTER_HWACCEL(HEVC_VDPAU,        hevc_vdpau);
     REGISTER_HWACCEL(MPEG1_XVMC,        mpeg1_xvmc);
     REGISTER_HWACCEL(MPEG1_VDPAU,       mpeg1_vdpau);
+    REGISTER_HWACCEL(MPEG1_VIDEOTOOLBOX, mpeg1_videotoolbox);
     REGISTER_HWACCEL(MPEG2_XVMC,        mpeg2_xvmc);
     REGISTER_HWACCEL(MPEG2_D3D11VA,     mpeg2_d3d11va);
     REGISTER_HWACCEL(MPEG2_DXVA2,       mpeg2_dxva2);
     REGISTER_HWACCEL(MPEG2_QSV,         mpeg2_qsv);
     REGISTER_HWACCEL(MPEG2_VAAPI,       mpeg2_vaapi);
     REGISTER_HWACCEL(MPEG2_VDPAU,       mpeg2_vdpau);
+    REGISTER_HWACCEL(MPEG2_VIDEOTOOLBOX, mpeg2_videotoolbox);
     REGISTER_HWACCEL(MPEG4_VAAPI,       mpeg4_vaapi);
     REGISTER_HWACCEL(MPEG4_VDPAU,       mpeg4_vdpau);
+    REGISTER_HWACCEL(MPEG4_VIDEOTOOLBOX, mpeg4_videotoolbox);
     REGISTER_HWACCEL(VC1_D3D11VA,       vc1_d3d11va);
     REGISTER_HWACCEL(VC1_DXVA2,         vc1_dxva2);
     REGISTER_HWACCEL(VC1_VAAPI,         vc1_vaapi);
diff --git a/libavcodec/h263dec.c b/libavcodec/h263dec.c
index c173152f9a..7fa7090623 100644
--- a/libavcodec/h263dec.c
+++ b/libavcodec/h263dec.c
@@ -723,6 +723,9 @@ const enum AVPixelFormat ff_h263_hwaccel_pixfmt_list_420[] = {
 #if CONFIG_H263_VDPAU_HWACCEL || CONFIG_MPEG4_VDPAU_HWACCEL
     AV_PIX_FMT_VDPAU,
 #endif
+#if CONFIG_H263_VIDEOTOOLBOX_HWACCEL || CONFIG_MPEG4_VIDEOTOOLBOX_HWACCEL
+    AV_PIX_FMT_VIDEOTOOLBOX,
+#endif
     AV_PIX_FMT_YUV420P,
     AV_PIX_FMT_NONE
 };
diff --git a/libavcodec/h264_slice.c b/libavcodec/h264_slice.c
index 9a8319bce0..b088392576 100644
--- a/libavcodec/h264_slice.c
+++ b/libavcodec/h264_slice.c
@@ -868,6 +868,7 @@ static enum AVPixelFormat get_pixel_format(H264Context *h, int force_callback)
                      CONFIG_H264_D3D11VA_HWACCEL + \
                      CONFIG_H264_VAAPI_HWACCEL + \
                      (CONFIG_H264_VDA_HWACCEL * 2) + \
+                     CONFIG_H264_VIDEOTOOLBOX_HWACCEL + \
                      CONFIG_H264_VDPAU_HWACCEL)
     enum AVPixelFormat pix_fmts[HWACCEL_MAX + 2], *fmt = pix_fmts;
     const enum AVPixelFormat *choices = pix_fmts;
@@ -948,6 +949,9 @@ static enum AVPixelFormat get_pixel_format(H264Context *h, int force_callback)
             *fmt++ = AV_PIX_FMT_VDA_VLD;
             *fmt++ = AV_PIX_FMT_VDA;
 #endif
+#if CONFIG_H264_VIDEOTOOLBOX_HWACCEL
+            *fmt++ = AV_PIX_FMT_VIDEOTOOLBOX;
+#endif
             if (h->avctx->codec->pix_fmts)
                 choices = h->avctx->codec->pix_fmts;
             else if (h->avctx->color_range == AVCOL_RANGE_JPEG)
diff --git a/libavcodec/mpeg12dec.c b/libavcodec/mpeg12dec.c
index 73ebdfa88b..1250d5343a 100644
--- a/libavcodec/mpeg12dec.c
+++ b/libavcodec/mpeg12dec.c
@@ -1211,6 +1211,9 @@ static const enum AVPixelFormat mpeg2_hwaccel_pixfmt_list_420[] = {
 #if CONFIG_MPEG2_VAAPI_HWACCEL
     AV_PIX_FMT_VAAPI_VLD,
 #endif
+#if CONFIG_MPEG2_VIDEOTOOLBOX_HWACCEL
+    AV_PIX_FMT_VIDEOTOOLBOX,
+#endif
     AV_PIX_FMT_YUV420P,
     AV_PIX_FMT_NONE
 };
diff --git a/libavcodec/vda.c b/libavcodec/vda.c
index 5867cae120..4670140c76 100644
--- a/libavcodec/vda.c
+++ b/libavcodec/vda.c
@@ -21,7 +21,7 @@
 #include "libavutil/mem.h"
 
 #include "vda.h"
-#include "vda_internal.h"
+#include "vda_vt_internal.h"
 
 #if CONFIG_H264_VDA_HWACCEL
 AVVDAContext *av_vda_alloc_context(void)
diff --git a/libavcodec/vda_h264.c b/libavcodec/vda_h264.c
index a1179f16e2..8c526c0754 100644
--- a/libavcodec/vda_h264.c
+++ b/libavcodec/vda_h264.c
@@ -32,20 +32,7 @@ struct vda_buffer {
     CVPixelBufferRef cv_buffer;
 };
 #include "internal.h"
-#include "vda_internal.h"
-
-typedef struct VDAContext {
-    // The current bitstream buffer.
-    uint8_t             *bitstream;
-
-    // The current size of the bitstream.
-    int                  bitstream_size;
-
-    // The reference size used for fast reallocation.
-    int                  allocated_size;
-
-    CVImageBufferRef frame;
-} VDAContext;
+#include "vda_vt_internal.h"
 
 /* Decoder callback that adds the vda frame to the queue in display order. */
 static void vda_decoder_callback(void *vda_hw_ctx,
@@ -68,7 +55,7 @@ static void vda_decoder_callback(void *vda_hw_ctx,
     vda_ctx->cv_buffer = CVPixelBufferRetain(image_buffer);
 }
 
-static int vda_sync_decode(VDAContext *ctx, struct vda_context *vda_ctx)
+static int vda_sync_decode(VTContext *ctx, struct vda_context *vda_ctx)
 {
     OSStatus status;
     CFDataRef coded_frame;
@@ -93,7 +80,7 @@ static int vda_old_h264_start_frame(AVCodecContext *avctx,
                                 av_unused const uint8_t *buffer,
                                 av_unused uint32_t size)
 {
-    VDAContext *vda = avctx->internal->hwaccel_priv_data;
+    VTContext *vda = avctx->internal->hwaccel_priv_data;
     struct vda_context *vda_ctx = avctx->hwaccel_context;
 
     if (!vda_ctx->decoder)
@@ -108,7 +95,7 @@ static int vda_old_h264_decode_slice(AVCodecContext *avctx,
                                  const uint8_t *buffer,
                                  uint32_t size)
 {
-    VDAContext *vda             = avctx->internal->hwaccel_priv_data;
+    VTContext *vda              = avctx->internal->hwaccel_priv_data;
     struct vda_context *vda_ctx = avctx->hwaccel_context;
     void *tmp;
 
@@ -141,7 +128,7 @@ static void vda_h264_release_buffer(void *opaque, uint8_t *data)
 static int vda_old_h264_end_frame(AVCodecContext *avctx)
 {
     H264Context *h                      = avctx->priv_data;
-    VDAContext *vda                     = avctx->internal->hwaccel_priv_data;
+    VTContext *vda                      = avctx->internal->hwaccel_priv_data;
     struct vda_context *vda_ctx         = avctx->hwaccel_context;
     AVFrame *frame                      = h->cur_pic_ptr->f;
     struct vda_buffer *context;
@@ -271,17 +258,6 @@ int ff_vda_destroy_decoder(struct vda_context *vda_ctx)
     return status;
 }
 
-static int vda_h264_uninit(AVCodecContext *avctx)
-{
-    VDAContext *vda = avctx->internal->hwaccel_priv_data;
-    if (vda) {
-        av_freep(&vda->bitstream);
-        if (vda->frame)
-            CVPixelBufferRelease(vda->frame);
-    }
-    return 0;
-}
-
 AVHWAccel ff_h264_vda_old_hwaccel = {
     .name           = "h264_vda",
     .type           = AVMEDIA_TYPE_VIDEO,
@@ -290,8 +266,8 @@ AVHWAccel ff_h264_vda_old_hwaccel = {
     .start_frame    = vda_old_h264_start_frame,
     .decode_slice   = vda_old_h264_decode_slice,
     .end_frame      = vda_old_h264_end_frame,
-    .uninit         = vda_h264_uninit,
-    .priv_data_size = sizeof(VDAContext),
+    .uninit         = ff_videotoolbox_uninit,
+    .priv_data_size = sizeof(VTContext),
 };
 
 void ff_vda_output_callback(void *opaque,
@@ -301,7 +277,7 @@ void ff_vda_output_callback(void *opaque,
                             CVImageBufferRef image_buffer)
 {
     AVCodecContext *ctx = opaque;
-    VDAContext *vda = ctx->internal->hwaccel_priv_data;
+    VTContext *vda = ctx->internal->hwaccel_priv_data;
 
 
     if (vda->frame) {
@@ -315,65 +291,10 @@ void ff_vda_output_callback(void *opaque,
     vda->frame = CVPixelBufferRetain(image_buffer);
 }
 
-static int vda_h264_start_frame(AVCodecContext *avctx,
-                                const uint8_t *buffer,
-                                uint32_t size)
-{
-    VDAContext *vda = avctx->internal->hwaccel_priv_data;
-    H264Context *h  = avctx->priv_data;
-
-    if (h->is_avc == 1) {
-        void *tmp;
-        vda->bitstream_size = 0;
-        tmp = av_fast_realloc(vda->bitstream,
-                              &vda->allocated_size,
-                              size);
-        vda->bitstream = tmp;
-        memcpy(vda->bitstream, buffer, size);
-        vda->bitstream_size = size;
-    } else {
-        vda->bitstream_size = 0;
-    }
-    return 0;
-}
-
-static int vda_h264_decode_slice(AVCodecContext *avctx,
-                                 const uint8_t *buffer,
-                                 uint32_t size)
-{
-    VDAContext *vda       = avctx->internal->hwaccel_priv_data;
-    H264Context *h  = avctx->priv_data;
-    void *tmp;
-
-    if (h->is_avc == 1)
-        return 0;
-
-    tmp = av_fast_realloc(vda->bitstream,
-                          &vda->allocated_size,
-                          vda->bitstream_size + size + 4);
-    if (!tmp)
-        return AVERROR(ENOMEM);
-
-    vda->bitstream = tmp;
-
-    AV_WB32(vda->bitstream + vda->bitstream_size, size);
-    memcpy(vda->bitstream + vda->bitstream_size + 4, buffer, size);
-
-    vda->bitstream_size += size + 4;
-
-    return 0;
-}
-
-static void release_buffer(void *opaque, uint8_t *data)
-{
-    CVImageBufferRef frame = (CVImageBufferRef)data;
-    CVPixelBufferRelease(frame);
-}
-
 static int vda_h264_end_frame(AVCodecContext *avctx)
 {
     H264Context *h        = avctx->priv_data;
-    VDAContext *vda       = avctx->internal->hwaccel_priv_data;
+    VTContext *vda        = avctx->internal->hwaccel_priv_data;
     AVVDAContext *vda_ctx = avctx->hwaccel_context;
     AVFrame *frame        = h->cur_pic_ptr->f;
     uint32_t flush_flags  = 1 << 0; ///< kVDADecoderFlush_emitFrames
@@ -403,19 +324,7 @@ static int vda_h264_end_frame(AVCodecContext *avctx)
         return AVERROR_UNKNOWN;
     }
 
-    av_buffer_unref(&frame->buf[0]);
-
-    frame->buf[0] = av_buffer_create((uint8_t*)vda->frame,
-                                     sizeof(vda->frame),
-                                     release_buffer, NULL,
-                                     AV_BUFFER_FLAG_READONLY);
-    if (!frame->buf[0])
-        return AVERROR(ENOMEM);
-
-    frame->data[3] = (uint8_t*)vda->frame;
-    vda->frame = NULL;
-
-    return 0;
+    return ff_videotoolbox_buffer_create(vda, frame);
 }
 
 int ff_vda_default_init(AVCodecContext *avctx)
@@ -434,26 +343,7 @@ int ff_vda_default_init(AVCodecContext *avctx)
 
     // kCVPixelFormatType_420YpCbCr8Planar;
 
-    /* Each VCL NAL in the bitstream sent to the decoder
-     * is preceded by a 4 bytes length header.
-     * Change the avcC atom header if needed, to signal headers of 4 bytes. */
-    if (avctx->extradata_size >= 4 && (avctx->extradata[4] & 0x03) != 0x03) {
-        uint8_t *rw_extradata;
-
-        if (!(rw_extradata = av_malloc(avctx->extradata_size)))
-            return AVERROR(ENOMEM);
-
-        memcpy(rw_extradata, avctx->extradata, avctx->extradata_size);
-
-        rw_extradata[4] |= 0x03;
-
-        avc_data = CFDataCreate(kCFAllocatorDefault, rw_extradata, avctx->extradata_size);
-
-        av_freep(&rw_extradata);
-    } else {
-        avc_data = CFDataCreate(kCFAllocatorDefault,
-                                avctx->extradata, avctx->extradata_size);
-    }
+    avc_data = ff_videotoolbox_avcc_extradata_create(avctx);
 
     config_info = CFDictionaryCreateMutable(kCFAllocatorDefault,
                                             4,
@@ -521,27 +411,15 @@ int ff_vda_default_init(AVCodecContext *avctx)
     }
 }
 
-static int vda_h264_alloc_frame(AVCodecContext *avctx, AVFrame *frame)
-{
-    frame->width  = avctx->width;
-    frame->height = avctx->height;
-    frame->format = avctx->pix_fmt;
-    frame->buf[0] = av_buffer_alloc(1);
-
-    if (!frame->buf[0])
-        return AVERROR(ENOMEM);
-    return 0;
-}
-
 AVHWAccel ff_h264_vda_hwaccel = {
     .name           = "h264_vda",
     .type           = AVMEDIA_TYPE_VIDEO,
     .id             = AV_CODEC_ID_H264,
     .pix_fmt        = AV_PIX_FMT_VDA,
-    .alloc_frame    = vda_h264_alloc_frame,
-    .start_frame    = vda_h264_start_frame,
-    .decode_slice   = vda_h264_decode_slice,
+    .alloc_frame    = ff_videotoolbox_alloc_frame,
+    .start_frame    = ff_videotoolbox_h264_start_frame,
+    .decode_slice   = ff_videotoolbox_h264_decode_slice,
     .end_frame      = vda_h264_end_frame,
-    .uninit         = vda_h264_uninit,
-    .priv_data_size = sizeof(VDAContext),
+    .uninit         = ff_videotoolbox_uninit,
+    .priv_data_size = sizeof(VTContext),
 };
diff --git a/libavcodec/vda_internal.h b/libavcodec/vda_vt_internal.h
index 457916b058..9ff63ccc52 100644
--- a/libavcodec/vda_internal.h
+++ b/libavcodec/vda_vt_internal.h
@@ -16,10 +16,8 @@
  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  */
 
-#ifndef AVCODEC_VDA_INTERNAL_H
-#define AVCODEC_VDA_INTERNAL_H
-
-#include "vda.h"
+#ifndef AVCODEC_VDA_VT_INTERNAL_H
+#define AVCODEC_VDA_VT_INTERNAL_H
 
 void ff_vda_output_callback(void *vda_hw_ctx,
                             CFDictionaryRef user_info,
@@ -30,4 +28,28 @@ void ff_vda_output_callback(void *vda_hw_ctx,
 int ff_vda_default_init(AVCodecContext *avctx);
 void ff_vda_default_free(AVCodecContext *avctx);
 
-#endif /* AVCODEC_VDA_INTERNAL_H */
+typedef struct VTContext {
+    // The current bitstream buffer.
+    uint8_t                     *bitstream;
+
+    // The current size of the bitstream.
+    int                         bitstream_size;
+
+    // The reference size used for fast reallocation.
+    int                         allocated_size;
+
+    // The core video buffer
+    CVImageBufferRef            frame;
+} VTContext;
+
+int ff_videotoolbox_alloc_frame(AVCodecContext *avctx, AVFrame *frame);
+int ff_videotoolbox_uninit(AVCodecContext *avctx);
+int ff_videotoolbox_buffer_create(VTContext *vtctx, AVFrame *frame);
+int ff_videotoolbox_h264_start_frame(AVCodecContext *avctx,
+                                     const uint8_t *buffer,
+                                     uint32_t size);
+int ff_videotoolbox_h264_decode_slice(AVCodecContext *avctx,
+                                      const uint8_t *buffer,
+                                      uint32_t size);
+CFDataRef ff_videotoolbox_avcc_extradata_create(AVCodecContext *avctx);
+#endif /* AVCODEC_VDA_VT_INTERNAL_H */
diff --git a/libavcodec/version.h b/libavcodec/version.h
index c6efd03efe..1b37a9e68a 100644
--- a/libavcodec/version.h
+++ b/libavcodec/version.h
@@ -29,8 +29,8 @@
 #include "libavutil/version.h"
 
 #define LIBAVCODEC_VERSION_MAJOR 56
-#define LIBAVCODEC_VERSION_MINOR  56
-#define LIBAVCODEC_VERSION_MICRO 101
+#define LIBAVCODEC_VERSION_MINOR  57
+#define LIBAVCODEC_VERSION_MICRO 100
 
 #define LIBAVCODEC_VERSION_INT  AV_VERSION_INT(LIBAVCODEC_VERSION_MAJOR, \
                                                LIBAVCODEC_VERSION_MINOR, \
diff --git a/libavcodec/videotoolbox.c b/libavcodec/videotoolbox.c
new file mode 100644
index 0000000000..cdb314b29b
--- /dev/null
+++ b/libavcodec/videotoolbox.c
@@ -0,0 +1,690 @@
+/*
+ * Videotoolbox hardware acceleration
+ *
+ * copyright (c) 2012 Sebastien Zwickert
+ *
+ * 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 "config.h"
+#if CONFIG_VIDEOTOOLBOX
+#  include "videotoolbox.h"
+#else
+#  include "vda.h"
+#endif
+#include "vda_vt_internal.h"
+#include "libavutil/avutil.h"
+#include "bytestream.h"
+#include "h264.h"
+#include "mpegvideo.h"
+
+#ifndef kVTVideoDecoderSpecification_EnableHardwareAcceleratedVideoDecoder
+#  define kVTVideoDecoderSpecification_EnableHardwareAcceleratedVideoDecoder CFSTR("EnableHardwareAcceleratedVideoDecoder")
+#endif
+
+#define VIDEOTOOLBOX_ESDS_EXTRADATA_PADDING  12
+
+static void videotoolbox_buffer_release(void *opaque, uint8_t *data)
+{
+    CVPixelBufferRef cv_buffer = (CVImageBufferRef)data;
+    CVPixelBufferRelease(cv_buffer);
+}
+
+static int videotoolbox_buffer_copy(VTContext *vtctx,
+                                    const uint8_t *buffer,
+                                    uint32_t size)
+{
+    void *tmp;
+
+    tmp = av_fast_realloc(vtctx->bitstream,
+                         &vtctx->allocated_size,
+                         size);
+
+    if (!tmp)
+        return AVERROR(ENOMEM);
+
+    vtctx->bitstream = tmp;
+    memcpy(vtctx->bitstream, buffer, size);
+    vtctx->bitstream_size = size;
+
+    return 0;
+}
+
+int ff_videotoolbox_alloc_frame(AVCodecContext *avctx, AVFrame *frame)
+{
+    frame->width  = avctx->width;
+    frame->height = avctx->height;
+    frame->format = avctx->pix_fmt;
+    frame->buf[0] = av_buffer_alloc(1);
+
+    if (!frame->buf[0])
+        return AVERROR(ENOMEM);
+
+    return 0;
+}
+
+CFDataRef ff_videotoolbox_avcc_extradata_create(AVCodecContext *avctx)
+{
+    CFDataRef data = NULL;
+
+    /* Each VCL NAL in the bistream sent to the decoder
+     * is preceded by a 4 bytes length header.
+     * Change the avcC atom header if needed, to signal headers of 4 bytes. */
+    if (avctx->extradata_size >= 4 && (avctx->extradata[4] & 0x03) != 0x03) {
+        uint8_t *rw_extradata = av_memdup(avctx->extradata, avctx->extradata_size);
+
+        if (!rw_extradata)
+            return NULL;
+
+        rw_extradata[4] |= 0x03;
+
+        data = CFDataCreate(kCFAllocatorDefault, rw_extradata, avctx->extradata_size);
+
+        av_freep(&rw_extradata);
+    } else {
+        data = CFDataCreate(kCFAllocatorDefault, avctx->extradata, avctx->extradata_size);
+    }
+
+    return data;
+}
+
+int ff_videotoolbox_buffer_create(VTContext *vtctx, AVFrame *frame)
+{
+    av_buffer_unref(&frame->buf[0]);
+
+    frame->buf[0] = av_buffer_create((uint8_t*)vtctx->frame,
+                                     sizeof(vtctx->frame),
+                                     videotoolbox_buffer_release,
+                                     NULL,
+                                     AV_BUFFER_FLAG_READONLY);
+    if (!frame->buf[0]) {
+        return AVERROR(ENOMEM);
+    }
+
+    frame->data[3] = (uint8_t*)vtctx->frame;
+    vtctx->frame = NULL;
+
+    return 0;
+}
+
+int ff_videotoolbox_h264_start_frame(AVCodecContext *avctx,
+                                     const uint8_t *buffer,
+                                     uint32_t size)
+{
+    VTContext *vtctx = avctx->internal->hwaccel_priv_data;
+    H264Context *h  = avctx->priv_data;
+
+    vtctx->bitstream_size = 0;
+
+    if (h->is_avc == 1) {
+        return videotoolbox_buffer_copy(vtctx, buffer, size);
+    }
+
+    return 0;
+}
+
+int ff_videotoolbox_h264_decode_slice(AVCodecContext *avctx,
+                                      const uint8_t *buffer,
+                                      uint32_t size)
+{
+    VTContext *vtctx = avctx->internal->hwaccel_priv_data;
+    H264Context *h  = avctx->priv_data;
+    void *tmp;
+
+    if (h->is_avc == 1)
+        return 0;
+
+    tmp = av_fast_realloc(vtctx->bitstream,
+                          &vtctx->allocated_size,
+                          vtctx->bitstream_size+size+4);
+    if (!tmp)
+        return AVERROR(ENOMEM);
+
+    vtctx->bitstream = tmp;
+
+    AV_WB32(vtctx->bitstream + vtctx->bitstream_size, size);
+    memcpy(vtctx->bitstream + vtctx->bitstream_size + 4, buffer, size);
+
+    vtctx->bitstream_size += size + 4;
+
+    return 0;
+}
+
+int ff_videotoolbox_uninit(AVCodecContext *avctx)
+{
+    VTContext *vtctx = avctx->internal->hwaccel_priv_data;
+    if (vtctx) {
+        av_freep(&vtctx->bitstream);
+        if (vtctx->frame)
+            CVPixelBufferRelease(vtctx->frame);
+    }
+
+    return 0;
+}
+
+#if CONFIG_VIDEOTOOLBOX
+static void videotoolbox_write_mp4_descr_length(PutByteContext *pb, int length)
+{
+    int i;
+    uint8_t b;
+
+    for (i = 3; i >= 0; i--) {
+        b = (length >> (i * 7)) & 0x7F;
+        if (i != 0)
+            b |= 0x80;
+
+        bytestream2_put_byteu(pb, b);
+    }
+}
+
+static CFDataRef videotoolbox_esds_extradata_create(AVCodecContext *avctx)
+{
+    CFDataRef data;
+    uint8_t *rw_extradata;
+    PutByteContext pb;
+    int full_size = 3 + 5 + 13 + 5 + avctx->extradata_size + 3;
+    // ES_DescrTag data + DecoderConfigDescrTag + data + DecSpecificInfoTag + size + SLConfigDescriptor
+    int config_size = 13 + 5 + avctx->extradata_size;
+    int s;
+
+    if (!(rw_extradata = av_mallocz(full_size + VIDEOTOOLBOX_ESDS_EXTRADATA_PADDING)))
+        return NULL;
+
+    bytestream2_init_writer(&pb, rw_extradata, full_size + VIDEOTOOLBOX_ESDS_EXTRADATA_PADDING);
+    bytestream2_put_byteu(&pb, 0);        // version
+    bytestream2_put_ne24(&pb, 0);         // flags
+
+    // elementary stream descriptor
+    bytestream2_put_byteu(&pb, 0x03);     // ES_DescrTag
+    videotoolbox_write_mp4_descr_length(&pb, full_size);
+    bytestream2_put_ne16(&pb, 0);         // esid
+    bytestream2_put_byteu(&pb, 0);        // stream priority (0-32)
+
+    // decoder configuration descriptor
+    bytestream2_put_byteu(&pb, 0x04);     // DecoderConfigDescrTag
+    videotoolbox_write_mp4_descr_length(&pb, config_size);
+    bytestream2_put_byteu(&pb, 32);       // object type indication. 32 = CODEC_ID_MPEG4
+    bytestream2_put_byteu(&pb, 0x11);     // stream type
+    bytestream2_put_ne24(&pb, 0);         // buffer size
+    bytestream2_put_ne32(&pb, 0);         // max bitrate
+    bytestream2_put_ne32(&pb, 0);         // avg bitrate
+
+    // decoder specific descriptor
+    bytestream2_put_byteu(&pb, 0x05);     ///< DecSpecificInfoTag
+    videotoolbox_write_mp4_descr_length(&pb, avctx->extradata_size);
+
+    bytestream2_put_buffer(&pb, avctx->extradata, avctx->extradata_size);
+
+    // SLConfigDescriptor
+    bytestream2_put_byteu(&pb, 0x06);     // SLConfigDescrTag
+    bytestream2_put_byteu(&pb, 0x01);     // length
+    bytestream2_put_byteu(&pb, 0x02);     //
+
+    s = bytestream2_size_p(&pb);
+
+    data = CFDataCreate(kCFAllocatorDefault, rw_extradata, s);
+
+    av_freep(&rw_extradata);
+    return data;
+}
+
+static CMSampleBufferRef videotoolbox_sample_buffer_create(CMFormatDescriptionRef fmt_desc,
+                                                           void *buffer,
+                                                           int size)
+{
+    OSStatus status;
+    CMBlockBufferRef  block_buf;
+    CMSampleBufferRef sample_buf;
+
+    block_buf  = NULL;
+    sample_buf = NULL;
+
+    status = CMBlockBufferCreateWithMemoryBlock(kCFAllocatorDefault,// structureAllocator
+                                                buffer,             // memoryBlock
+                                                size,               // blockLength
+                                                kCFAllocatorNull,   // blockAllocator
+                                                NULL,               // customBlockSource
+                                                0,                  // offsetToData
+                                                size,               // dataLength
+                                                0,                  // flags
+                                                &block_buf);
+
+    if (!status) {
+        status = CMSampleBufferCreate(kCFAllocatorDefault,  // allocator
+                                      block_buf,            // dataBuffer
+                                      TRUE,                 // dataReady
+                                      0,                    // makeDataReadyCallback
+                                      0,                    // makeDataReadyRefcon
+                                      fmt_desc,             // formatDescription
+                                      1,                    // numSamples
+                                      0,                    // numSampleTimingEntries
+                                      NULL,                 // sampleTimingArray
+                                      0,                    // numSampleSizeEntries
+                                      NULL,                 // sampleSizeArray
+                                      &sample_buf);
+    }
+
+    if (block_buf)
+        CFRelease(block_buf);
+
+    return sample_buf;
+}
+
+static void videotoolbox_decoder_callback(void *opaque,
+                                          void *sourceFrameRefCon,
+                                          OSStatus status,
+                                          VTDecodeInfoFlags flags,
+                                          CVImageBufferRef image_buffer,
+                                          CMTime pts,
+                                          CMTime duration)
+{
+    AVCodecContext *avctx = opaque;
+    VTContext *vtctx = avctx->internal->hwaccel_priv_data;
+
+    if (vtctx->frame) {
+        CVPixelBufferRelease(vtctx->frame);
+        vtctx->frame = NULL;
+    }
+
+    if (!image_buffer) {
+        av_log(NULL, AV_LOG_DEBUG, "vt decoder cb: output image buffer is null\n");
+        return;
+    }
+
+    vtctx->frame = CVPixelBufferRetain(image_buffer);
+}
+
+static OSStatus videotoolbox_session_decode_frame(AVCodecContext *avctx)
+{
+    OSStatus status;
+    CMSampleBufferRef sample_buf;
+    AVVideotoolboxContext *videotoolbox = avctx->hwaccel_context;
+    VTContext *vtctx = avctx->internal->hwaccel_priv_data;
+
+    sample_buf = videotoolbox_sample_buffer_create(videotoolbox->cm_fmt_desc,
+                                                   vtctx->bitstream,
+                                                   vtctx->bitstream_size);
+
+    if (!sample_buf)
+        return -1;
+
+    status = VTDecompressionSessionDecodeFrame(videotoolbox->session,
+                                               sample_buf,
+                                               0,       // decodeFlags
+                                               NULL,    // sourceFrameRefCon
+                                               0);      // infoFlagsOut
+    if (status == noErr)
+        status = VTDecompressionSessionWaitForAsynchronousFrames(videotoolbox->session);
+
+    CFRelease(sample_buf);
+
+    return status;
+}
+
+static int videotoolbox_common_end_frame(AVCodecContext *avctx, AVFrame *frame)
+{
+    int status;
+    AVVideotoolboxContext *videotoolbox = avctx->hwaccel_context;
+    VTContext *vtctx = avctx->internal->hwaccel_priv_data;
+
+    if (!videotoolbox->session || !vtctx->bitstream)
+        return AVERROR_INVALIDDATA;
+
+    status = videotoolbox_session_decode_frame(avctx);
+
+    if (status) {
+        av_log(avctx, AV_LOG_ERROR, "Failed to decode frame (%d)\n", status);
+        return AVERROR_UNKNOWN;
+    }
+
+    if (!vtctx->frame)
+        return AVERROR_UNKNOWN;
+
+    return ff_videotoolbox_buffer_create(vtctx, frame);
+}
+
+static int videotoolbox_h264_end_frame(AVCodecContext *avctx)
+{
+    H264Context *h = avctx->priv_data;
+    AVFrame *frame = h->cur_pic_ptr->f;
+
+    return videotoolbox_common_end_frame(avctx, frame);
+}
+
+static int videotoolbox_mpeg_start_frame(AVCodecContext *avctx,
+                                         const uint8_t *buffer,
+                                         uint32_t size)
+{
+    VTContext *vtctx = avctx->internal->hwaccel_priv_data;
+
+    return videotoolbox_buffer_copy(vtctx, buffer, size);
+}
+
+static int videotoolbox_mpeg_decode_slice(AVCodecContext *avctx,
+                                          const uint8_t *buffer,
+                                          uint32_t size)
+{
+    return 0;
+}
+
+static int videotoolbox_mpeg_end_frame(AVCodecContext *avctx)
+{
+    MpegEncContext *s = avctx->priv_data;
+    AVFrame *frame = s->current_picture_ptr->f;
+
+    return videotoolbox_common_end_frame(avctx, frame);
+}
+
+static CFDictionaryRef videotoolbox_decoder_config_create(CMVideoCodecType codec_type,
+                                                          AVCodecContext *avctx)
+{
+    CFMutableDictionaryRef config_info = CFDictionaryCreateMutable(kCFAllocatorDefault,
+                                                                   1,
+                                                                   &kCFTypeDictionaryKeyCallBacks,
+                                                                   &kCFTypeDictionaryValueCallBacks);
+
+    CFDictionarySetValue(config_info,
+                         kVTVideoDecoderSpecification_EnableHardwareAcceleratedVideoDecoder,
+                         kCFBooleanTrue);
+
+    if (avctx->extradata_size) {
+        CFMutableDictionaryRef avc_info;
+        CFDataRef data = NULL;
+
+        avc_info = CFDictionaryCreateMutable(kCFAllocatorDefault,
+                                             1,
+                                             &kCFTypeDictionaryKeyCallBacks,
+                                             &kCFTypeDictionaryValueCallBacks);
+
+        switch (codec_type) {
+        case kCMVideoCodecType_MPEG4Video :
+            data = videotoolbox_esds_extradata_create(avctx);
+            if (data)
+                CFDictionarySetValue(avc_info, CFSTR("esds"), data);
+            break;
+        case kCMVideoCodecType_H264 :
+            data = ff_videotoolbox_avcc_extradata_create(avctx);
+            if (data)
+                CFDictionarySetValue(avc_info, CFSTR("avcC"), data);
+            break;
+        default:
+            break;
+        }
+
+        CFDictionarySetValue(config_info,
+                kCMFormatDescriptionExtension_SampleDescriptionExtensionAtoms,
+                avc_info);
+
+        if (data)
+            CFRelease(data);
+
+        CFRelease(avc_info);
+    }
+    return config_info;
+}
+
+static CFDictionaryRef videotoolbox_buffer_attributes_create(int width,
+                                                             int height,
+                                                             OSType pix_fmt)
+{
+    CFMutableDictionaryRef buffer_attributes;
+    CFMutableDictionaryRef io_surface_properties;
+    CFNumberRef cv_pix_fmt;
+    CFNumberRef w;
+    CFNumberRef h;
+
+    w = CFNumberCreate(kCFAllocatorDefault, kCFNumberSInt32Type, &width);
+    h = CFNumberCreate(kCFAllocatorDefault, kCFNumberSInt32Type, &height);
+    cv_pix_fmt = CFNumberCreate(kCFAllocatorDefault, kCFNumberSInt32Type, &pix_fmt);
+
+    buffer_attributes = CFDictionaryCreateMutable(kCFAllocatorDefault,
+                                                  4,
+                                                  &kCFTypeDictionaryKeyCallBacks,
+                                                  &kCFTypeDictionaryValueCallBacks);
+    io_surface_properties = CFDictionaryCreateMutable(kCFAllocatorDefault,
+                                                      0,
+                                                      &kCFTypeDictionaryKeyCallBacks,
+                                                      &kCFTypeDictionaryValueCallBacks);
+
+    CFDictionarySetValue(buffer_attributes, kCVPixelBufferPixelFormatTypeKey, cv_pix_fmt);
+    CFDictionarySetValue(buffer_attributes, kCVPixelBufferIOSurfacePropertiesKey, io_surface_properties);
+    CFDictionarySetValue(buffer_attributes, kCVPixelBufferWidthKey, w);
+    CFDictionarySetValue(buffer_attributes, kCVPixelBufferHeightKey, h);
+
+    CFRelease(io_surface_properties);
+    CFRelease(cv_pix_fmt);
+    CFRelease(w);
+    CFRelease(h);
+
+    return buffer_attributes;
+}
+
+static CMVideoFormatDescriptionRef videotoolbox_format_desc_create(CMVideoCodecType codec_type,
+                                                                   CFDictionaryRef decoder_spec,
+                                                                   int width,
+                                                                   int height)
+{
+    CMFormatDescriptionRef cm_fmt_desc;
+    OSStatus status;
+
+    status = CMVideoFormatDescriptionCreate(kCFAllocatorDefault,
+                                            codec_type,
+                                            width,
+                                            height,
+                                            decoder_spec, // Dictionary of extension
+                                            &cm_fmt_desc);
+
+    if (status)
+        return NULL;
+
+    return cm_fmt_desc;
+}
+
+static int videotoolbox_default_init(AVCodecContext *avctx)
+{
+    AVVideotoolboxContext *videotoolbox = avctx->hwaccel_context;
+    OSStatus status;
+    VTDecompressionOutputCallbackRecord decoder_cb;
+    CFDictionaryRef decoder_spec;
+    CFDictionaryRef buf_attr;
+    int32_t pix_fmt;
+
+    if (!videotoolbox) {
+        av_log(avctx, AV_LOG_ERROR, "hwaccel context is not set\n");
+        return -1;
+    }
+
+    switch( avctx->codec_id ) {
+    case AV_CODEC_ID_H263 :
+        videotoolbox->cm_codec_type = kCMVideoCodecType_H263;
+        break;
+    case AV_CODEC_ID_H264 :
+        videotoolbox->cm_codec_type = kCMVideoCodecType_H264;
+        break;
+    case AV_CODEC_ID_MPEG1VIDEO :
+        videotoolbox->cm_codec_type = kCMVideoCodecType_MPEG1Video;
+        break;
+    case AV_CODEC_ID_MPEG2VIDEO :
+        videotoolbox->cm_codec_type = kCMVideoCodecType_MPEG2Video;
+        break;
+    case AV_CODEC_ID_MPEG4 :
+        videotoolbox->cm_codec_type = kCMVideoCodecType_MPEG4Video;
+        break;
+    default :
+        break;
+    }
+
+    pix_fmt = videotoolbox->cv_pix_fmt_type;
+
+    decoder_spec = videotoolbox_decoder_config_create(videotoolbox->cm_codec_type, avctx);
+
+    videotoolbox->cm_fmt_desc = videotoolbox_format_desc_create(videotoolbox->cm_codec_type,
+                                                                decoder_spec,
+                                                                avctx->width,
+                                                                avctx->height);
+    if (!videotoolbox->cm_fmt_desc) {
+        if (decoder_spec)
+            CFRelease(decoder_spec);
+
+        av_log(avctx, AV_LOG_ERROR, "format description creation failed\n");
+        return -1;
+    }
+
+    buf_attr = videotoolbox_buffer_attributes_create(avctx->width,
+                                                     avctx->height,
+                                                     videotoolbox->cv_pix_fmt_type);
+
+    decoder_cb.decompressionOutputCallback = videotoolbox_decoder_callback;
+    decoder_cb.decompressionOutputRefCon   = avctx;
+
+    status = VTDecompressionSessionCreate(NULL,                      // allocator
+                                          videotoolbox->cm_fmt_desc, // videoFormatDescription
+                                          decoder_spec,              // videoDecoderSpecification
+                                          buf_attr,                  // destinationImageBufferAttributes
+                                          &decoder_cb,               // outputCallback
+                                          &videotoolbox->session);   // decompressionSessionOut
+
+    if (decoder_spec)
+        CFRelease(decoder_spec);
+    if (buf_attr)
+        CFRelease(buf_attr);
+
+    switch (status) {
+    case kVTVideoDecoderNotAvailableNowErr:
+    case kVTVideoDecoderUnsupportedDataFormatErr:
+        return AVERROR(ENOSYS);
+    case kVTVideoDecoderMalfunctionErr:
+        return AVERROR(EINVAL);
+    case kVTVideoDecoderBadDataErr :
+        return AVERROR_INVALIDDATA;
+    case 0:
+        return 0;
+    default:
+        return AVERROR_UNKNOWN;
+    }
+}
+
+static void videotoolbox_default_free(AVCodecContext *avctx)
+{
+    AVVideotoolboxContext *videotoolbox = avctx->hwaccel_context;
+
+    if (videotoolbox) {
+        if (videotoolbox->cm_fmt_desc)
+            CFRelease(videotoolbox->cm_fmt_desc);
+
+        if (videotoolbox->session)
+            VTDecompressionSessionInvalidate(videotoolbox->session);
+    }
+}
+
+AVHWAccel ff_h263_videotoolbox_hwaccel = {
+    .name           = "h263_videotoolbox",
+    .type           = AVMEDIA_TYPE_VIDEO,
+    .id             = AV_CODEC_ID_H263,
+    .pix_fmt        = AV_PIX_FMT_VIDEOTOOLBOX,
+    .alloc_frame    = ff_videotoolbox_alloc_frame,
+    .start_frame    = videotoolbox_mpeg_start_frame,
+    .decode_slice   = videotoolbox_mpeg_decode_slice,
+    .end_frame      = videotoolbox_mpeg_end_frame,
+    .uninit         = ff_videotoolbox_uninit,
+    .priv_data_size = sizeof(VTContext),
+};
+
+AVHWAccel ff_h264_videotoolbox_hwaccel = {
+    .name           = "h264_videotoolbox",
+    .type           = AVMEDIA_TYPE_VIDEO,
+    .id             = AV_CODEC_ID_H264,
+    .pix_fmt        = AV_PIX_FMT_VIDEOTOOLBOX,
+    .alloc_frame    = ff_videotoolbox_alloc_frame,
+    .start_frame    = ff_videotoolbox_h264_start_frame,
+    .decode_slice   = ff_videotoolbox_h264_decode_slice,
+    .end_frame      = videotoolbox_h264_end_frame,
+    .uninit         = ff_videotoolbox_uninit,
+    .priv_data_size = sizeof(VTContext),
+};
+
+AVHWAccel ff_mpeg1_videotoolbox_hwaccel = {
+    .name           = "mpeg1_videotoolbox",
+    .type           = AVMEDIA_TYPE_VIDEO,
+    .id             = AV_CODEC_ID_MPEG1VIDEO,
+    .pix_fmt        = AV_PIX_FMT_VIDEOTOOLBOX,
+    .alloc_frame    = ff_videotoolbox_alloc_frame,
+    .start_frame    = videotoolbox_mpeg_start_frame,
+    .decode_slice   = videotoolbox_mpeg_decode_slice,
+    .end_frame      = videotoolbox_mpeg_end_frame,
+    .uninit         = ff_videotoolbox_uninit,
+    .priv_data_size = sizeof(VTContext),
+};
+
+AVHWAccel ff_mpeg2_videotoolbox_hwaccel = {
+    .name           = "mpeg2_videotoolbox",
+    .type           = AVMEDIA_TYPE_VIDEO,
+    .id             = AV_CODEC_ID_MPEG2VIDEO,
+    .pix_fmt        = AV_PIX_FMT_VIDEOTOOLBOX,
+    .alloc_frame    = ff_videotoolbox_alloc_frame,
+    .start_frame    = videotoolbox_mpeg_start_frame,
+    .decode_slice   = videotoolbox_mpeg_decode_slice,
+    .end_frame      = videotoolbox_mpeg_end_frame,
+    .uninit         = ff_videotoolbox_uninit,
+    .priv_data_size = sizeof(VTContext),
+};
+
+AVHWAccel ff_mpeg4_videotoolbox_hwaccel = {
+    .name           = "mpeg4_videotoolbox",
+    .type           = AVMEDIA_TYPE_VIDEO,
+    .id             = AV_CODEC_ID_MPEG4,
+    .pix_fmt        = AV_PIX_FMT_VIDEOTOOLBOX,
+    .alloc_frame    = ff_videotoolbox_alloc_frame,
+    .start_frame    = videotoolbox_mpeg_start_frame,
+    .decode_slice   = videotoolbox_mpeg_decode_slice,
+    .end_frame      = videotoolbox_mpeg_end_frame,
+    .uninit         = ff_videotoolbox_uninit,
+    .priv_data_size = sizeof(VTContext),
+};
+
+AVVideotoolboxContext *av_videotoolbox_alloc_context(void)
+{
+    AVVideotoolboxContext *ret = av_mallocz(sizeof(*ret));
+
+    if (ret) {
+        ret->output_callback = videotoolbox_decoder_callback;
+        ret->cv_pix_fmt_type = kCVPixelFormatType_420YpCbCr8BiPlanarVideoRange;
+    }
+
+    return ret;
+}
+
+int av_videotoolbox_default_init(AVCodecContext *avctx)
+{
+    return av_videotoolbox_default_init2(avctx, NULL);
+}
+
+int av_videotoolbox_default_init2(AVCodecContext *avctx, AVVideotoolboxContext *vtctx)
+{
+    avctx->hwaccel_context = vtctx ?: av_videotoolbox_alloc_context();
+    if (!avctx->hwaccel_context)
+        return AVERROR(ENOMEM);
+    return videotoolbox_default_init(avctx);
+}
+
+void av_videotoolbox_default_free(AVCodecContext *avctx)
+{
+
+    videotoolbox_default_free(avctx);
+    av_freep(&avctx->hwaccel_context);
+}
+#endif /* CONFIG_VIDEOTOOLBOX */
diff --git a/libavcodec/videotoolbox.h b/libavcodec/videotoolbox.h
new file mode 100644
index 0000000000..a48638e2b2
--- /dev/null
+++ b/libavcodec/videotoolbox.h
@@ -0,0 +1,126 @@
+/*
+ * Videotoolbox hardware acceleration
+ *
+ * copyright (c) 2012 Sebastien Zwickert
+ *
+ * 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
+ */
+
+#ifndef AVCODEC_VIDEOTOOLBOX_H
+#define AVCODEC_VIDEOTOOLBOX_H
+
+/**
+ * @file
+ * @ingroup lavc_codec_hwaccel_videotoolbox
+ * Public libavcodec Videotoolbox header.
+ */
+
+#include <stdint.h>
+
+#define Picture QuickdrawPicture
+#include <VideoToolbox/VideoToolbox.h>
+#undef Picture
+
+#include "libavcodec/avcodec.h"
+
+/**
+ * This struct holds all the information that needs to be passed
+ * between the caller and libavcodec for initializing Videotoolbox decoding.
+ * Its size is not a part of the public ABI, it must be allocated with
+ * av_videotoolbox_alloc_context() and freed with av_free().
+ */
+typedef struct AVVideotoolboxContext {
+    /**
+     * Videotoolbox decompression session object.
+     * Created and freed the caller.
+     */
+    VTDecompressionSessionRef session;
+
+    /**
+     * The output callback that must be passed to the session.
+     * Set by av_videottoolbox_default_init()
+     */
+    VTDecompressionOutputCallback output_callback;
+
+    /**
+     * CVPixelBuffer Format Type that Videotoolbox will use for decoded frames.
+     * set by the caller.
+     */
+    OSType cv_pix_fmt_type;
+
+    /**
+     * CoreMedia Format Description that Videotoolbox will use to create the decompression session.
+     * Set by the caller.
+     */
+    CMVideoFormatDescriptionRef cm_fmt_desc;
+
+    /**
+     * CoreMedia codec type that Videotoolbox will use to create the decompression session.
+     * Set by the caller.
+     */
+    int cm_codec_type;
+} AVVideotoolboxContext;
+
+/**
+ * Allocate and initialize a Videotoolbox context.
+ *
+ * This function should be called from the get_format() callback when the caller
+ * selects the AV_PIX_FMT_VIDETOOLBOX format. The caller must then create
+ * the decoder object (using the output callback provided by libavcodec) that
+ * will be used for Videotoolbox-accelerated decoding.
+ *
+ * When decoding with Videotoolbox is finished, the caller must destroy the decoder
+ * object and free the Videotoolbox context using av_free().
+ *
+ * @return the newly allocated context or NULL on failure
+ */
+AVVideotoolboxContext *av_videotoolbox_alloc_context(void);
+
+/**
+ * This is a convenience function that creates and sets up the Videotoolbox context using
+ * an internal implementation.
+ *
+ * @param avctx the corresponding codec context
+ *
+ * @return >= 0 on success, a negative AVERROR code on failure
+ */
+int av_videotoolbox_default_init(AVCodecContext *avctx);
+
+/**
+ * This is a convenience function that creates and sets up the Videotoolbox context using
+ * an internal implementation.
+ *
+ * @param avctx the corresponding codec context
+ * @param vtctx the Videotoolbox context to use
+ *
+ * @return >= 0 on success, a negative AVERROR code on failure
+ */
+int av_videotoolbox_default_init2(AVCodecContext *avctx, AVVideotoolboxContext *vtctx);
+
+/**
+ * This function must be called to free the Videotoolbox context initialized with
+ * av_videotoolbox_default_init().
+ *
+ * @param avctx the corresponding codec context
+ */
+void av_videotoolbox_default_free(AVCodecContext *avctx);
+
+/**
+ * @}
+ */
+
+#endif /* AVCODEC_VIDEOTOOLBOX_H */
diff --git a/libavutil/pixdesc.c b/libavutil/pixdesc.c
index 1699646093..eb0dc9d311 100644
--- a/libavutil/pixdesc.c
+++ b/libavutil/pixdesc.c
@@ -1632,6 +1632,10 @@ const AVPixFmtDescriptor av_pix_fmt_descriptors[AV_PIX_FMT_NB] = {
         },
         .flags = AV_PIX_FMT_FLAG_BE | AV_PIX_FMT_FLAG_ALPHA,
     },
+    [AV_PIX_FMT_VIDEOTOOLBOX] = {
+        .name = "videotoolbox_vld",
+        .flags = AV_PIX_FMT_FLAG_HWACCEL,
+    },
     [AV_PIX_FMT_GBRP] = {
         .name = "gbrp",
         .nb_components = 3,
diff --git a/libavutil/pixfmt.h b/libavutil/pixfmt.h
index a6c8d6f7da..5cdbeb9678 100644
--- a/libavutil/pixfmt.h
+++ b/libavutil/pixfmt.h
@@ -311,6 +311,8 @@ enum AVPixelFormat {
     AV_PIX_FMT_AYUV64LE,    ///< packed AYUV 4:4:4,64bpp (1 Cr & Cb sample per 1x1 Y & A samples), little-endian
     AV_PIX_FMT_AYUV64BE,    ///< packed AYUV 4:4:4,64bpp (1 Cr & Cb sample per 1x1 Y & A samples), big-endian
 
+    AV_PIX_FMT_VIDEOTOOLBOX, ///< hardware decoding through Videotoolbox
+
     AV_PIX_FMT_NB,        ///< number of pixel formats, DO NOT USE THIS if you want to link with shared libav* because the number of formats might differ between versions
 
 #if FF_API_PIX_FMT