Merge branch 'master' of git://git.ffmpeg.org/ffmpeg

Merged-by: Michael Niedermayer <michaelni@gmx.at>

6 files changed, 112 insertions, 95 deletions

diff --git a/libavcodec/Makefile b/libavcodec/Makefile
index 2a2197d1fb..eefe60c772 100644
--- a/libavcodec/Makefile
+++ b/libavcodec/Makefile
@@ -26,6 +26,7 @@ OBJS = allcodecs.o                                                      \
 
 # parts needed for many different codecs
 OBJS-$(CONFIG_AANDCT)                  += aandcttab.o
+OBJS-$(CONFIG_AC3DSP)                  += ac3dsp.o
 OBJS-$(CONFIG_CRYSTALHD)               += crystalhd.o
 OBJS-$(CONFIG_ENCODERS)                += faandct.o jfdctfst.o jfdctint.o
 OBJS-$(CONFIG_DCT)                     += dct.o
@@ -56,10 +57,8 @@ OBJS-$(CONFIG_AAC_ENCODER)             += aacenc.o aaccoder.o    \
                                           mpeg4audio.o
 OBJS-$(CONFIG_AASC_DECODER)            += aasc.o msrledec.o
 OBJS-$(CONFIG_AC3_DECODER)             += ac3dec.o ac3dec_data.o ac3.o
-OBJS-$(CONFIG_AC3_ENCODER)             += ac3enc_float.o ac3tab.o ac3.o \
-                                          ac3dsp.o
-OBJS-$(CONFIG_AC3_FIXED_ENCODER)       += ac3enc_fixed.o ac3tab.o ac3.o \
-                                          ac3dsp.o
+OBJS-$(CONFIG_AC3_ENCODER)             += ac3enc_float.o ac3tab.o ac3.o
+OBJS-$(CONFIG_AC3_FIXED_ENCODER)       += ac3enc_fixed.o ac3tab.o ac3.o
 OBJS-$(CONFIG_ALAC_DECODER)            += alac.o
 OBJS-$(CONFIG_ALAC_ENCODER)            += alacenc.o
 OBJS-$(CONFIG_ALS_DECODER)             += alsdec.o bgmc.o mpeg4audio.o
diff --git a/libavcodec/avcodec.h b/libavcodec/avcodec.h
index 3e85567982..3ea1e76989 100644
--- a/libavcodec/avcodec.h
+++ b/libavcodec/avcodec.h
@@ -2323,6 +2323,11 @@ typedef struct AVCodecContext {
 #define FF_PROFILE_H264_HIGH_444_INTRA       (244|FF_PROFILE_H264_INTRA)
 #define FF_PROFILE_H264_CAVLC_444            44
 
+#define FF_PROFILE_VC1_SIMPLE   0
+#define FF_PROFILE_VC1_MAIN     1
+#define FF_PROFILE_VC1_COMPLEX  2
+#define FF_PROFILE_VC1_ADVANCED 3
+
     /**
      * level
      * - encoding: Set by user.
diff --git a/libavcodec/vc1dec.c b/libavcodec/vc1dec.c
index b27b6d01c1..1a437a892a 100644
--- a/libavcodec/vc1dec.c
+++ b/libavcodec/vc1dec.c
@@ -3098,6 +3098,10 @@ static av_cold int vc1_decode_init(AVCodecContext *avctx)
             return -1;
         }
     }
+    avctx->profile = v->profile;
+    if (v->profile == PROFILE_ADVANCED)
+        avctx->level = v->level;
+
     avctx->has_b_frames= !!(avctx->max_b_frames);
     s->low_delay = !avctx->has_b_frames;
 
@@ -3345,6 +3349,13 @@ static av_cold int vc1_decode_end(AVCodecContext *avctx)
     return 0;
 }
 
+static const AVProfile profiles[] = {
+    { FF_PROFILE_VC1_SIMPLE,   "Simple"   },
+    { FF_PROFILE_VC1_MAIN,     "Main"     },
+    { FF_PROFILE_VC1_COMPLEX,  "Complex"  },
+    { FF_PROFILE_VC1_ADVANCED, "Advanced" },
+    { FF_PROFILE_UNKNOWN },
+};
 
 AVCodec ff_vc1_decoder = {
     "vc1",
@@ -3358,7 +3369,8 @@ AVCodec ff_vc1_decoder = {
     CODEC_CAP_DR1 | CODEC_CAP_DELAY,
     NULL,
     .long_name = NULL_IF_CONFIG_SMALL("SMPTE VC-1"),
-    .pix_fmts = ff_hwaccel_pixfmt_list_420
+    .pix_fmts = ff_hwaccel_pixfmt_list_420,
+    .profiles = NULL_IF_CONFIG_SMALL(profiles)
 };
 
 #if CONFIG_WMV3_DECODER
@@ -3374,7 +3386,8 @@ AVCodec ff_wmv3_decoder = {
     CODEC_CAP_DR1 | CODEC_CAP_DELAY,
     NULL,
     .long_name = NULL_IF_CONFIG_SMALL("Windows Media Video 9"),
-    .pix_fmts = ff_hwaccel_pixfmt_list_420
+    .pix_fmts = ff_hwaccel_pixfmt_list_420,
+    .profiles = NULL_IF_CONFIG_SMALL(profiles)
 };
 #endif
 
@@ -3391,7 +3404,8 @@ AVCodec ff_wmv3_vdpau_decoder = {
     CODEC_CAP_DR1 | CODEC_CAP_DELAY | CODEC_CAP_HWACCEL_VDPAU,
     NULL,
     .long_name = NULL_IF_CONFIG_SMALL("Windows Media Video 9 VDPAU"),
-    .pix_fmts = (const enum PixelFormat[]){PIX_FMT_VDPAU_WMV3, PIX_FMT_NONE}
+    .pix_fmts = (const enum PixelFormat[]){PIX_FMT_VDPAU_WMV3, PIX_FMT_NONE},
+    .profiles = NULL_IF_CONFIG_SMALL(profiles)
 };
 #endif
 
@@ -3408,6 +3422,7 @@ AVCodec ff_vc1_vdpau_decoder = {
     CODEC_CAP_DR1 | CODEC_CAP_DELAY | CODEC_CAP_HWACCEL_VDPAU,
     NULL,
     .long_name = NULL_IF_CONFIG_SMALL("SMPTE VC-1 VDPAU"),
-    .pix_fmts = (const enum PixelFormat[]){PIX_FMT_VDPAU_VC1, PIX_FMT_NONE}
+    .pix_fmts = (const enum PixelFormat[]){PIX_FMT_VDPAU_VC1, PIX_FMT_NONE},
+    .profiles = NULL_IF_CONFIG_SMALL(profiles)
 };
 #endif
diff --git a/libavcodec/vp8.c b/libavcodec/vp8.c
index f71a7b7783..42f401d2b1 100644
--- a/libavcodec/vp8.c
+++ b/libavcodec/vp8.c
@@ -474,7 +474,7 @@ void decode_mvs(VP8Context *s, VP8Macroblock *mb, int mb_x, int mb_y)
     enum { EDGE_TOP, EDGE_LEFT, EDGE_TOPLEFT };
     int idx = CNT_ZERO;
     int cur_sign_bias = s->sign_bias[mb->ref_frame];
-    int *sign_bias = s->sign_bias;
+    int8_t *sign_bias = s->sign_bias;
     VP56mv near_mv[4];
     uint8_t cnt[4] = { 0 };
     VP56RangeCoder *c = &s->c;
@@ -640,7 +640,7 @@ void decode_mb_mode(VP8Context *s, VP8Macroblock *mb, int mb_x, int mb_y, uint8_
  *         otherwise, the index of the last coeff decoded plus one
  */
 static int decode_block_coeffs_internal(VP56RangeCoder *c, DCTELEM block[16],
-                                        uint8_t probs[8][3][NUM_DCT_TOKENS-1],
+                                        uint8_t probs[16][3][NUM_DCT_TOKENS-1],
                                         int i, uint8_t *token_prob, int16_t qmul[2])
 {
     goto skip_eob;
@@ -1282,7 +1282,7 @@ static av_always_inline void idct_mb(VP8Context *s, uint8_t *dst[3], VP8Macroblo
                             s->vp8dsp.vp8_idct_add(ch_dst+4*x, s->block[4+ch][(y<<1)+x], s->uvlinesize);
                         nnz4 >>= 8;
                         if (!nnz4)
-                            break;
+                            goto chroma_idct_end;
                     }
                     ch_dst += 4*s->uvlinesize;
                 }
@@ -1290,6 +1290,7 @@ static av_always_inline void idct_mb(VP8Context *s, uint8_t *dst[3], VP8Macroblo
                 s->vp8dsp.vp8_idct_dc_add4uv(ch_dst, s->block[4+ch], s->uvlinesize);
             }
         }
+chroma_idct_end: ;
     }
 }
 
diff --git a/libavcodec/vp8.h b/libavcodec/vp8.h
index d3b8705ea0..2db056f073 100644
--- a/libavcodec/vp8.h
+++ b/libavcodec/vp8.h
@@ -85,83 +85,24 @@ typedef struct {
 
 typedef struct {
     AVCodecContext *avctx;
-    DSPContext dsp;
-    VP8DSPContext vp8dsp;
-    H264PredContext hpc;
-    vp8_mc_func put_pixels_tab[3][3][3];
-    AVFrame frames[4];
     AVFrame *framep[4];
     uint8_t *edge_emu_buffer;
-    VP56RangeCoder c;   ///< header context, includes mb modes and motion vectors
-    int profile;
 
-    int mb_width;   /* number of horizontal MB */
-    int mb_height;  /* number of vertical MB */
+    uint16_t mb_width;   /* number of horizontal MB */
+    uint16_t mb_height;  /* number of vertical MB */
     int linesize;
     int uvlinesize;
 
-    int keyframe;
-    int invisible;
-    int update_last;    ///< update VP56_FRAME_PREVIOUS with the current one
-    int update_golden;  ///< VP56_FRAME_NONE if not updated, or which frame to copy if so
-    int update_altref;
-    int deblock_filter;
-
-    /**
-     * If this flag is not set, all the probability updates
-     * are discarded after this frame is decoded.
-     */
-    int update_probabilities;
-
-    /**
-     * All coefficients are contained in separate arith coding contexts.
-     * There can be 1, 2, 4, or 8 of these after the header context.
-     */
-    int num_coeff_partitions;
-    VP56RangeCoder coeff_partition[8];
-
-    VP8Macroblock *macroblocks;
-    VP8Macroblock *macroblocks_base;
-    VP8FilterStrength *filter_strength;
-
-    uint8_t *intra4x4_pred_mode_top;
-    uint8_t intra4x4_pred_mode_left[4];
-    uint8_t *segmentation_map;
-
-    /**
-     * Cache of the top row needed for intra prediction
-     * 16 for luma, 8 for each chroma plane
-     */
-    uint8_t (*top_border)[16+8+8];
-
-    /**
-     * For coeff decode, we need to know whether the above block had non-zero
-     * coefficients. This means for each macroblock, we need data for 4 luma
-     * blocks, 2 u blocks, 2 v blocks, and the luma dc block, for a total of 9
-     * per macroblock. We keep the last row in top_nnz.
-     */
-    uint8_t (*top_nnz)[9];
-    DECLARE_ALIGNED(8, uint8_t, left_nnz)[9];
-
-    /**
-     * This is the index plus one of the last non-zero coeff
-     * for each of the blocks in the current macroblock.
-     * So, 0 -> no coeffs
-     *     1 -> dc-only (special transform)
-     *     2+-> full transform
-     */
-    DECLARE_ALIGNED(16, uint8_t, non_zero_count_cache)[6][4];
-    DECLARE_ALIGNED(16, DCTELEM, block)[6][4][16];
-    DECLARE_ALIGNED(16, DCTELEM, block_dc)[16];
-    uint8_t intra4x4_pred_mode_mb[16];
-
-    int chroma_pred_mode;    ///< 8x8c pred mode of the current macroblock
-    int segment;             ///< segment of the current macroblock
+    uint8_t keyframe;
+    uint8_t deblock_filter;
+    uint8_t mbskip_enabled;
+    uint8_t segment;             ///< segment of the current macroblock
+    uint8_t chroma_pred_mode;    ///< 8x8c pred mode of the current macroblock
+    uint8_t profile;
     VP56mv mv_min;
     VP56mv mv_max;
 
-    int mbskip_enabled;
-    int sign_bias[4]; ///< one state [0, 1] per ref frame type
+    int8_t sign_bias[4]; ///< one state [0, 1] per ref frame type
     int ref_count[3];
 
     /**
@@ -170,13 +111,26 @@ typedef struct {
      * a frame, since the values persist between interframes.
      */
     struct {
-        int enabled;
-        int absolute_vals;
-        int update_map;
+        uint8_t enabled;
+        uint8_t absolute_vals;
+        uint8_t update_map;
         int8_t base_quant[4];
         int8_t filter_level[4];     ///< base loop filter level
     } segmentation;
 
+    struct {
+        uint8_t simple;
+        uint8_t level;
+        uint8_t sharpness;
+    } filter;
+
+    VP8Macroblock *macroblocks;
+    VP8FilterStrength *filter_strength;
+
+    uint8_t *intra4x4_pred_mode_top;
+    uint8_t intra4x4_pred_mode_left[4];
+    uint8_t *segmentation_map;
+
     /**
      * Macroblocks can have one of 4 different quants in a frame when
      * segmentation is enabled.
@@ -190,13 +144,7 @@ typedef struct {
     } qmat[4];
 
     struct {
-        int simple;
-        int level;
-        int sharpness;
-    } filter;
-
-    struct {
-        int enabled;    ///< whether each mb can have a different strength based on mode/ref
+        uint8_t enabled;    ///< whether each mb can have a different strength based on mode/ref
 
         /**
          * filter strength adjustment for the following macroblock modes:
@@ -220,6 +168,34 @@ typedef struct {
     } lf_delta;
 
     /**
+     * Cache of the top row needed for intra prediction
+     * 16 for luma, 8 for each chroma plane
+     */
+    uint8_t (*top_border)[16+8+8];
+
+    /**
+     * For coeff decode, we need to know whether the above block had non-zero
+     * coefficients. This means for each macroblock, we need data for 4 luma
+     * blocks, 2 u blocks, 2 v blocks, and the luma dc block, for a total of 9
+     * per macroblock. We keep the last row in top_nnz.
+     */
+    uint8_t (*top_nnz)[9];
+    DECLARE_ALIGNED(8, uint8_t, left_nnz)[9];
+
+    /**
+     * This is the index plus one of the last non-zero coeff
+     * for each of the blocks in the current macroblock.
+     * So, 0 -> no coeffs
+     *     1 -> dc-only (special transform)
+     *     2+-> full transform
+     */
+    DECLARE_ALIGNED(16, uint8_t, non_zero_count_cache)[6][4];
+    VP56RangeCoder c;   ///< header context, includes mb modes and motion vectors
+    DECLARE_ALIGNED(16, DCTELEM, block)[6][4][16];
+    DECLARE_ALIGNED(16, DCTELEM, block_dc)[16];
+    uint8_t intra4x4_pred_mode_mb[16];
+
+    /**
      * These are all of the updatable probabilities for binary decisions.
      * They are only implictly reset on keyframes, making it quite likely
      * for an interframe to desync if a prior frame's header was corrupt
@@ -233,10 +209,33 @@ typedef struct {
         uint8_t golden;
         uint8_t pred16x16[4];
         uint8_t pred8x8c[3];
-        /* Padded to allow overreads */
-        uint8_t token[4][17][3][NUM_DCT_TOKENS-1];
+        uint8_t token[4][16][3][NUM_DCT_TOKENS-1];
         uint8_t mvc[2][19];
     } prob[2];
+
+    VP8Macroblock *macroblocks_base;
+    int invisible;
+    int update_last;    ///< update VP56_FRAME_PREVIOUS with the current one
+    int update_golden;  ///< VP56_FRAME_NONE if not updated, or which frame to copy if so
+    int update_altref;
+
+    /**
+     * If this flag is not set, all the probability updates
+     * are discarded after this frame is decoded.
+     */
+    int update_probabilities;
+
+    /**
+     * All coefficients are contained in separate arith coding contexts.
+     * There can be 1, 2, 4, or 8 of these after the header context.
+     */
+    int num_coeff_partitions;
+    VP56RangeCoder coeff_partition[8];
+    DSPContext dsp;
+    VP8DSPContext vp8dsp;
+    H264PredContext hpc;
+    vp8_mc_func put_pixels_tab[3][3][3];
+    AVFrame frames[4];
 } VP8Context;
 
 #endif
diff --git a/libavcodec/x86/Makefile b/libavcodec/x86/Makefile
index 1b58fa121f..f8d456d3ea 100644
--- a/libavcodec/x86/Makefile
+++ b/libavcodec/x86/Makefile
@@ -17,10 +17,8 @@ MMX-OBJS-$(CONFIG_H264PRED)            += x86/h264_intrapred_init.o
 
 YASM-OBJS-$(CONFIG_VC1_DECODER)        += x86/vc1dsp_yasm.o
 
-MMX-OBJS-$(CONFIG_AC3_ENCODER)         += x86/ac3dsp_mmx.o
-MMX-OBJS-$(CONFIG_AC3_FIXED_ENCODER)   += x86/ac3dsp_mmx.o
-YASM-OBJS-$(CONFIG_AC3_ENCODER)        += x86/ac3dsp.o
-YASM-OBJS-$(CONFIG_AC3_FIXED_ENCODER)  += x86/ac3dsp.o
+MMX-OBJS-$(CONFIG_AC3DSP)              += x86/ac3dsp_mmx.o
+YASM-OBJS-$(CONFIG_AC3DSP)             += x86/ac3dsp.o
 MMX-OBJS-$(CONFIG_CAVS_DECODER)        += x86/cavsdsp_mmx.o
 MMX-OBJS-$(CONFIG_MP1FLOAT_DECODER)    += x86/mpegaudiodec_mmx.o
 MMX-OBJS-$(CONFIG_MP2FLOAT_DECODER)    += x86/mpegaudiodec_mmx.o