Merge remote-tracking branch 'qatar/master'

* qatar/master: (27 commits)
  cmdutils: use new avcodec_is_decoder/encoder() functions.
  lavc: make codec_is_decoder/encoder() public.
  lavc: deprecate AVCodecContext.sub_id.
  libcdio: add a forgotten AVClass to the private context.
  swscale: remove "cpu flags" from -sws_flags description.
  proresenc: give user a possibility to alter some encoding parameters
  vorbisenc: add output buffer overwrite protection
  libopencore-amrnbenc: fix end-of-stream handling
  ra144enc: fix end-of-stream handling
  nellymoserenc: zero any leftover packet bytes
  nellymoserenc: use proper MDCT overlap delay
  qpeg: Use bytestream2 functions to prevent buffer overreads.
  swscale: make %rep unconditional.
  vp8: convert simple loopfilter x86 assembly to use named arguments.
  vp8: convert idct x86 assembly to use named arguments.
  vp8: convert mc x86 assembly to use named arguments.
  vp8: convert loopfilter x86 assembly to use cpuflags().
  vp8: convert idct/mc x86 assembly to use cpuflags().
  swscale: remove now unnecessary hack.
  x86inc: don't "bake" stack_offset in named arguments.
  ...

Conflicts:
	cmdutils.c
	doc/APIchanges
	libavcodec/mpeg12.c
	libavcodec/options.c
	libavcodec/qpeg.c
	libavcodec/utils.c
	libavcodec/version.h
	libavdevice/libcdio.c
	tests/lavf-regression.sh

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

cmdutils.c

@@ -815,9 +815,9 @@ int opt_codecs(const char *opt, const char *arg)
                 decode = encode = cap = 0;
             }
             if (p2 && strcmp(p->name, p2->name) == 0) {
-                if (p->decode)
+                if (av_codec_is_decoder(p))
                     decode = 1;
-                if (p->encode || p->encode2)
+                if (av_codec_is_encoder(p))
                     encode = 1;
                 cap |= p->capabilities;
             }

doc/APIchanges

@@ -35,6 +35,9 @@ API changes, most recent first:
 2012-01-24 - xxxxxxx - lavfi 2.60.100
   Add avfilter_graph_dump.
 
+2012-03-xx - xxxxxxx - lavc 54.7.0 - avcodec.h
+  Add av_codec_is_encoder/decoder().
+
 2012-xx-xx - xxxxxxx - lavc 54.3.0 - avcodec.h
   Add av_packet_shrink_side_data.
 
@@ -42,7 +45,6 @@ API changes, most recent first:
   Add AVStream.attached_pic and AV_DISPOSITION_ATTACHED_PIC,
   used for dealing with attached pictures/cover art.
 
->>>>>>> qatar/master
 2012-02-25 - c9bca80 - lavu 51.24.0 - error.h
   Add AVERROR_UNKNOWN
 

libavcodec/avcodec.h

@@ -1280,15 +1280,12 @@ typedef struct AVCodecContext {
      */
     unsigned int stream_codec_tag;
 
+#if FF_API_SUB_ID
     /**
-     * Some codecs need additional format info. It is stored here.
-     * If any muxer uses this then ALL demuxers/parsers AND encoders for the
-     * specific codec MUST set it correctly otherwise stream copy breaks.
-     * In general use of this field by muxers is not recommended.
-     * - encoding: Set by libavcodec.
-     * - decoding: Set by libavcodec. (FIXME: Is this OK?)
+     * @deprecated this field is unused
      */
-    int sub_id;
+    attribute_deprecated int sub_id;
+#endif
 
     void *priv_data;
 
@@ -4504,4 +4501,14 @@ const AVClass *avcodec_get_frame_class(void);
  */
 int avcodec_is_open(AVCodecContext *s);
 
+/**
+ * @return a non-zero number if codec is an encoder, zero otherwise
+ */
+int av_codec_is_encoder(AVCodec *codec);
+
+/**
+ * @return a non-zero number if codec is a decoder, zero otherwise
+ */
+int av_codec_is_decoder(AVCodec *codec);
+
 #endif /* AVCODEC_AVCODEC_H */

libavcodec/libopencore-amr.c

@@ -85,6 +85,7 @@ typedef struct AMRContext {
     int   enc_bitrate;
     int   enc_mode;
     int   enc_dtx;
+    int   enc_last_frame;
 } AMRContext;
 
 static const AVOption options[] = {
@@ -195,6 +196,7 @@ static av_cold int amr_nb_encode_init(AVCodecContext *avctx)
     }
 
     avctx->frame_size  = 160;
+    avctx->delay       =  50;
     avctx->coded_frame = avcodec_alloc_frame();
     if (!avctx->coded_frame)
         return AVERROR(ENOMEM);
@@ -227,17 +229,40 @@ static int amr_nb_encode_frame(AVCodecContext *avctx,
 {
     AMRContext *s = avctx->priv_data;
     int written;
+    int16_t *flush_buf = NULL;
+    const int16_t *samples = data;
 
     if (s->enc_bitrate != avctx->bit_rate) {
         s->enc_mode    = get_bitrate_mode(avctx->bit_rate, avctx);
         s->enc_bitrate = avctx->bit_rate;
     }
 
-    written = Encoder_Interface_Encode(s->enc_state, s->enc_mode, data,
+    if (data) {
+        if (avctx->frame_size < 160) {
+            flush_buf = av_mallocz(160 * sizeof(*flush_buf));
+            if (!flush_buf)
+                return AVERROR(ENOMEM);
+            memcpy(flush_buf, samples, avctx->frame_size * sizeof(*flush_buf));
+            samples = flush_buf;
+            if (avctx->frame_size < 110)
+                s->enc_last_frame = -1;
+        }
+    } else {
+        if (s->enc_last_frame < 0)
+            return 0;
+        flush_buf = av_mallocz(160 * sizeof(*flush_buf));
+        if (!flush_buf)
+            return AVERROR(ENOMEM);
+        samples = flush_buf;
+        s->enc_last_frame = -1;
+    }
+
+    written = Encoder_Interface_Encode(s->enc_state, s->enc_mode, samples,
                                        frame, 0);
     av_dlog(avctx, "amr_nb_encode_frame encoded %u bytes, bitrate %u, first byte was %#02x\n",
             written, s->enc_mode, frame[0]);
 
+    av_freep(&flush_buf);
     return written;
 }
 
@@ -249,6 +274,7 @@ AVCodec ff_libopencore_amrnb_encoder = {
     .init           = amr_nb_encode_init,
     .encode         = amr_nb_encode_frame,
     .close          = amr_nb_encode_close,
+    .capabilities   = CODEC_CAP_DELAY | CODEC_CAP_SMALL_LAST_FRAME,
     .sample_fmts = (const enum AVSampleFormat[]){AV_SAMPLE_FMT_S16,AV_SAMPLE_FMT_NONE},
     .long_name = NULL_IF_CONFIG_SMALL("OpenCORE Adaptive Multi-Rate (AMR) Narrow-Band"),
     .priv_class = &class,

libavcodec/mpeg12.c

@@ -1276,7 +1276,6 @@ static int mpeg_decode_postinit(AVCodecContext *avctx)
          * that behave like P-frames. */
         avctx->has_b_frames = !s->low_delay;
 
-        assert((avctx->sub_id == 1) == (avctx->codec_id == CODEC_ID_MPEG1VIDEO));
         if (avctx->codec_id == CODEC_ID_MPEG1VIDEO) {
             //MPEG-1 fps
             avctx->time_base.den = avpriv_frame_rate_tab[s->frame_rate_index].num;
@@ -1420,7 +1419,6 @@ static void mpeg_decode_sequence_extension(Mpeg1Context *s1)
 
     av_dlog(s->avctx, "sequence extension\n");
     s->codec_id      = s->avctx->codec_id = CODEC_ID_MPEG2VIDEO;
-    s->avctx->sub_id = 2; /* indicates MPEG-2 found */
 
     if (s->avctx->debug & FF_DEBUG_PICT_INFO)
         av_log(s->avctx, AV_LOG_DEBUG, "profile: %d, level: %d vbv buffer: %d, bitrate:%d\n",
@@ -2038,7 +2036,6 @@ static int mpeg1_decode_sequence(AVCodecContext *avctx,
     s->frame_pred_frame_dct = 1;
     s->chroma_format        = 1;
     s->codec_id             = s->avctx->codec_id = CODEC_ID_MPEG1VIDEO;
-    avctx->sub_id           = 1; /* indicates MPEG-1 */
     s->out_format           = FMT_MPEG1;
     s->swap_uv              = 0; // AFAIK VCR2 does not have SEQ_HEADER
     if (s->flags & CODEC_FLAG_LOW_DELAY)
@@ -2097,12 +2094,10 @@ static int vcr2_init_sequence(AVCodecContext *avctx)
     s->chroma_format         = 1;
     if (s->codec_tag == AV_RL32("BW10")) {
         s->codec_id              = s->avctx->codec_id = CODEC_ID_MPEG1VIDEO;
-        avctx->sub_id            = 1; /* indicates MPEG-1 */
     } else {
         exchange_uv(s); // common init reset pblocks, so we swap them here
         s->swap_uv = 1; // in case of xvmc we need to swap uv for each MB
         s->codec_id              = s->avctx->codec_id = CODEC_ID_MPEG2VIDEO;
-        avctx->sub_id            = 2; /* indicates MPEG-2 */
     }
     s1->save_width           = s->width;
     s1->save_height          = s->height;

libavcodec/mpegaudiodec.c

@@ -1659,7 +1659,6 @@ static int decode_frame(AVCodecContext * avctx, void *data, int *got_frame_ptr,
     avctx->channel_layout = s->nb_channels == 1 ? AV_CH_LAYOUT_MONO : AV_CH_LAYOUT_STEREO;
     if (!avctx->bit_rate)
         avctx->bit_rate = s->bit_rate;
-    avctx->sub_id = s->layer;
 
     if (s->frame_size <= 0 || s->frame_size > buf_size) {
         av_log(avctx, AV_LOG_ERROR, "incomplete frame\n");
@@ -1732,7 +1731,6 @@ static int decode_frame_adu(AVCodecContext *avctx, void *data,
     avctx->channels    = s->nb_channels;
     if (!avctx->bit_rate)
         avctx->bit_rate = s->bit_rate;
-    avctx->sub_id = s->layer;
 
     s->frame_size = len;
 

libavcodec/mpegaudiodecheader.c

@@ -142,6 +142,5 @@ int avpriv_mpa_decode_header(AVCodecContext *avctx, uint32_t head, int *sample_r
     *sample_rate = s->sample_rate;
     *channels = s->nb_channels;
     *bit_rate = s->bit_rate;
-    avctx->sub_id = s->layer;
     return s->frame_size;
 }

libavcodec/mpegvideo_parser.c

@@ -69,7 +69,6 @@ static void mpegvideo_extract_headers(AVCodecParserContext *s,
                 pc->frame_rate.num = avctx->time_base.num = avpriv_frame_rate_tab[frame_rate_index].den;
                 avctx->bit_rate = ((buf[4]<<10) | (buf[5]<<2) | (buf[6]>>6))*400;
                 avctx->codec_id = CODEC_ID_MPEG1VIDEO;
-                avctx->sub_id = 1;
             }
             break;
         case EXT_START_CODE:
@@ -94,7 +93,6 @@ static void mpegvideo_extract_headers(AVCodecParserContext *s,
                         avctx->time_base.den = pc->frame_rate.den * (frame_rate_ext_n + 1) * 2;
                         avctx->time_base.num = pc->frame_rate.num * (frame_rate_ext_d + 1);
                         avctx->codec_id = CODEC_ID_MPEG2VIDEO;
-                        avctx->sub_id = 2; /* forces MPEG2 */
                     }
                     break;
                 case 0x8: /* picture coding extension */

libavcodec/nellymoserenc.c

@@ -52,13 +52,11 @@
 typedef struct NellyMoserEncodeContext {
     AVCodecContext  *avctx;
     int             last_frame;
-    int             bufsel;
-    int             have_saved;
     DSPContext      dsp;
     FFTContext      mdct_ctx;
     DECLARE_ALIGNED(32, float, mdct_out)[NELLY_SAMPLES];
     DECLARE_ALIGNED(32, float, in_buff)[NELLY_SAMPLES];
-    DECLARE_ALIGNED(32, float, buf)[2][3 * NELLY_BUF_LEN];     ///< sample buffer
+    DECLARE_ALIGNED(32, float, buf)[3 * NELLY_BUF_LEN];     ///< sample buffer
     float           (*opt )[NELLY_BANDS];
     uint8_t         (*path)[NELLY_BANDS];
 } NellyMoserEncodeContext;
@@ -115,16 +113,17 @@ static const uint8_t quant_lut_offset[8] = { 0, 0, 1, 4, 11, 32, 81, 230 };
 
 static void apply_mdct(NellyMoserEncodeContext *s)
 {
-    s->dsp.vector_fmul(s->in_buff, s->buf[s->bufsel], ff_sine_128, NELLY_BUF_LEN);
-    s->dsp.vector_fmul_reverse(s->in_buff + NELLY_BUF_LEN, s->buf[s->bufsel] + NELLY_BUF_LEN, ff_sine_128,
-                               NELLY_BUF_LEN);
+    float *in0 = s->buf;
+    float *in1 = s->buf + NELLY_BUF_LEN;
+    float *in2 = s->buf + 2 * NELLY_BUF_LEN;
+
+    s->dsp.vector_fmul        (s->in_buff,                 in0, ff_sine_128, NELLY_BUF_LEN);
+    s->dsp.vector_fmul_reverse(s->in_buff + NELLY_BUF_LEN, in1, ff_sine_128, NELLY_BUF_LEN);
     s->mdct_ctx.mdct_calc(&s->mdct_ctx, s->mdct_out, s->in_buff);
 
-    s->dsp.vector_fmul(s->buf[s->bufsel] + NELLY_BUF_LEN, s->buf[s->bufsel] + NELLY_BUF_LEN,
-                       ff_sine_128, NELLY_BUF_LEN);
-    s->dsp.vector_fmul_reverse(s->buf[s->bufsel] + 2 * NELLY_BUF_LEN, s->buf[1 - s->bufsel], ff_sine_128,
-                               NELLY_BUF_LEN);
-    s->mdct_ctx.mdct_calc(&s->mdct_ctx, s->mdct_out + NELLY_BUF_LEN, s->buf[s->bufsel] + NELLY_BUF_LEN);
+    s->dsp.vector_fmul        (s->in_buff,                 in1, ff_sine_128, NELLY_BUF_LEN);
+    s->dsp.vector_fmul_reverse(s->in_buff + NELLY_BUF_LEN, in2, ff_sine_128, NELLY_BUF_LEN);
+    s->mdct_ctx.mdct_calc(&s->mdct_ctx, s->mdct_out + NELLY_BUF_LEN, s->in_buff);
 }
 
 static av_cold int encode_end(AVCodecContext *avctx)
@@ -161,6 +160,7 @@ static av_cold int encode_init(AVCodecContext *avctx)
     }
 
     avctx->frame_size = NELLY_SAMPLES;
+    avctx->delay      = NELLY_BUF_LEN;
     s->avctx = avctx;
     if ((ret = ff_mdct_init(&s->mdct_ctx, 8, 0, 32768.0)) < 0)
         goto error;
@@ -363,38 +363,33 @@ static void encode_block(NellyMoserEncodeContext *s, unsigned char *output, int
     }
 
     flush_put_bits(&pb);
+    memset(put_bits_ptr(&pb), 0, output + output_size - put_bits_ptr(&pb));
 }
 
 static int encode_frame(AVCodecContext *avctx, uint8_t *frame, int buf_size, void *data)
 {
     NellyMoserEncodeContext *s = avctx->priv_data;
     const float *samples = data;
-    int i;
 
     if (s->last_frame)
         return 0;
 
+    memcpy(s->buf, s->buf + NELLY_SAMPLES, NELLY_BUF_LEN * sizeof(*s->buf));
     if (data) {
-        memcpy(s->buf[s->bufsel], samples, avctx->frame_size * sizeof(*samples));
-        for (i = avctx->frame_size; i < NELLY_SAMPLES; i++) {
-            s->buf[s->bufsel][i] = 0;
-        }
-        s->bufsel = 1 - s->bufsel;
-        if (!s->have_saved) {
-            s->have_saved = 1;
-            return 0;
+        memcpy(s->buf + NELLY_BUF_LEN, samples, avctx->frame_size * sizeof(*s->buf));
+        if (avctx->frame_size < NELLY_SAMPLES) {
+            memset(s->buf + NELLY_BUF_LEN + avctx->frame_size, 0,
+                   (NELLY_SAMPLES - avctx->frame_size) * sizeof(*s->buf));
+            if (avctx->frame_size >= NELLY_BUF_LEN)
+                s->last_frame = 1;
         }
     } else {
-        memset(s->buf[s->bufsel], 0, sizeof(s->buf[0][0]) * NELLY_BUF_LEN);
-        s->bufsel = 1 - s->bufsel;
+        memset(s->buf + NELLY_BUF_LEN, 0, NELLY_SAMPLES * sizeof(*s->buf));
         s->last_frame = 1;
     }
 
-    if (s->have_saved) {
-        encode_block(s, frame, buf_size);
-        return NELLY_BLOCK_LEN;
-    }
-    return 0;
+    encode_block(s, frame, buf_size);
+    return NELLY_BLOCK_LEN;
 }
 
 AVCodec ff_nellymoser_encoder = {

libavcodec/options.c

@@ -111,7 +111,9 @@ static const AVOption options[]={
 {"noout", "skip bitstream encoding", 0, AV_OPT_TYPE_CONST, {.dbl = CODEC_FLAG2_NO_OUTPUT }, INT_MIN, INT_MAX, V|E, "flags2"},
 {"local_header", "place global headers at every keyframe instead of in extradata", 0, AV_OPT_TYPE_CONST, {.dbl = CODEC_FLAG2_LOCAL_HEADER }, INT_MIN, INT_MAX, V|E, "flags2"},
 {"showall", "Show all frames before the first keyframe", 0, AV_OPT_TYPE_CONST, {.dbl = CODEC_FLAG2_SHOW_ALL }, INT_MIN, INT_MAX, V|D, "flags2"},
+#if FF_API_SUB_ID
 {"sub_id", NULL, OFFSET(sub_id), AV_OPT_TYPE_INT, {.dbl = DEFAULT }, INT_MIN, INT_MAX},
+#endif
 {"me_method", "set motion estimation method", OFFSET(me_method), AV_OPT_TYPE_INT, {.dbl = ME_EPZS }, INT_MIN, INT_MAX, V|E, "me_method"},
 {"zero", "zero motion estimation (fastest)", 0, AV_OPT_TYPE_CONST, {.dbl = ME_ZERO }, INT_MIN, INT_MAX, V|E, "me_method" },
 {"full", "full motion estimation (slowest)", 0, AV_OPT_TYPE_CONST, {.dbl = ME_FULL }, INT_MIN, INT_MAX, V|E, "me_method" },

libavcodec/proresenc_kostya.c

@@ -42,6 +42,67 @@ enum {
     PRORES_PROFILE_HQ,
 };
 
+enum {
+    QUANT_MAT_PROXY = 0,
+    QUANT_MAT_LT,
+    QUANT_MAT_STANDARD,
+    QUANT_MAT_HQ,
+    QUANT_MAT_DEFAULT,
+};
+
+static const uint8_t prores_quant_matrices[][64] = {
+    { // proxy
+         4,  7,  9, 11, 13, 14, 15, 63,
+         7,  7, 11, 12, 14, 15, 63, 63,
+         9, 11, 13, 14, 15, 63, 63, 63,
+        11, 11, 13, 14, 63, 63, 63, 63,
+        11, 13, 14, 63, 63, 63, 63, 63,
+        13, 14, 63, 63, 63, 63, 63, 63,
+        13, 63, 63, 63, 63, 63, 63, 63,
+        63, 63, 63, 63, 63, 63, 63, 63,
+    },
+    { // LT
+         4,  5,  6,  7,  9, 11, 13, 15,
+         5,  5,  7,  8, 11, 13, 15, 17,
+         6,  7,  9, 11, 13, 15, 15, 17,
+         7,  7,  9, 11, 13, 15, 17, 19,
+         7,  9, 11, 13, 14, 16, 19, 23,
+         9, 11, 13, 14, 16, 19, 23, 29,
+         9, 11, 13, 15, 17, 21, 28, 35,
+        11, 13, 16, 17, 21, 28, 35, 41,
+    },
+    { // standard
+         4,  4,  5,  5,  6,  7,  7,  9,
+         4,  4,  5,  6,  7,  7,  9,  9,
+         5,  5,  6,  7,  7,  9,  9, 10,
+         5,  5,  6,  7,  7,  9,  9, 10,
+         5,  6,  7,  7,  8,  9, 10, 12,
+         6,  7,  7,  8,  9, 10, 12, 15,
+         6,  7,  7,  9, 10, 11, 14, 17,
+         7,  7,  9, 10, 11, 14, 17, 21,
+    },
+    { // high quality
+         4,  4,  4,  4,  4,  4,  4,  4,
+         4,  4,  4,  4,  4,  4,  4,  4,
+         4,  4,  4,  4,  4,  4,  4,  4,
+         4,  4,  4,  4,  4,  4,  4,  5,
+         4,  4,  4,  4,  4,  4,  5,  5,
+         4,  4,  4,  4,  4,  5,  5,  6,
+         4,  4,  4,  4,  5,  5,  6,  7,
+         4,  4,  4,  4,  5,  6,  7,  7,
+    },
+    { // codec default
+         4,  4,  4,  4,  4,  4,  4,  4,
+         4,  4,  4,  4,  4,  4,  4,  4,
+         4,  4,  4,  4,  4,  4,  4,  4,
+         4,  4,  4,  4,  4,  4,  4,  4,
+         4,  4,  4,  4,  4,  4,  4,  4,
+         4,  4,  4,  4,  4,  4,  4,  4,
+         4,  4,  4,  4,  4,  4,  4,  4,
+         4,  4,  4,  4,  4,  4,  4,  4,
+    },
+};
+
 #define NUM_MB_LIMITS 4
 static const int prores_mb_limits[NUM_MB_LIMITS] = {
     1620, // up to 720x576
@@ -56,7 +117,7 @@ static const struct prores_profile {
     int         min_quant;
     int         max_quant;
     int         br_tab[NUM_MB_LIMITS];
-    uint8_t     quant[64];
+    int         quant;
 } prores_profile_info[4] = {
     {
         .full_name = "proxy",
@@ -64,16 +125,7 @@ static const struct prores_profile {
         .min_quant = 4,
         .max_quant = 8,
         .br_tab    = { 300, 242, 220, 194 },
-        .quant     = {
-             4,  7,  9, 11, 13, 14, 15, 63,
-             7,  7, 11, 12, 14, 15, 63, 63,
-             9, 11, 13, 14, 15, 63, 63, 63,
-            11, 11, 13, 14, 63, 63, 63, 63,
-            11, 13, 14, 63, 63, 63, 63, 63,
-            13, 14, 63, 63, 63, 63, 63, 63,
-            13, 63, 63, 63, 63, 63, 63, 63,
-            63, 63, 63, 63, 63, 63, 63, 63,
-        },
+        .quant     = QUANT_MAT_PROXY,
     },
     {
         .full_name = "LT",
@@ -81,16 +133,7 @@ static const struct prores_profile {
         .min_quant = 1,
         .max_quant = 9,
         .br_tab    = { 720, 560, 490, 440 },
-        .quant     = {
-             4,  5,  6,  7,  9, 11, 13, 15,
-             5,  5,  7,  8, 11, 13, 15, 17,
-             6,  7,  9, 11, 13, 15, 15, 17,
-             7,  7,  9, 11, 13, 15, 17, 19,
-             7,  9, 11, 13, 14, 16, 19, 23,
-             9, 11, 13, 14, 16, 19, 23, 29,
-             9, 11, 13, 15, 17, 21, 28, 35,
-            11, 13, 16, 17, 21, 28, 35, 41,
-        },
+        .quant     = QUANT_MAT_LT,
     },
     {
         .full_name = "standard",
@@ -98,16 +141,7 @@ static const struct prores_profile {
         .min_quant = 1,
         .max_quant = 6,
         .br_tab    = { 1050, 808, 710, 632 },
-        .quant     = {
-             4,  4,  5,  5,  6,  7,  7,  9,
-             4,  4,  5,  6,  7,  7,  9,  9,
-             5,  5,  6,  7,  7,  9,  9, 10,
-             5,  5,  6,  7,  7,  9,  9, 10,
-             5,  6,  7,  7,  8,  9, 10, 12,
-             6,  7,  7,  8,  9, 10, 12, 15,
-             6,  7,  7,  9, 10, 11, 14, 17,
-             7,  7,  9, 10, 11, 14, 17, 21,
-        },
+        .quant     = QUANT_MAT_STANDARD,
     },
     {
         .full_name = "high quality",
@@ -115,16 +149,7 @@ static const struct prores_profile {
         .min_quant = 1,
         .max_quant = 6,
         .br_tab    = { 1566, 1216, 1070, 950 },
-        .quant     = {
-             4,  4,  4,  4,  4,  4,  4,  4,
-             4,  4,  4,  4,  4,  4,  4,  4,
-             4,  4,  4,  4,  4,  4,  4,  4,
-             4,  4,  4,  4,  4,  4,  4,  5,
-             4,  4,  4,  4,  4,  4,  5,  5,
-             4,  4,  4,  4,  4,  5,  5,  6,
-             4,  4,  4,  4,  5,  5,  6,  7,
-             4,  4,  4,  4,  5,  6,  7,  7,
-        },
+        .quant     = QUANT_MAT_HQ,
     }
 // for 4444 profile bitrate numbers are { 2350, 1828, 1600, 1425 }
 };
@@ -147,6 +172,7 @@ typedef struct ProresContext {
     DECLARE_ALIGNED(16, uint16_t, emu_buf)[16*16];
     int16_t quants[MAX_STORED_Q][64];
     int16_t custom_q[64];
+    const uint8_t *quant_mat;
 
     ProresDSPContext dsp;
     ScanTable  scantable;
@@ -159,6 +185,9 @@ typedef struct ProresContext {
     int num_planes;
     int bits_per_mb;
 
+    char *vendor;
+    int quant_sel;
+
     int frame_size;
 
     int profile;
@@ -373,7 +402,7 @@ static int encode_slice(AVCodecContext *avctx, const AVFrame *pic,
     } else {
         qmat = ctx->custom_q;
         for (i = 0; i < 64; i++)
-            qmat[i] = ctx->profile_info->quant[i] * quant;
+            qmat[i] = ctx->quant_mat[i] * quant;
     }
 
     for (i = 0; i < ctx->num_planes; i++) {
@@ -591,7 +620,7 @@ static int find_slice_quant(AVCodecContext *avctx, const AVFrame *pic,
             } else {
                 qmat = ctx->custom_q;
                 for (i = 0; i < 64; i++)
-                    qmat[i] = ctx->profile_info->quant[i] * q;
+                    qmat[i] = ctx->quant_mat[i] * q;
             }
             for (i = 0; i < ctx->num_planes; i++) {
                 bits += estimate_slice_plane(ctx, &error, i,
@@ -684,7 +713,7 @@ static int encode_frame(AVCodecContext *avctx, AVPacket *pkt,
     tmp = buf;
     buf += 2;                                   // frame header size will be stored here
     bytestream_put_be16  (&buf, 0);             // version 1
-    bytestream_put_buffer(&buf, "Lavc", 4);     // creator
+    bytestream_put_buffer(&buf, ctx->vendor, 4);
     bytestream_put_be16  (&buf, avctx->width);
     bytestream_put_be16  (&buf, avctx->height);
     bytestream_put_byte  (&buf, ctx->chroma_factor << 6); // frame flags
@@ -694,13 +723,17 @@ static int encode_frame(AVCodecContext *avctx, AVPacket *pkt,
     bytestream_put_byte  (&buf, avctx->colorspace);
     bytestream_put_byte  (&buf, 0x40);          // source format and alpha information
     bytestream_put_byte  (&buf, 0);             // reserved
-    bytestream_put_byte  (&buf, 0x03);          // matrix flags - both matrices are present
-    // luma quantisation matrix
-    for (i = 0; i < 64; i++)
-        bytestream_put_byte(&buf, ctx->profile_info->quant[i]);
-    // chroma quantisation matrix
-    for (i = 0; i < 64; i++)
-        bytestream_put_byte(&buf, ctx->profile_info->quant[i]);
+    if (ctx->quant_sel != QUANT_MAT_DEFAULT) {
+        bytestream_put_byte  (&buf, 0x03);      // matrix flags - both matrices are present
+        // luma quantisation matrix
+        for (i = 0; i < 64; i++)
+            bytestream_put_byte(&buf, ctx->quant_mat[i]);
+        // chroma quantisation matrix
+        for (i = 0; i < 64; i++)
+            bytestream_put_byte(&buf, ctx->quant_mat[i]);
+    } else {
+        bytestream_put_byte  (&buf, 0x00);      // matrix flags - default matrices are used
+    }
     bytestream_put_be16  (&tmp, buf - orig_buf); // write back frame header size
 
     // picture header
@@ -816,10 +849,25 @@ static av_cold int encode_init(AVCodecContext *avctx)
     ctx->slices_width += av_popcount(ctx->mb_width - ctx->slices_width * mps);
     ctx->num_slices    = ctx->mb_height * ctx->slices_width;
 
-    for (i = 0; i < NUM_MB_LIMITS - 1; i++)
-        if (prores_mb_limits[i] >= ctx->mb_width * ctx->mb_height)
-            break;
-    ctx->bits_per_mb   = ctx->profile_info->br_tab[i];
+    if (ctx->quant_sel == -1)
+        ctx->quant_mat = prores_quant_matrices[ctx->profile_info->quant];
+    else
+        ctx->quant_mat = prores_quant_matrices[ctx->quant_sel];
+
+    if (strlen(ctx->vendor) != 4) {
+        av_log(avctx, AV_LOG_ERROR, "vendor ID should be 4 bytes\n");
+        return AVERROR_INVALIDDATA;
+    }
+
+    if (!ctx->bits_per_mb) {
+        for (i = 0; i < NUM_MB_LIMITS - 1; i++)
+            if (prores_mb_limits[i] >= ctx->mb_width * ctx->mb_height)
+                break;
+        ctx->bits_per_mb   = ctx->profile_info->br_tab[i];
+    } else if (ctx->bits_per_mb < 128) {
+        av_log(avctx, AV_LOG_ERROR, "too few bits per MB, please set at least 128\n");
+        return AVERROR_INVALIDDATA;
+    }
 
     ctx->frame_size = ctx->num_slices * (2 + 2 * ctx->num_planes
                                          + (2 * mps * ctx->bits_per_mb) / 8)
@@ -829,7 +877,7 @@ static av_cold int encode_init(AVCodecContext *avctx)
     max_quant = ctx->profile_info->max_quant;
     for (i = min_quant; i < MAX_STORED_Q; i++) {
         for (j = 0; j < 64; j++)
-            ctx->quants[i][j] = ctx->profile_info->quant[j] * i;
+            ctx->quants[i][j] = ctx->quant_mat[j] * i;
     }
 
     avctx->codec_tag   = ctx->profile_info->tag;
@@ -877,6 +925,24 @@ static const AVOption options[] = {
         0, 0, VE, "profile" },
     { "hq",            NULL, 0, AV_OPT_TYPE_CONST, { PRORES_PROFILE_HQ },
         0, 0, VE, "profile" },
+    { "vendor", "vendor ID", OFFSET(vendor),
+        AV_OPT_TYPE_STRING, { .str = "Lavc" }, CHAR_MIN, CHAR_MAX, VE },
+    { "bits_per_mb", "desired bits per macroblock", OFFSET(bits_per_mb),
+        AV_OPT_TYPE_INT, { 0 }, 0, 8192, VE },
+    { "quant_mat", "quantiser matrix", OFFSET(quant_sel), AV_OPT_TYPE_INT,
+        { -1 }, -1, QUANT_MAT_DEFAULT, VE, "quant_mat" },
+    { "auto",          NULL, 0, AV_OPT_TYPE_CONST, { -1 },
+        0, 0, VE, "quant_mat" },
+    { "proxy",         NULL, 0, AV_OPT_TYPE_CONST, { QUANT_MAT_PROXY },
+        0, 0, VE, "quant_mat" },
+    { "lt",            NULL, 0, AV_OPT_TYPE_CONST, { QUANT_MAT_LT },
+        0, 0, VE, "quant_mat" },
+    { "standard",      NULL, 0, AV_OPT_TYPE_CONST, { QUANT_MAT_STANDARD },
+        0, 0, VE, "quant_mat" },
+    { "hq",            NULL, 0, AV_OPT_TYPE_CONST, { QUANT_MAT_HQ },
+        0, 0, VE, "quant_mat" },
+    { "default",       NULL, 0, AV_OPT_TYPE_CONST, { QUANT_MAT_DEFAULT },
+        0, 0, VE, "quant_mat" },
     { NULL }
 };
 

libavcodec/pthread.c

@@ -416,7 +416,6 @@ static int update_context_from_thread(AVCodecContext *dst, AVCodecContext *src,
     int err = 0;
 
     if (dst != src) {
-        dst->sub_id    = src->sub_id;
         dst->time_base = src->time_base;
         dst->width     = src->width;
         dst->height    = src->height;

libavcodec/qpeg.c

@@ -25,15 +25,17 @@
  */
 
 #include "avcodec.h"
+#include "bytestream.h"
 
 typedef struct QpegContext{
     AVCodecContext *avctx;
     AVFrame pic, ref;
     uint32_t pal[256];
+    GetByteContext buffer;
 } QpegContext;
 
-static int qpeg_decode_intra(const uint8_t *src, uint8_t *dst, int size,
-                            int stride, int width, int height)
+static void qpeg_decode_intra(QpegContext *qctx, uint8_t *dst,
+                              int stride, int width, int height)
 {
     int i;
     int code;
@@ -46,31 +48,26 @@ static int qpeg_decode_intra(const uint8_t *src, uint8_t *dst, int size,
     height--;
     dst = dst + height * stride;
 
-    while((size > 0) && (rows_to_go > 0)) {
-        code = *src++;
-        size--;
+    while ((bytestream2_get_bytes_left(&qctx->buffer) > 0) && (rows_to_go > 0)) {
+        code = bytestream2_get_byte(&qctx->buffer);
         run = copy = 0;
         if(code == 0xFC) /* end-of-picture code */
             break;
         if(code >= 0xF8) { /* very long run */
-            c0 = *src++;
-            c1 = *src++;
-            size -= 2;
+            c0 = bytestream2_get_byte(&qctx->buffer);
+            c1 = bytestream2_get_byte(&qctx->buffer);
             run = ((code & 0x7) << 16) + (c0 << 8) + c1 + 2;
         } else if (code >= 0xF0) { /* long run */
-            c0 = *src++;
-            size--;
+            c0 = bytestream2_get_byte(&qctx->buffer);
             run = ((code & 0xF) << 8) + c0 + 2;
         } else if (code >= 0xE0) { /* short run */
             run = (code & 0x1F) + 2;
         } else if (code >= 0xC0) { /* very long copy */
-            c0 = *src++;
-            c1 = *src++;
-            size -= 2;
+            c0 = bytestream2_get_byte(&qctx->buffer);
+            c1 = bytestream2_get_byte(&qctx->buffer);
             copy = ((code & 0x3F) << 16) + (c0 << 8) + c1 + 1;
         } else if (code >= 0x80) { /* long copy */
-            c0 = *src++;
-            size--;
+            c0 = bytestream2_get_byte(&qctx->buffer);
             copy = ((code & 0x7F) << 8) + c0 + 1;
         } else { /* short copy */
             copy = code + 1;
@@ -80,8 +77,7 @@ static int qpeg_decode_intra(const uint8_t *src, uint8_t *dst, int size,
         if(run) {
             int p;
 
-            p = *src++;
-            size--;
+            p = bytestream2_get_byte(&qctx->buffer);
             for(i = 0; i < run; i++) {
                 dst[filled++] = p;
                 if (filled >= width) {
@@ -93,11 +89,8 @@ static int qpeg_decode_intra(const uint8_t *src, uint8_t *dst, int size,
                 }
             }
         } else {
-            size -= copy;
-            if (size<0)
-                return AVERROR_INVALIDDATA;
             for(i = 0; i < copy; i++) {
-                dst[filled++] = *src++;
+                dst[filled++] = bytestream2_get_byte(&qctx->buffer);
                 if (filled >= width) {
                     filled = 0;
                     dst -= stride;
@@ -108,7 +101,6 @@ static int qpeg_decode_intra(const uint8_t *src, uint8_t *dst, int size,
             }
         }
     }
-    return 0;
 }
 
 static const int qpeg_table_h[16] =
@@ -117,9 +109,10 @@ static const int qpeg_table_w[16] =
  { 0x00, 0x20, 0x18, 0x08, 0x18, 0x10, 0x20, 0x10, 0x08, 0x10, 0x20, 0x20, 0x08, 0x10, 0x18, 0x04};
 
 /* Decodes delta frames */
-static void qpeg_decode_inter(const uint8_t *src, uint8_t *dst, int size,
-                            int stride, int width, int height,
-                            int delta, const uint8_t *ctable, uint8_t *refdata)
+static void qpeg_decode_inter(QpegContext *qctx, uint8_t *dst,
+                              int stride, int width, int height,
+                              int delta, const uint8_t *ctable,
+                              uint8_t *refdata)
 {
     int i, j;
     int code;
@@ -137,13 +130,12 @@ static void qpeg_decode_inter(const uint8_t *src, uint8_t *dst, int size,
     height--;
     dst = dst + height * stride;
 
-    while((size > 0) && (height >= 0)) {
-        code = *src++;
-        size--;
+    while ((bytestream2_get_bytes_left(&qctx->buffer) > 0) && (height >= 0)) {
+        code = bytestream2_get_byte(&qctx->buffer);
 
         if(delta) {
             /* motion compensation */
-            while(size > 0 && (code & 0xF0) == 0xF0) {
+            while(bytestream2_get_bytes_left(&qctx->buffer) > 0 && (code & 0xF0) == 0xF0) {
                 if(delta == 1) {
                     int me_idx;
                     int me_w, me_h, me_x, me_y;
@@ -156,8 +148,7 @@ static void qpeg_decode_inter(const uint8_t *src, uint8_t *dst, int size,
                     me_h = qpeg_table_h[me_idx];
 
                     /* extract motion vector */
-                    corr = *src++;
-                    size--;
+                    corr = bytestream2_get_byte(&qctx->buffer);
 
                     val = corr >> 4;
                     if(val > 7)
@@ -184,8 +175,7 @@ static void qpeg_decode_inter(const uint8_t *src, uint8_t *dst, int size,
                         }
                     }
                 }
-                code = *src++;
-                size--;
+                code = bytestream2_get_byte(&qctx->buffer);
             }
         }
 
@@ -195,8 +185,7 @@ static void qpeg_decode_inter(const uint8_t *src, uint8_t *dst, int size,
             int p;
 
             code &= 0x1F;
-            p = *src++;
-            size--;
+            p = bytestream2_get_byte(&qctx->buffer);
             for(i = 0; i <= code; i++) {
                 dst[filled++] = p;
                 if(filled >= width) {
@@ -210,11 +199,11 @@ static void qpeg_decode_inter(const uint8_t *src, uint8_t *dst, int size,
         } else if(code >= 0xC0) { /* copy code: 0xC0..0xDF */
             code &= 0x1F;
 
-            if(code + 1 > size)
+            if(code + 1 > bytestream2_get_bytes_left(&qctx->buffer))
                 break;
 
             for(i = 0; i <= code; i++) {
-                dst[filled++] = *src++;
+                dst[filled++] = bytestream2_get_byte(&qctx->buffer);
                 if(filled >= width) {
                     filled = 0;
                     dst -= stride;
@@ -223,18 +212,17 @@ static void qpeg_decode_inter(const uint8_t *src, uint8_t *dst, int size,
                         break;
                 }
             }
-            size -= code + 1;
         } else if(code >= 0x80) { /* skip code: 0x80..0xBF */
             int skip;
 
             code &= 0x3F;
             /* codes 0x80 and 0x81 are actually escape codes,
                skip value minus constant is in the next byte */
-            if(!code) {
-                skip = (*src++) + 64; size--;
-            } else if(code == 1) {
-                skip = (*src++) + 320; size--;
-            } else
+            if(!code)
+                skip = bytestream2_get_byte(&qctx->buffer) +  64;
+            else if(code == 1)
+                skip = bytestream2_get_byte(&qctx->buffer) + 320;
+            else
                 skip = code;
             filled += skip;
             while( filled >= width) {
@@ -246,8 +234,9 @@ static void qpeg_decode_inter(const uint8_t *src, uint8_t *dst, int size,
             }
         } else {
             /* zero code treated as one-pixel skip */
-            if(code)
+            if(code) {
                 dst[filled++] = ctable[code & 0x7F];
+            }
             else
                 filled++;
             if(filled >= width) {
@@ -263,8 +252,7 @@ static int decode_frame(AVCodecContext *avctx,
                         void *data, int *data_size,
                         AVPacket *avpkt)
 {
-    const uint8_t *buf = avpkt->data;
-    int buf_size = avpkt->size;
+    uint8_t ctable[128];
     QpegContext * const a = avctx->priv_data;
     AVFrame *  p = &a->pic;
     AVFrame * ref= &a->ref;
@@ -272,6 +260,13 @@ static int decode_frame(AVCodecContext *avctx,
     int delta, ret = 0;
     const uint8_t *pal = av_packet_get_side_data(avpkt, AV_PKT_DATA_PALETTE, NULL);
 
+    if (avpkt->size < 0x86) {
+        av_log(avctx, AV_LOG_ERROR, "Packet is too small\n");
+        return AVERROR_INVALIDDATA;
+    }
+
+    bytestream2_init(&a->buffer, avpkt->data, avpkt->size);
+
     if(ref->data[0])
         avctx->release_buffer(avctx, ref);
     FFSWAP(AVFrame, *ref, *p);
@@ -282,16 +277,17 @@ static int decode_frame(AVCodecContext *avctx,
         return -1;
     }
     outdata = a->pic.data[0];
-    if(buf[0x85] == 0x10) {
-        ret = qpeg_decode_intra(buf+0x86, outdata, buf_size - 0x86, a->pic.linesize[0], avctx->width, avctx->height);
+    bytestream2_skip(&a->buffer, 4);
+    bytestream2_get_buffer(&a->buffer, ctable, 128);
+    bytestream2_skip(&a->buffer, 1);
+
+    delta = bytestream2_get_byte(&a->buffer);
+    if(delta == 0x10) {
+        qpeg_decode_intra(a, outdata, a->pic.linesize[0], avctx->width, avctx->height);
     } else {
-        delta = buf[0x85];
-        qpeg_decode_inter(buf+0x86, outdata, buf_size - 0x86, a->pic.linesize[0], avctx->width, avctx->height, delta, buf + 4, a->ref.data[0]);
+        qpeg_decode_inter(a, outdata, a->pic.linesize[0], avctx->width, avctx->height, delta, ctable, a->ref.data[0]);
     }
 
-    if (ret<0)
-        return ret;
-
     /* make the palette available on the way out */
     if (pal) {
         a->pic.palette_has_changed = 1;
@@ -302,7 +298,7 @@ static int decode_frame(AVCodecContext *avctx,
     *data_size = sizeof(AVFrame);
     *(AVFrame*)data = a->pic;
 
-    return buf_size;
+    return avpkt->size;
 }
 
 static av_cold int decode_init(AVCodecContext *avctx){

libavcodec/ra144.h

@@ -36,6 +36,7 @@ typedef struct {
     AVCodecContext *avctx;
     AVFrame frame;
     LPCContext lpc_ctx;
+    int last_frame;
 
     unsigned int     old_energy;        ///< previous frame energy
 

libavcodec/ra144enc.c

@@ -53,6 +53,7 @@ static av_cold int ra144_encode_init(AVCodecContext * avctx)
         return -1;
     }
     avctx->frame_size = NBLOCKS * BLOCKSIZE;
+    avctx->delay      = avctx->frame_size;
     avctx->bit_rate = 8000;
     ractx = avctx->priv_data;
     ractx->lpc_coef[0] = ractx->lpc_tables[0];
@@ -433,7 +434,7 @@ static int ra144_encode_frame(AVCodecContext *avctx, uint8_t *frame,
 {
     static const uint8_t sizes[LPC_ORDER] = {64, 32, 32, 16, 16, 8, 8, 8, 8, 4};
     static const uint8_t bit_sizes[LPC_ORDER] = {6, 5, 5, 4, 4, 3, 3, 3, 3, 2};
-    RA144Context *ractx;
+    RA144Context *ractx = avctx->priv_data;
     PutBitContext pb;
     int32_t lpc_data[NBLOCKS * BLOCKSIZE];
     int32_t lpc_coefs[LPC_ORDER][MAX_LPC_ORDER];
@@ -445,11 +446,13 @@ static int ra144_encode_frame(AVCodecContext *avctx, uint8_t *frame,
     int energy = 0;
     int i, idx;
 
+    if (ractx->last_frame)
+        return 0;
+
     if (buf_size < FRAMESIZE) {
         av_log(avctx, AV_LOG_ERROR, "output buffer too small\n");
         return 0;
     }
-    ractx = avctx->priv_data;
 
     /**
      * Since the LPC coefficients are calculated on a frame centered over the
@@ -462,11 +465,15 @@ static int ra144_encode_frame(AVCodecContext *avctx, uint8_t *frame,
         lpc_data[i] = ractx->curr_block[BLOCKSIZE + BLOCKSIZE / 2 + i];
         energy += (lpc_data[i] * lpc_data[i]) >> 4;
     }
-    for (i = 2 * BLOCKSIZE + BLOCKSIZE / 2; i < NBLOCKS * BLOCKSIZE; i++) {
-        lpc_data[i] = *((int16_t *)data + i - 2 * BLOCKSIZE - BLOCKSIZE / 2) >>
-                      2;
-        energy += (lpc_data[i] * lpc_data[i]) >> 4;
+    if (data) {
+        int j;
+        for (j = 0; j < avctx->frame_size && i < NBLOCKS * BLOCKSIZE; i++, j++) {
+            lpc_data[i] = samples[j] >> 2;
+            energy += (lpc_data[i] * lpc_data[i]) >> 4;
+        }
     }
+    if (i < NBLOCKS * BLOCKSIZE)
+        memset(&lpc_data[i], 0, (NBLOCKS * BLOCKSIZE - i) * sizeof(*lpc_data));
     energy = ff_energy_tab[quantize(ff_t_sqrt(energy >> 5) >> 10, ff_energy_tab,
                                     32)];
 
@@ -515,8 +522,17 @@ static int ra144_encode_frame(AVCodecContext *avctx, uint8_t *frame,
     ractx->old_energy = energy;
     ractx->lpc_refl_rms[1] = ractx->lpc_refl_rms[0];
     FFSWAP(unsigned int *, ractx->lpc_coef[0], ractx->lpc_coef[1]);
-    for (i = 0; i < NBLOCKS * BLOCKSIZE; i++)
-        ractx->curr_block[i] = samples[i] >> 2;
+
+    /* copy input samples to current block for processing in next call */
+    i = 0;
+    if (data) {
+        for (; i < avctx->frame_size; i++)
+            ractx->curr_block[i] = samples[i] >> 2;
+    } else
+        ractx->last_frame = 1;
+    memset(&ractx->curr_block[i], 0,
+           (NBLOCKS * BLOCKSIZE - i) * sizeof(*ractx->curr_block));
+
     return FRAMESIZE;
 }
 
@@ -529,6 +545,7 @@ AVCodec ff_ra_144_encoder = {
     .init           = ra144_encode_init,
     .encode         = ra144_encode_frame,
     .close          = ra144_encode_close,
+    .capabilities   = CODEC_CAP_DELAY | CODEC_CAP_SMALL_LAST_FRAME,
     .sample_fmts    = (const enum AVSampleFormat[]){ AV_SAMPLE_FMT_S16,
                                                      AV_SAMPLE_FMT_NONE },
     .long_name      = NULL_IF_CONFIG_SMALL("RealAudio 1.0 (14.4K)"),

libavcodec/rv10.c

@@ -40,6 +40,11 @@
 
 #define DC_VLC_BITS 14 //FIXME find a better solution
 
+typedef struct RVDecContext {
+    MpegEncContext m;
+    int sub_id;
+} RVDecContext;
+
 static const uint16_t rv_lum_code[256] =
 {
  0x3e7f, 0x0f00, 0x0f01, 0x0f02, 0x0f03, 0x0f04, 0x0f05, 0x0f06,
@@ -293,8 +298,9 @@ static int rv10_decode_picture_header(MpegEncContext *s)
     return mb_count;
 }
 
-static int rv20_decode_picture_header(MpegEncContext *s)
+static int rv20_decode_picture_header(RVDecContext *rv)
 {
+    MpegEncContext *s = &rv->m;
     int seq, mb_pos, i;
     int rpr_bits;
 
@@ -342,10 +348,10 @@ static int rv20_decode_picture_header(MpegEncContext *s)
         return -1;
     }
 
-    if(RV_GET_MINOR_VER(s->avctx->sub_id) >= 2)
+    if(RV_GET_MINOR_VER(rv->sub_id) >= 2)
         s->loop_filter = get_bits1(&s->gb);
 
-    if(RV_GET_MINOR_VER(s->avctx->sub_id) <= 1)
+    if(RV_GET_MINOR_VER(rv->sub_id) <= 1)
         seq = get_bits(&s->gb, 8) << 7;
     else
         seq = get_bits(&s->gb, 13) << 2;
@@ -410,7 +416,7 @@ static int rv20_decode_picture_header(MpegEncContext *s)
 av_log(s->avctx, AV_LOG_DEBUG, "\n");*/
     s->no_rounding= get_bits1(&s->gb);
 
-    if(RV_GET_MINOR_VER(s->avctx->sub_id) <= 1 && s->pict_type == AV_PICTURE_TYPE_B)
+    if(RV_GET_MINOR_VER(rv->sub_id) <= 1 && s->pict_type == AV_PICTURE_TYPE_B)
         skip_bits(&s->gb, 5); // binary decoder reads 3+2 bits here but they don't seem to be used
 
     s->f_code = 1;
@@ -435,7 +441,8 @@ av_log(s->avctx, AV_LOG_DEBUG, "\n");*/
 
 static av_cold int rv10_decode_init(AVCodecContext *avctx)
 {
-    MpegEncContext *s = avctx->priv_data;
+    RVDecContext  *rv = avctx->priv_data;
+    MpegEncContext *s = &rv->m;
     static int done=0;
     int major_ver, minor_ver, micro_ver;
 
@@ -454,11 +461,11 @@ static av_cold int rv10_decode_init(AVCodecContext *avctx)
     s->orig_height= s->height = avctx->coded_height;
 
     s->h263_long_vectors= ((uint8_t*)avctx->extradata)[3] & 1;
-    avctx->sub_id= AV_RB32((uint8_t*)avctx->extradata + 4);
+    rv->sub_id = AV_RB32((uint8_t*)avctx->extradata + 4);
 
-    major_ver = RV_GET_MAJOR_VER(avctx->sub_id);
-    minor_ver = RV_GET_MINOR_VER(avctx->sub_id);
-    micro_ver = RV_GET_MICRO_VER(avctx->sub_id);
+    major_ver = RV_GET_MAJOR_VER(rv->sub_id);
+    minor_ver = RV_GET_MINOR_VER(rv->sub_id);
+    micro_ver = RV_GET_MICRO_VER(rv->sub_id);
 
     s->low_delay = 1;
     switch (major_ver) {
@@ -473,13 +480,13 @@ static av_cold int rv10_decode_init(AVCodecContext *avctx)
         }
         break;
     default:
-        av_log(s->avctx, AV_LOG_ERROR, "unknown header %X\n", avctx->sub_id);
+        av_log(s->avctx, AV_LOG_ERROR, "unknown header %X\n", rv->sub_id);
         av_log_missing_feature(avctx, "RV1/2 version", 1);
         return AVERROR_PATCHWELCOME;
     }
 
     if(avctx->debug & FF_DEBUG_PICT_INFO){
-        av_log(avctx, AV_LOG_DEBUG, "ver:%X ver0:%X\n", avctx->sub_id, avctx->extradata_size >= 4 ? ((uint32_t*)avctx->extradata)[0] : -1);
+        av_log(avctx, AV_LOG_DEBUG, "ver:%X ver0:%X\n", rv->sub_id, avctx->extradata_size >= 4 ? ((uint32_t*)avctx->extradata)[0] : -1);
     }
 
     avctx->pix_fmt = PIX_FMT_YUV420P;
@@ -514,7 +521,8 @@ static av_cold int rv10_decode_end(AVCodecContext *avctx)
 static int rv10_decode_packet(AVCodecContext *avctx,
                              const uint8_t *buf, int buf_size, int buf_size2)
 {
-    MpegEncContext *s = avctx->priv_data;
+    RVDecContext  *rv = avctx->priv_data;
+    MpegEncContext *s = &rv->m;
     int mb_count, mb_pos, left, start_mb_x, active_bits_size;
 
     active_bits_size = buf_size * 8;
@@ -522,7 +530,7 @@ static int rv10_decode_packet(AVCodecContext *avctx,
     if(s->codec_id ==CODEC_ID_RV10)
         mb_count = rv10_decode_picture_header(s);
     else
-        mb_count = rv20_decode_picture_header(s);
+        mb_count = rv20_decode_picture_header(rv);
     if (mb_count < 0) {
         av_log(s->avctx, AV_LOG_ERROR, "HEADER ERROR\n");
         return -1;
@@ -733,7 +741,7 @@ AVCodec ff_rv10_decoder = {
     .name           = "rv10",
     .type           = AVMEDIA_TYPE_VIDEO,
     .id             = CODEC_ID_RV10,
-    .priv_data_size = sizeof(MpegEncContext),
+    .priv_data_size = sizeof(RVDecContext),
     .init           = rv10_decode_init,
     .close          = rv10_decode_end,
     .decode         = rv10_decode_frame,
@@ -747,7 +755,7 @@ AVCodec ff_rv20_decoder = {
     .name           = "rv20",
     .type           = AVMEDIA_TYPE_VIDEO,
     .id             = CODEC_ID_RV20,
-    .priv_data_size = sizeof(MpegEncContext),
+    .priv_data_size = sizeof(RVDecContext),
     .init           = rv10_decode_init,
     .close          = rv10_decode_end,
     .decode         = rv10_decode_frame,

libavcodec/utils.c

@@ -118,12 +118,12 @@ static void avcodec_init(void)
     ff_dsputil_static_init();
 }
 
-static av_always_inline int codec_is_encoder(AVCodec *codec)
+int av_codec_is_encoder(AVCodec *codec)
 {
     return codec && (codec->encode || codec->encode2);
 }
 
-static av_always_inline int codec_is_decoder(AVCodec *codec)
+int av_codec_is_decoder(AVCodec *codec)
 {
     return codec && codec->decode;
 }
@@ -798,7 +798,7 @@ int attribute_align_arg avcodec_open2(AVCodecContext *avctx, AVCodec *codec, AVD
 
     /* if the decoder init function was already called previously,
        free the already allocated subtitle_header before overwriting it */
-    if (codec_is_decoder(codec))
+    if (av_codec_is_decoder(codec))
         av_freep(&avctx->subtitle_header);
 
 #define SANE_NB_CHANNELS 128U
@@ -845,7 +845,7 @@ int attribute_align_arg avcodec_open2(AVCodecContext *avctx, AVCodec *codec, AVD
         ret = AVERROR(EINVAL);
         goto free_and_end;
     }
-    if (codec_is_encoder(avctx->codec)) {
+    if (av_codec_is_encoder(avctx->codec)) {
         int i;
         if (avctx->codec->sample_fmts) {
             for (i = 0; avctx->codec->sample_fmts[i] != AV_SAMPLE_FMT_NONE; i++)
@@ -914,7 +914,7 @@ int attribute_align_arg avcodec_open2(AVCodecContext *avctx, AVCodec *codec, AVD
         }
     }
 
-    if (codec_is_decoder(avctx->codec) && !avctx->bit_rate)
+    if (av_codec_is_decoder(avctx->codec) && !avctx->bit_rate)
         avctx->bit_rate = get_bit_rate(avctx);
 
     ret=0;
@@ -1527,7 +1527,7 @@ av_cold int avcodec_close(AVCodecContext *avctx)
         av_opt_free(avctx->priv_data);
     av_opt_free(avctx);
     av_freep(&avctx->priv_data);
-    if (codec_is_encoder(avctx->codec))
+    if (av_codec_is_encoder(avctx->codec))
         av_freep(&avctx->extradata);
     avctx->codec = NULL;
     avctx->active_thread_type = 0;
@@ -1556,7 +1556,7 @@ AVCodec *avcodec_find_encoder(enum CodecID id)
     p = first_avcodec;
     id= remap_deprecated_codec_id(id);
     while (p) {
-        if (codec_is_encoder(p) && p->id == id) {
+        if (av_codec_is_encoder(p) && p->id == id) {
             if (p->capabilities & CODEC_CAP_EXPERIMENTAL && !experimental) {
                 experimental = p;
             } else
@@ -1574,7 +1574,7 @@ AVCodec *avcodec_find_encoder_by_name(const char *name)
         return NULL;
     p = first_avcodec;
     while (p) {
-        if (codec_is_encoder(p) && strcmp(name,p->name) == 0)
+        if (av_codec_is_encoder(p) && strcmp(name,p->name) == 0)
             return p;
         p = p->next;
     }
@@ -1587,7 +1587,7 @@ AVCodec *avcodec_find_decoder(enum CodecID id)
     p = first_avcodec;
     id= remap_deprecated_codec_id(id);
     while (p) {
-        if (codec_is_decoder(p) && p->id == id) {
+        if (av_codec_is_decoder(p) && p->id == id) {
             if (p->capabilities & CODEC_CAP_EXPERIMENTAL && !experimental) {
                 experimental = p;
             } else
@@ -1605,7 +1605,7 @@ AVCodec *avcodec_find_decoder_by_name(const char *name)
         return NULL;
     p = first_avcodec;
     while (p) {
-        if (codec_is_decoder(p) && strcmp(name,p->name) == 0)
+        if (av_codec_is_decoder(p) && strcmp(name,p->name) == 0)
             return p;
         p = p->next;
     }

libavcodec/version.h

@@ -21,7 +21,7 @@
 #define AVCODEC_VERSION_H
 
 #define LIBAVCODEC_VERSION_MAJOR 54
-#define LIBAVCODEC_VERSION_MINOR  7
+#define LIBAVCODEC_VERSION_MINOR  8
 #define LIBAVCODEC_VERSION_MICRO 100
 
 #define LIBAVCODEC_VERSION_INT  AV_VERSION_INT(LIBAVCODEC_VERSION_MAJOR, \
@@ -69,5 +69,8 @@
 #ifndef FF_API_INTER_THRESHOLD
 #define FF_API_INTER_THRESHOLD  (LIBAVCODEC_VERSION_MAJOR < 55)
 #endif
+#ifndef FF_API_SUB_ID
+#define FF_API_SUB_ID           (LIBAVCODEC_VERSION_MAJOR < 55)
+#endif
 
 #endif /* AVCODEC_VERSION_H */

libavcodec/vorbisenc.c

@@ -137,13 +137,16 @@ typedef struct {
 #define RESIDUE_PART_SIZE      32
 #define NUM_RESIDUE_PARTITIONS (RESIDUE_SIZE/RESIDUE_PART_SIZE)
 
-static inline void put_codeword(PutBitContext *pb, vorbis_enc_codebook *cb,
-                                int entry)
+static inline int put_codeword(PutBitContext *pb, vorbis_enc_codebook *cb,
+                               int entry)
 {
     assert(entry >= 0);
     assert(entry < cb->nentries);
     assert(cb->lens[entry]);
+    if (pb->size_in_bits - put_bits_count(pb) < cb->lens[entry])
+        return AVERROR(EINVAL);
     put_bits(pb, cb->lens[entry], cb->codewords[entry]);
+    return 0;
 }
 
 static int cb_lookup_vals(int lookup, int dimentions, int entries)
@@ -751,14 +754,16 @@ static int render_point(int x0, int y0, int x1, int y1, int x)
     return y0 +  (x - x0) * (y1 - y0) / (x1 - x0);
 }
 
-static void floor_encode(vorbis_enc_context *venc, vorbis_enc_floor *fc,
-                         PutBitContext *pb, uint16_t *posts,
-                         float *floor, int samples)
+static int floor_encode(vorbis_enc_context *venc, vorbis_enc_floor *fc,
+                        PutBitContext *pb, uint16_t *posts,
+                        float *floor, int samples)
 {
     int range = 255 / fc->multiplier + 1;
     int coded[MAX_FLOOR_VALUES]; // first 2 values are unused
     int i, counter;
 
+    if (pb->size_in_bits - put_bits_count(pb) < 1 + 2 * ilog(range - 1))
+        return AVERROR(EINVAL);
     put_bits(pb, 1, 1); // non zero
     put_bits(pb, ilog(range - 1), posts[0]);
     put_bits(pb, ilog(range - 1), posts[1]);
@@ -816,7 +821,8 @@ static void floor_encode(vorbis_enc_context *venc, vorbis_enc_floor *fc,
                 cval   |= l << cshift;
                 cshift += c->subclass;
             }
-            put_codeword(pb, book, cval);
+            if (put_codeword(pb, book, cval))
+                return AVERROR(EINVAL);
         }
         for (k = 0; k < c->dim; k++) {
             int book  = c->books[cval & (csub-1)];
@@ -826,12 +832,15 @@ static void floor_encode(vorbis_enc_context *venc, vorbis_enc_floor *fc,
                 continue;
             if (entry == -1)
                 entry = 0;
-            put_codeword(pb, &venc->codebooks[book], entry);
+            if (put_codeword(pb, &venc->codebooks[book], entry))
+                return AVERROR(EINVAL);
         }
     }
 
     ff_vorbis_floor1_render_list(fc->list, fc->values, posts, coded,
                                  fc->multiplier, floor, samples);
+
+    return 0;
 }
 
 static float *put_vector(vorbis_enc_codebook *book, PutBitContext *pb,
@@ -852,13 +861,14 @@ static float *put_vector(vorbis_enc_codebook *book, PutBitContext *pb,
             distance = d;
         }
     }
-    put_codeword(pb, book, entry);
+    if (put_codeword(pb, book, entry))
+        return NULL;
     return &book->dimentions[entry * book->ndimentions];
 }
 
-static void residue_encode(vorbis_enc_context *venc, vorbis_enc_residue *rc,
-                           PutBitContext *pb, float *coeffs, int samples,
-                           int real_ch)
+static int residue_encode(vorbis_enc_context *venc, vorbis_enc_residue *rc,
+                          PutBitContext *pb, float *coeffs, int samples,
+                          int real_ch)
 {
     int pass, i, j, p, k;
     int psize      = rc->partition_size;
@@ -894,7 +904,8 @@ static void residue_encode(vorbis_enc_context *venc, vorbis_enc_residue *rc,
                         entry *= rc->classifications;
                         entry += classes[j][p + i];
                     }
-                    put_codeword(pb, book, entry);
+                    if (put_codeword(pb, book, entry))
+                        return AVERROR(EINVAL);
                 }
             for (i = 0; i < classwords && p < partitions; i++, p++) {
                 for (j = 0; j < channels; j++) {
@@ -909,8 +920,10 @@ static void residue_encode(vorbis_enc_context *venc, vorbis_enc_residue *rc,
 
                     if (rc->type == 0) {
                         for (k = 0; k < psize; k += book->ndimentions) {
-                            float *a = put_vector(book, pb, &buf[k]);
                             int l;
+                            float *a = put_vector(book, pb, &buf[k]);
+                            if (!a)
+                                return AVERROR(EINVAL);
                             for (l = 0; l < book->ndimentions; l++)
                                 buf[k + l] -= a[l];
                         }
@@ -930,6 +943,8 @@ static void residue_encode(vorbis_enc_context *venc, vorbis_enc_residue *rc,
                                 }
                             }
                             pv = put_vector(book, pb, vec);
+                            if (!pv)
+                                return AVERROR(EINVAL);
                             for (dim = book->ndimentions; dim--; ) {
                                 coeffs[a1 + b1] -= *pv++;
                                 if ((a1 += samples) == s) {
@@ -943,6 +958,7 @@ static void residue_encode(vorbis_enc_context *venc, vorbis_enc_residue *rc,
             }
         }
     }
+    return 0;
 }
 
 static int apply_window_and_mdct(vorbis_enc_context *venc, const signed short *audio,
@@ -1016,6 +1032,11 @@ static int vorbis_encode_frame(AVCodecContext *avccontext,
 
     init_put_bits(&pb, packets, buf_size);
 
+    if (pb.size_in_bits - put_bits_count(&pb) < 1 + ilog(venc->nmodes - 1)) {
+        av_log(avccontext, AV_LOG_ERROR, "output buffer is too small\n");
+        return AVERROR(EINVAL);
+    }
+
     put_bits(&pb, 1, 0); // magic bit
 
     put_bits(&pb, ilog(venc->nmodes - 1), 0); // 0 bits, the mode
@@ -1031,7 +1052,10 @@ static int vorbis_encode_frame(AVCodecContext *avccontext,
         vorbis_enc_floor *fc = &venc->floors[mapping->floor[mapping->mux[i]]];
         uint16_t posts[MAX_FLOOR_VALUES];
         floor_fit(venc, fc, &venc->coeffs[i * samples], posts, samples);
-        floor_encode(venc, fc, &pb, posts, &venc->floor[i * samples], samples);
+        if (floor_encode(venc, fc, &pb, posts, &venc->floor[i * samples], samples)) {
+            av_log(avccontext, AV_LOG_ERROR, "output buffer is too small\n");
+            return AVERROR(EINVAL);
+        }
     }
 
     for (i = 0; i < venc->channels * samples; i++)
@@ -1051,8 +1075,11 @@ static int vorbis_encode_frame(AVCodecContext *avccontext,
         }
     }
 
-    residue_encode(venc, &venc->residues[mapping->residue[mapping->mux[0]]],
-                   &pb, venc->coeffs, samples, venc->channels);
+    if (residue_encode(venc, &venc->residues[mapping->residue[mapping->mux[0]]],
+                       &pb, venc->coeffs, samples, venc->channels)) {
+        av_log(avccontext, AV_LOG_ERROR, "output buffer is too small\n");
+        return AVERROR(EINVAL);
+    }
 
     avccontext->coded_frame->pts = venc->sample_count;
     venc->sample_count += avccontext->frame_size;

libavcodec/wmaenc.c

@@ -39,6 +39,12 @@ static int encode_init(AVCodecContext * avctx){
         return AVERROR(EINVAL);
     }
 
+    if (avctx->sample_rate > 48000) {
+        av_log(avctx, AV_LOG_ERROR, "sample rate is too high: %d > 48kHz",
+               avctx->sample_rate);
+        return AVERROR(EINVAL);
+    }
+
     if(avctx->bit_rate < 24*1000) {
         av_log(avctx, AV_LOG_ERROR, "bitrate too low: got %i, need 24000 or higher\n",
                avctx->bit_rate);
@@ -64,6 +70,8 @@ static int encode_init(AVCodecContext * avctx){
     s->use_exp_vlc = flags2 & 0x0001;
     s->use_bit_reservoir = flags2 & 0x0002;
     s->use_variable_block_len = flags2 & 0x0004;
+    if (avctx->channels == 2)
+        s->ms_stereo = 1;
 
     ff_wma_init(avctx, flags2);
 
@@ -71,8 +79,12 @@ static int encode_init(AVCodecContext * avctx){
     for(i = 0; i < s->nb_block_sizes; i++)
         ff_mdct_init(&s->mdct_ctx[i], s->frame_len_bits - i + 1, 0, 1.0);
 
-    avctx->block_align=
-    s->block_align= avctx->bit_rate*(int64_t)s->frame_len / (avctx->sample_rate*8);
+    s->block_align     = avctx->bit_rate * (int64_t)s->frame_len /
+                         (avctx->sample_rate * 8);
+    s->block_align     = FFMIN(s->block_align, MAX_CODED_SUPERFRAME_SIZE);
+    avctx->block_align = s->block_align;
+    avctx->bit_rate    = avctx->block_align * 8LL * avctx->sample_rate /
+                         s->frame_len;
 //av_log(NULL, AV_LOG_ERROR, "%d %d %d %d\n", s->block_align, avctx->bit_rate, s->frame_len, avctx->sample_rate);
     avctx->frame_size= s->frame_len;
 
@@ -181,7 +193,7 @@ static int encode_block(WMACodecContext *s, float (*src_coefs)[BLOCK_MAX_SIZE],
     }
 
     if (s->nb_channels == 2) {
-        put_bits(&s->pb, 1, s->ms_stereo= 1);
+        put_bits(&s->pb, 1, !!s->ms_stereo);
     }
 
     for(ch = 0; ch < s->nb_channels; ch++) {
@@ -355,6 +367,11 @@ static int encode_superframe(AVCodecContext *avctx,
         }
     }
 
+    if (buf_size < 2 * MAX_CODED_SUPERFRAME_SIZE) {
+        av_log(avctx, AV_LOG_ERROR, "output buffer size is too small\n");
+        return AVERROR(EINVAL);
+    }
+
 #if 1
     total_gain= 128;
     for(i=64; i; i>>=1){
@@ -379,15 +396,17 @@ static int encode_superframe(AVCodecContext *avctx,
     }
 #endif
 
-    encode_frame(s, s->coefs, buf, buf_size, total_gain);
+    if ((i = encode_frame(s, s->coefs, buf, buf_size, total_gain)) >= 0) {
+        av_log(avctx, AV_LOG_ERROR, "required frame size too large. please "
+               "use a higher bit rate.\n");
+        return AVERROR(EINVAL);
+    }
     assert((put_bits_count(&s->pb) & 7) == 0);
-    i= s->block_align - (put_bits_count(&s->pb)+7)/8;
-    assert(i>=0);
-    while(i--)
+    while (i++)
         put_bits(&s->pb, 8, 'N');
 
     flush_put_bits(&s->pb);
-    return put_bits_ptr(&s->pb) - s->pb.buf;
+    return s->block_align;
 }
 
 AVCodec ff_wmav1_encoder = {

libavcodec/x86/vp8dsp-init.c

@@ -29,16 +29,16 @@
 /*
  * MC functions
  */
-extern void ff_put_vp8_epel4_h4_mmxext(uint8_t *dst, ptrdiff_t dststride,
+extern void ff_put_vp8_epel4_h4_mmx2  (uint8_t *dst, ptrdiff_t dststride,
                                        uint8_t *src, ptrdiff_t srcstride,
                                        int height, int mx, int my);
-extern void ff_put_vp8_epel4_h6_mmxext(uint8_t *dst, ptrdiff_t dststride,
+extern void ff_put_vp8_epel4_h6_mmx2  (uint8_t *dst, ptrdiff_t dststride,
                                        uint8_t *src, ptrdiff_t srcstride,
                                        int height, int mx, int my);
-extern void ff_put_vp8_epel4_v4_mmxext(uint8_t *dst, ptrdiff_t dststride,
+extern void ff_put_vp8_epel4_v4_mmx2  (uint8_t *dst, ptrdiff_t dststride,
                                        uint8_t *src, ptrdiff_t srcstride,
                                        int height, int mx, int my);
-extern void ff_put_vp8_epel4_v6_mmxext(uint8_t *dst, ptrdiff_t dststride,
+extern void ff_put_vp8_epel4_v6_mmx2  (uint8_t *dst, ptrdiff_t dststride,
                                        uint8_t *src, ptrdiff_t srcstride,
                                        int height, int mx, int my);
 
@@ -80,7 +80,7 @@ extern void ff_put_vp8_epel8_v6_ssse3 (uint8_t *dst, ptrdiff_t dststride,
                                        uint8_t *src, ptrdiff_t srcstride,
                                        int height, int mx, int my);
 
-extern void ff_put_vp8_bilinear4_h_mmxext(uint8_t *dst, ptrdiff_t dststride,
+extern void ff_put_vp8_bilinear4_h_mmx2  (uint8_t *dst, ptrdiff_t dststride,
                                           uint8_t *src, ptrdiff_t srcstride,
                                           int height, int mx, int my);
 extern void ff_put_vp8_bilinear8_h_sse2  (uint8_t *dst, ptrdiff_t dststride,
@@ -93,7 +93,7 @@ extern void ff_put_vp8_bilinear8_h_ssse3 (uint8_t *dst, ptrdiff_t dststride,
                                           uint8_t *src, ptrdiff_t srcstride,
                                           int height, int mx, int my);
 
-extern void ff_put_vp8_bilinear4_v_mmxext(uint8_t *dst, ptrdiff_t dststride,
+extern void ff_put_vp8_bilinear4_v_mmx2  (uint8_t *dst, ptrdiff_t dststride,
                                           uint8_t *src, ptrdiff_t srcstride,
                                           int height, int mx, int my);
 extern void ff_put_vp8_bilinear8_v_sse2  (uint8_t *dst, ptrdiff_t dststride,
@@ -139,27 +139,27 @@ static void ff_put_vp8_ ## FILTERTYPE ## 8_ ## TAPTYPE ## _ ## OPT( \
 }
 
 #if ARCH_X86_32
-TAP_W8 (mmxext, epel, h4)
-TAP_W8 (mmxext, epel, h6)
-TAP_W16(mmxext, epel, h6)
-TAP_W8 (mmxext, epel, v4)
-TAP_W8 (mmxext, epel, v6)
-TAP_W16(mmxext, epel, v6)
-TAP_W8 (mmxext, bilinear, h)
-TAP_W16(mmxext, bilinear, h)
-TAP_W8 (mmxext, bilinear, v)
-TAP_W16(mmxext, bilinear, v)
+TAP_W8 (mmx2,  epel, h4)
+TAP_W8 (mmx2,  epel, h6)
+TAP_W16(mmx2,  epel, h6)
+TAP_W8 (mmx2,  epel, v4)
+TAP_W8 (mmx2,  epel, v6)
+TAP_W16(mmx2,  epel, v6)
+TAP_W8 (mmx2,  bilinear, h)
+TAP_W16(mmx2,  bilinear, h)
+TAP_W8 (mmx2,  bilinear, v)
+TAP_W16(mmx2,  bilinear, v)
 #endif
 
-TAP_W16(sse2,   epel, h6)
-TAP_W16(sse2,   epel, v6)
-TAP_W16(sse2,   bilinear, h)
-TAP_W16(sse2,   bilinear, v)
+TAP_W16(sse2,  epel, h6)
+TAP_W16(sse2,  epel, v6)
+TAP_W16(sse2,  bilinear, h)
+TAP_W16(sse2,  bilinear, v)
 
-TAP_W16(ssse3,  epel, h6)
-TAP_W16(ssse3,  epel, v6)
-TAP_W16(ssse3,  bilinear, h)
-TAP_W16(ssse3,  bilinear, v)
+TAP_W16(ssse3, epel, h6)
+TAP_W16(ssse3, epel, v6)
+TAP_W16(ssse3, bilinear, h)
+TAP_W16(ssse3, bilinear, v)
 
 #define HVTAP(OPT, ALIGN, TAPNUMX, TAPNUMY, SIZE, MAXHEIGHT) \
 static void ff_put_vp8_epel ## SIZE ## _h ## TAPNUMX ## v ## TAPNUMY ## _ ## OPT( \
@@ -177,13 +177,13 @@ static void ff_put_vp8_epel ## SIZE ## _h ## TAPNUMX ## v ## TAPNUMY ## _ ## OPT
 
 #if ARCH_X86_32
 #define HVTAPMMX(x, y) \
-HVTAP(mmxext, 8, x, y,  4,  8) \
-HVTAP(mmxext, 8, x, y,  8, 16)
+HVTAP(mmx2, 8, x, y,  4,  8) \
+HVTAP(mmx2, 8, x, y,  8, 16)
 
-HVTAP(mmxext, 8, 6, 6, 16, 16)
+HVTAP(mmx2, 8, 6, 6, 16, 16)
 #else
 #define HVTAPMMX(x, y) \
-HVTAP(mmxext, 8, x, y,  4,  8)
+HVTAP(mmx2, 8, x, y,  4,  8)
 #endif
 
 HVTAPMMX(4, 4)
@@ -218,16 +218,16 @@ static void ff_put_vp8_bilinear ## SIZE ## _hv_ ## OPT( \
         dst, dststride, tmp, SIZE,      height,     mx, my); \
 }
 
-HVBILIN(mmxext, 8,  4,  8)
+HVBILIN(mmx2,  8,  4,  8)
 #if ARCH_X86_32
-HVBILIN(mmxext, 8,  8, 16)
-HVBILIN(mmxext, 8, 16, 16)
+HVBILIN(mmx2,  8,  8, 16)
+HVBILIN(mmx2,  8, 16, 16)
 #endif
-HVBILIN(sse2,   8,  8, 16)
-HVBILIN(sse2,   8, 16, 16)
-HVBILIN(ssse3,  8,  4,  8)
-HVBILIN(ssse3,  8,  8, 16)
-HVBILIN(ssse3,  8, 16, 16)
+HVBILIN(sse2,  8,  8, 16)
+HVBILIN(sse2,  8, 16, 16)
+HVBILIN(ssse3, 8,  4,  8)
+HVBILIN(ssse3, 8,  8, 16)
+HVBILIN(ssse3, 8, 16, 16)
 
 extern void ff_vp8_idct_dc_add_mmx(uint8_t *dst, DCTELEM block[16],
                                    ptrdiff_t stride);
@@ -283,7 +283,7 @@ extern void ff_vp8_h_loop_filter8uv_mbedge_ ## NAME(uint8_t *dstU, \
                                                     int e, int i, int hvt);
 
 DECLARE_LOOP_FILTER(mmx)
-DECLARE_LOOP_FILTER(mmxext)
+DECLARE_LOOP_FILTER(mmx2)
 DECLARE_LOOP_FILTER(sse2)
 DECLARE_LOOP_FILTER(ssse3)
 DECLARE_LOOP_FILTER(sse4)
@@ -351,26 +351,26 @@ av_cold void ff_vp8dsp_init_x86(VP8DSPContext* c)
     /* note that 4-tap width=16 functions are missing because w=16
      * is only used for luma, and luma is always a copy or sixtap. */
     if (mm_flags & AV_CPU_FLAG_MMX2) {
-        VP8_MC_FUNC(2, 4, mmxext);
-        VP8_BILINEAR_MC_FUNC(2, 4, mmxext);
+        VP8_MC_FUNC(2, 4, mmx2);
+        VP8_BILINEAR_MC_FUNC(2, 4, mmx2);
 #if ARCH_X86_32
-        VP8_LUMA_MC_FUNC(0, 16, mmxext);
-        VP8_MC_FUNC(1, 8, mmxext);
-        VP8_BILINEAR_MC_FUNC(0, 16, mmxext);
-        VP8_BILINEAR_MC_FUNC(1, 8, mmxext);
-
-        c->vp8_v_loop_filter_simple = ff_vp8_v_loop_filter_simple_mmxext;
-        c->vp8_h_loop_filter_simple = ff_vp8_h_loop_filter_simple_mmxext;
-
-        c->vp8_v_loop_filter16y_inner = ff_vp8_v_loop_filter16y_inner_mmxext;
-        c->vp8_h_loop_filter16y_inner = ff_vp8_h_loop_filter16y_inner_mmxext;
-        c->vp8_v_loop_filter8uv_inner = ff_vp8_v_loop_filter8uv_inner_mmxext;
-        c->vp8_h_loop_filter8uv_inner = ff_vp8_h_loop_filter8uv_inner_mmxext;
-
-        c->vp8_v_loop_filter16y       = ff_vp8_v_loop_filter16y_mbedge_mmxext;
-        c->vp8_h_loop_filter16y       = ff_vp8_h_loop_filter16y_mbedge_mmxext;
-        c->vp8_v_loop_filter8uv       = ff_vp8_v_loop_filter8uv_mbedge_mmxext;
-        c->vp8_h_loop_filter8uv       = ff_vp8_h_loop_filter8uv_mbedge_mmxext;
+        VP8_LUMA_MC_FUNC(0, 16, mmx2);
+        VP8_MC_FUNC(1, 8, mmx2);
+        VP8_BILINEAR_MC_FUNC(0, 16, mmx2);
+        VP8_BILINEAR_MC_FUNC(1, 8, mmx2);
+
+        c->vp8_v_loop_filter_simple = ff_vp8_v_loop_filter_simple_mmx2;
+        c->vp8_h_loop_filter_simple = ff_vp8_h_loop_filter_simple_mmx2;
+
+        c->vp8_v_loop_filter16y_inner = ff_vp8_v_loop_filter16y_inner_mmx2;
+        c->vp8_h_loop_filter16y_inner = ff_vp8_h_loop_filter16y_inner_mmx2;
+        c->vp8_v_loop_filter8uv_inner = ff_vp8_v_loop_filter8uv_inner_mmx2;
+        c->vp8_h_loop_filter8uv_inner = ff_vp8_h_loop_filter8uv_inner_mmx2;
+
+        c->vp8_v_loop_filter16y       = ff_vp8_v_loop_filter16y_mbedge_mmx2;
+        c->vp8_h_loop_filter16y       = ff_vp8_h_loop_filter16y_mbedge_mmx2;
+        c->vp8_v_loop_filter8uv       = ff_vp8_v_loop_filter8uv_mbedge_mmx2;
+        c->vp8_h_loop_filter8uv       = ff_vp8_h_loop_filter8uv_mbedge_mmx2;
 #endif
     }
 

libavutil/x86/x86inc.asm

@@ -251,6 +251,8 @@ DECLARE_REG_TMP_SIZE 0,1,2,3,4,5,6,7,8,9
         %endrep
     %endif
 
+    %xdefine %%stack_offset stack_offset
+    %undef stack_offset ; so that the current value of stack_offset doesn't get baked in by xdefine
     %assign %%i 0
     %rep %0
         %xdefine %1q r %+ %%i %+ q
@@ -262,7 +264,8 @@ DECLARE_REG_TMP_SIZE 0,1,2,3,4,5,6,7,8,9
         %assign %%i %%i+1
         %rotate 1
     %endrep
-    %assign n_arg_names %%i
+    %xdefine stack_offset %%stack_offset
+    %assign n_arg_names %0
 %endmacro
 
 %if WIN64 ; Windows x64 ;=================================================

libswscale/options.c

@@ -34,7 +34,7 @@ static const char *sws_context_to_name(void *ptr)
 #define VE AV_OPT_FLAG_VIDEO_PARAM | AV_OPT_FLAG_ENCODING_PARAM
 
 static const AVOption options[] = {
-    { "sws_flags",       "scaler/cpu flags",              OFFSET(flags),     AV_OPT_TYPE_FLAGS,  { .dbl = DEFAULT            }, 0,       UINT_MAX,       VE, "sws_flags" },
+    { "sws_flags",       "scaler flags",                  OFFSET(flags),     AV_OPT_TYPE_FLAGS,  { .dbl = DEFAULT            }, 0,       UINT_MAX,       VE, "sws_flags" },
     { "fast_bilinear",   "fast bilinear",                 0,                 AV_OPT_TYPE_CONST,  { .dbl = SWS_FAST_BILINEAR  }, INT_MIN, INT_MAX,        VE, "sws_flags" },
     { "bilinear",        "bilinear",                      0,                 AV_OPT_TYPE_CONST,  { .dbl = SWS_BILINEAR       }, INT_MIN, INT_MAX,        VE, "sws_flags" },
     { "bicubic",         "bicubic",                       0,                 AV_OPT_TYPE_CONST,  { .dbl = SWS_BICUBIC        }, INT_MIN, INT_MAX,        VE, "sws_flags" },

libswscale/x86/output.asm

@@ -131,8 +131,12 @@ cglobal yuv2planeX_%1, %3, 7, %2, filter, fltsize, src, dst, w, dither, offset
     ; pixels per iteration. In order to not have to keep track of where
     ; we are w.r.t. dithering, we unroll the mmx/8bit loop x2.
 %if %1 == 8
-%rep 16/mmsize
-%endif ; %1 == 8
+%assign %%repcnt 16/mmsize
+%else
+%assign %%repcnt 1
+%endif
+
+%rep %%repcnt
 
 %if %1 == 8
 %if ARCH_X86_32
@@ -146,7 +150,7 @@ cglobal yuv2planeX_%1, %3, 7, %2, filter, fltsize, src, dst, w, dither, offset
     mova            m1, [yuv2yuvX_%1_start]
     mova            m2,  m1
 %endif ; %1 == 8/9/10/16
-    movsx     cntr_reg,  r1m ; FIXME should be fltsizem, but the assembler does the wrong thing b/c of SUB above
+    movsx     cntr_reg,  fltsizem
 .filterloop_ %+ %%i:
     ; input pixels
     mov             r6, [srcq+gprsize*cntr_reg-2*gprsize]
@@ -226,10 +230,9 @@ cglobal yuv2planeX_%1, %3, 7, %2, filter, fltsize, src, dst, w, dither, offset
 
     add             r5,  mmsize/2
     sub             wd,  mmsize/2
-%if %1 == 8
+
 %assign %%i %%i+2
 %endrep
-%endif ; %1 == 8
     jg .pixelloop
 
 %if %1 == 8

tests/lavf-regression.sh

@@ -223,7 +223,7 @@ if [ -n "$do_xwd" ] ; then
 do_image_formats xwd
 fi
 
-if [ -n "$do_sun" ] ; then
+if [ -n "$do_sunrast" ] ; then
 do_image_formats sun
 fi
 

tests/ref/acodec/wmav1

@@ -1,4 +1,4 @@
-26a7f6b0f0b7181df8df3fa589f6bf81 *./tests/data/acodec/wmav1.asf
+0260385b8a54df11ad349f9ba8240fd8 *./tests/data/acodec/wmav1.asf
 106004 ./tests/data/acodec/wmav1.asf
-stddev:12245.52 PSNR: 14.57 MAXDIFF:65521 bytes:  1064960/  1058400
-stddev: 2095.89 PSNR: 29.90 MAXDIFF:27658 bytes:  1056768/  1058400
+stddev:12241.90 PSNR: 14.57 MAXDIFF:65521 bytes:  1064960/  1058400
+stddev: 2074.79 PSNR: 29.99 MAXDIFF:27658 bytes:  1056768/  1058400

tests/ref/acodec/wmav2

@@ -1,4 +1,4 @@
-7c6c0cb692af01b312ae345723674b5f *./tests/data/acodec/wmav2.asf
+bdb4c312fb109f990be83a70f8ec9bdc *./tests/data/acodec/wmav2.asf
 106044 ./tests/data/acodec/wmav2.asf
-stddev:12249.93 PSNR: 14.57 MAXDIFF:65521 bytes:  1064960/  1058400
-stddev: 2089.21 PSNR: 29.93 MAXDIFF:27650 bytes:  1056768/  1058400
+stddev:12246.35 PSNR: 14.57 MAXDIFF:65521 bytes:  1064960/  1058400
+stddev: 2068.08 PSNR: 30.02 MAXDIFF:27650 bytes:  1056768/  1058400