Merge remote-tracking branch 'qatar/master'

* qatar/master:
  qdm2: remove broken and disabled dump_context() debug function
  x86: h264_intrapred: use newly introduced SPLAT* and PSHUFLW macros
  x86inc: add SPLATB_LOAD, SPLATB_REG, PSHUFLW macros
  x86inc: modify ALIGN to not generate long nops on i586
  x86: h264_intrapred: port to cpuflag macros
  avplay: update input filter pointer when the filtergraph is reset.
  avconv: fix parsing of -force_key_frames option.
  h264: use templates to avoid excessive inlining
  xtea: Make the count parameter match the documentation
  blowfish: Make the count parameter match the documentation
  mpegvideo: Don't use ff_mspel_motion() for vc1
  xtea: invert branch and loop precedence
  blowfish: invert branch and loop precedence
  flvdec: optionally trust the metadata
  avconv: Set audio filter time base to the sample rate
  vp8: Add ifdef guards around the sse2 loopfilter in the sse2slow branch too

Conflicts:
	ffmpeg.c
	ffplay.c
	libavcodec/h264.c
	libavcodec/mpegvideo_common.h
Merged-by: Michael Niedermayer <michaelni@gmx.at>

ffmpeg.c

@@ -1077,7 +1077,7 @@ static int configure_input_audio_filter(FilterGraph *fg, InputFilter *ifilter,
 
     snprintf(args, sizeof(args), "time_base=%d/%d:sample_rate=%d:sample_fmt=%s"
              ":channel_layout=0x%"PRIx64,
-             ist->st->time_base.num, ist->st->time_base.den,
+             1, ist->st->codec->sample_rate,
              ist->st->codec->sample_rate,
              av_get_sample_fmt_name(ist->st->codec->sample_fmt),
              ist->st->codec->channel_layout);
@@ -2451,6 +2451,10 @@ static int decode_audio(InputStream *ist, AVPacket *pkt, int *got_output)
             }
     }
 
+    if (decoded_frame->pts != AV_NOPTS_VALUE)
+        decoded_frame->pts = av_rescale_q(decoded_frame->pts,
+                                          ist->st->time_base,
+                                          (AVRational){1, ist->st->codec->sample_rate});
     for (i = 0; i < ist->nb_filters; i++)
         av_buffersrc_add_frame(ist->filters[i]->filter, decoded_frame, 0);
 
@@ -2804,12 +2808,17 @@ static void parse_forced_key_frames(char *kf, OutputStream *ost,
         av_log(NULL, AV_LOG_FATAL, "Could not allocate forced key frames array.\n");
         exit_program(1);
     }
+
     p = kf;
     for (i = 0; i < n; i++) {
         char *next = strchr(p, ',');
-        if (next) *next++ = 0;
+
+        if (next)
+            *next++ = 0;
+
         t = parse_time_or_die("force_key_frames", p, 1);
         ost->forced_kf_pts[i] = av_rescale_q(t, AV_TIME_BASE_Q, avctx->time_base);
+
         p = next;
     }
 }

libavcodec/h264_mc_template.c

+/*
+ * H.26L/H.264/AVC/JVT/14496-10/... decoder
+ * Copyright (c) 2003 Michael Niedermayer <michaelni@gmx.at>
+ *
+ * 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
+ */
+
+#undef MCFUNC
+
+#if   CHROMA_IDC == 1
+#   define MCFUNC(n) FUNC(n ## _420)
+#elif CHROMA_IDC == 2
+#   define MCFUNC(n) FUNC(n ## _422)
+#elif CHROMA_IDC == 3
+#   define MCFUNC(n) FUNC(n ## _444)
+#endif
+
+#undef  mc_part
+#define mc_part MCFUNC(mc_part)
+
+static void mc_part(H264Context *h, int n, int square,
+                    int height, int delta,
+                    uint8_t *dest_y, uint8_t *dest_cb,
+                    uint8_t *dest_cr,
+                    int x_offset, int y_offset,
+                    qpel_mc_func *qpix_put,
+                    h264_chroma_mc_func chroma_put,
+                    qpel_mc_func *qpix_avg,
+                    h264_chroma_mc_func chroma_avg,
+                    h264_weight_func *weight_op,
+                    h264_biweight_func *weight_avg,
+                    int list0, int list1)
+{
+    if ((h->use_weight == 2 && list0 && list1 &&
+         (h->implicit_weight[h->ref_cache[0][scan8[n]]][h->ref_cache[1][scan8[n]]][h->s.mb_y & 1] != 32)) ||
+        h->use_weight == 1)
+        mc_part_weighted(h, n, square, height, delta, dest_y, dest_cb, dest_cr,
+                         x_offset, y_offset, qpix_put, chroma_put,
+                         weight_op[0], weight_op[1], weight_avg[0],
+                         weight_avg[1], list0, list1, PIXEL_SHIFT, CHROMA_IDC);
+    else
+        mc_part_std(h, n, square, height, delta, dest_y, dest_cb, dest_cr,
+                    x_offset, y_offset, qpix_put, chroma_put, qpix_avg,
+                    chroma_avg, list0, list1, PIXEL_SHIFT, CHROMA_IDC);
+}
+
+static void MCFUNC(hl_motion)(H264Context *h, uint8_t *dest_y,
+                              uint8_t *dest_cb, uint8_t *dest_cr,
+                              qpel_mc_func(*qpix_put)[16],
+                              h264_chroma_mc_func(*chroma_put),
+                              qpel_mc_func(*qpix_avg)[16],
+                              h264_chroma_mc_func(*chroma_avg),
+                              h264_weight_func *weight_op,
+                              h264_biweight_func *weight_avg)
+{
+    MpegEncContext *const s = &h->s;
+    const int mb_xy   = h->mb_xy;
+    const int mb_type = s->current_picture.f.mb_type[mb_xy];
+
+    assert(IS_INTER(mb_type));
+
+    if (HAVE_THREADS && (s->avctx->active_thread_type & FF_THREAD_FRAME))
+        await_references(h);
+    prefetch_motion(h, 0, PIXEL_SHIFT, CHROMA_IDC);
+
+    if (IS_16X16(mb_type)) {
+        mc_part(h, 0, 1, 16, 0, dest_y, dest_cb, dest_cr, 0, 0,
+                qpix_put[0], chroma_put[0], qpix_avg[0], chroma_avg[0],
+                weight_op, weight_avg,
+                IS_DIR(mb_type, 0, 0), IS_DIR(mb_type, 0, 1));
+    } else if (IS_16X8(mb_type)) {
+        mc_part(h, 0, 0, 8, 8 << PIXEL_SHIFT, dest_y, dest_cb, dest_cr, 0, 0,
+                qpix_put[1], chroma_put[0], qpix_avg[1], chroma_avg[0],
+                weight_op, weight_avg,
+                IS_DIR(mb_type, 0, 0), IS_DIR(mb_type, 0, 1));
+        mc_part(h, 8, 0, 8, 8 << PIXEL_SHIFT, dest_y, dest_cb, dest_cr, 0, 4,
+                qpix_put[1], chroma_put[0], qpix_avg[1], chroma_avg[0],
+                weight_op, weight_avg,
+                IS_DIR(mb_type, 1, 0), IS_DIR(mb_type, 1, 1));
+    } else if (IS_8X16(mb_type)) {
+        mc_part(h, 0, 0, 16, 8 * h->mb_linesize, dest_y, dest_cb, dest_cr, 0, 0,
+                qpix_put[1], chroma_put[1], qpix_avg[1], chroma_avg[1],
+                &weight_op[1], &weight_avg[1],
+                IS_DIR(mb_type, 0, 0), IS_DIR(mb_type, 0, 1));
+        mc_part(h, 4, 0, 16, 8 * h->mb_linesize, dest_y, dest_cb, dest_cr, 4, 0,
+                qpix_put[1], chroma_put[1], qpix_avg[1], chroma_avg[1],
+                &weight_op[1], &weight_avg[1],
+                IS_DIR(mb_type, 1, 0), IS_DIR(mb_type, 1, 1));
+    } else {
+        int i;
+
+        assert(IS_8X8(mb_type));
+
+        for (i = 0; i < 4; i++) {
+            const int sub_mb_type = h->sub_mb_type[i];
+            const int n  = 4 * i;
+            int x_offset = (i & 1) << 2;
+            int y_offset = (i & 2) << 1;
+
+            if (IS_SUB_8X8(sub_mb_type)) {
+                mc_part(h, n, 1, 8, 0, dest_y, dest_cb, dest_cr,
+                        x_offset, y_offset,
+                        qpix_put[1], chroma_put[1], qpix_avg[1], chroma_avg[1],
+                        &weight_op[1], &weight_avg[1],
+                        IS_DIR(sub_mb_type, 0, 0), IS_DIR(sub_mb_type, 0, 1));
+            } else if (IS_SUB_8X4(sub_mb_type)) {
+                mc_part(h, n, 0, 4, 4 << PIXEL_SHIFT, dest_y, dest_cb, dest_cr,
+                        x_offset, y_offset,
+                        qpix_put[2], chroma_put[1], qpix_avg[2], chroma_avg[1],
+                        &weight_op[1], &weight_avg[1],
+                        IS_DIR(sub_mb_type, 0, 0), IS_DIR(sub_mb_type, 0, 1));
+                mc_part(h, n + 2, 0, 4, 4 << PIXEL_SHIFT,
+                        dest_y, dest_cb, dest_cr, x_offset, y_offset + 2,
+                        qpix_put[2], chroma_put[1], qpix_avg[2], chroma_avg[1],
+                        &weight_op[1], &weight_avg[1],
+                        IS_DIR(sub_mb_type, 0, 0), IS_DIR(sub_mb_type, 0, 1));
+            } else if (IS_SUB_4X8(sub_mb_type)) {
+                mc_part(h, n, 0, 8, 4 * h->mb_linesize,
+                        dest_y, dest_cb, dest_cr, x_offset, y_offset,
+                        qpix_put[2], chroma_put[2], qpix_avg[2], chroma_avg[2],
+                        &weight_op[2], &weight_avg[2],
+                        IS_DIR(sub_mb_type, 0, 0), IS_DIR(sub_mb_type, 0, 1));
+                mc_part(h, n + 1, 0, 8, 4 * h->mb_linesize,
+                        dest_y, dest_cb, dest_cr, x_offset + 2, y_offset,
+                        qpix_put[2], chroma_put[2], qpix_avg[2], chroma_avg[2],
+                        &weight_op[2], &weight_avg[2],
+                        IS_DIR(sub_mb_type, 0, 0), IS_DIR(sub_mb_type, 0, 1));
+            } else {
+                int j;
+                assert(IS_SUB_4X4(sub_mb_type));
+                for (j = 0; j < 4; j++) {
+                    int sub_x_offset = x_offset + 2 * (j & 1);
+                    int sub_y_offset = y_offset + (j & 2);
+                    mc_part(h, n + j, 1, 4, 0,
+                            dest_y, dest_cb, dest_cr, sub_x_offset, sub_y_offset,
+                            qpix_put[2], chroma_put[2], qpix_avg[2], chroma_avg[2],
+                            &weight_op[2], &weight_avg[2],
+                            IS_DIR(sub_mb_type, 0, 0), IS_DIR(sub_mb_type, 0, 1));
+                }
+            }
+        }
+    }
+
+    prefetch_motion(h, 1, PIXEL_SHIFT, CHROMA_IDC);
+}
+

libavcodec/mpegvideo_common.h

@@ -720,7 +720,8 @@ static av_always_inline void MPV_motion_internal(MpegEncContext *s,
                         0, 0, 0,
                         ref_picture, pix_op, qpix_op,
                         s->mv[dir][0][0], s->mv[dir][0][1], 16);
-        }else if(!is_mpeg12 && (CONFIG_WMV2_DECODER || CONFIG_WMV2_ENCODER) && s->mspel && s->codec_id == CODEC_ID_WMV2){
+        } else if (!is_mpeg12 && (CONFIG_WMV2_DECODER || CONFIG_WMV2_ENCODER) &&
+                   s->mspel && s->codec_id == CODEC_ID_WMV2) {
             ff_mspel_motion(s, dest_y, dest_cb, dest_cr,
                         ref_picture, pix_op,
                         s->mv[dir][0][0], s->mv[dir][0][1], 16);

libavcodec/qdm2.c

@@ -1696,51 +1696,6 @@ static av_cold void qdm2_init(QDM2Context *q) {
 }
 
 
-#if 0
-static void dump_context(QDM2Context *q)
-{
-    int i;
-#define PRINT(a,b) av_log(NULL,AV_LOG_DEBUG," %s = %d\n", a, b);
-    PRINT("compressed_data",q->compressed_data);
-    PRINT("compressed_size",q->compressed_size);
-    PRINT("frame_size",q->frame_size);
-    PRINT("checksum_size",q->checksum_size);
-    PRINT("channels",q->channels);
-    PRINT("nb_channels",q->nb_channels);
-    PRINT("fft_size",q->fft_size);
-    PRINT("sub_sampling",q->sub_sampling);
-    PRINT("fft_order",q->fft_order);
-    PRINT("group_order",q->group_order);
-    PRINT("group_size",q->group_size);
-    PRINT("sub_packet",q->sub_packet);
-    PRINT("frequency_range",q->frequency_range);
-    PRINT("has_errors",q->has_errors);
-    PRINT("fft_tone_end",q->fft_tone_end);
-    PRINT("fft_tone_start",q->fft_tone_start);
-    PRINT("fft_coefs_index",q->fft_coefs_index);
-    PRINT("coeff_per_sb_select",q->coeff_per_sb_select);
-    PRINT("cm_table_select",q->cm_table_select);
-    PRINT("noise_idx",q->noise_idx);
-
-    for (i = q->fft_tone_start; i < q->fft_tone_end; i++)
-    {
-    FFTTone *t = &q->fft_tones[i];
-
-    av_log(NULL,AV_LOG_DEBUG,"Tone (%d) dump:\n", i);
-    av_log(NULL,AV_LOG_DEBUG,"  level = %f\n", t->level);
-//  PRINT(" level", t->level);
-    PRINT(" phase", t->phase);
-    PRINT(" phase_shift", t->phase_shift);
-    PRINT(" duration", t->duration);
-    PRINT(" samples_im", t->samples_im);
-    PRINT(" samples_re", t->samples_re);
-    PRINT(" table", t->table);
-    }
-
-}
-#endif
-
-
 /**
  * Init parameters from codec extradata
  */
@@ -1922,7 +1877,6 @@ static av_cold int qdm2_decode_init(AVCodecContext *avctx)
     avcodec_get_frame_defaults(&s->frame);
     avctx->coded_frame = &s->frame;
 
-//    dump_context(s);
     return 0;
 }
 
@@ -1949,8 +1903,6 @@ static int qdm2_decode (QDM2Context *q, const uint8_t *in, int16_t *out)
     q->compressed_data = in;
     q->compressed_size = q->checksum_size;
 
-//  dump_context(q);
-
     /* copy old block, clear new block of output samples */
     memmove(q->output_buffer, &q->output_buffer[frame_size], frame_size * sizeof(float));
     memset(&q->output_buffer[frame_size], 0, frame_size * sizeof(float));

libavcodec/x86/h264_intrapred_init.c

@@ -96,9 +96,9 @@ PRED16x16(horizontal, 10, sse2)
 void ff_pred16x16_vertical_mmx     (uint8_t *src, int stride);
 void ff_pred16x16_vertical_sse     (uint8_t *src, int stride);
 void ff_pred16x16_horizontal_mmx   (uint8_t *src, int stride);
-void ff_pred16x16_horizontal_mmxext(uint8_t *src, int stride);
+void ff_pred16x16_horizontal_mmx2  (uint8_t *src, int stride);
 void ff_pred16x16_horizontal_ssse3 (uint8_t *src, int stride);
-void ff_pred16x16_dc_mmxext        (uint8_t *src, int stride);
+void ff_pred16x16_dc_mmx2          (uint8_t *src, int stride);
 void ff_pred16x16_dc_sse2          (uint8_t *src, int stride);
 void ff_pred16x16_dc_ssse3         (uint8_t *src, int stride);
 void ff_pred16x16_plane_h264_mmx   (uint8_t *src, int stride);
@@ -114,21 +114,21 @@ void ff_pred16x16_plane_svq3_mmx2  (uint8_t *src, int stride);
 void ff_pred16x16_plane_svq3_sse2  (uint8_t *src, int stride);
 void ff_pred16x16_plane_svq3_ssse3 (uint8_t *src, int stride);
 void ff_pred16x16_tm_vp8_mmx       (uint8_t *src, int stride);
-void ff_pred16x16_tm_vp8_mmxext    (uint8_t *src, int stride);
+void ff_pred16x16_tm_vp8_mmx2      (uint8_t *src, int stride);
 void ff_pred16x16_tm_vp8_sse2      (uint8_t *src, int stride);
 void ff_pred8x8_top_dc_mmxext      (uint8_t *src, int stride);
 void ff_pred8x8_dc_rv40_mmxext     (uint8_t *src, int stride);
 void ff_pred8x8_dc_mmxext          (uint8_t *src, int stride);
 void ff_pred8x8_vertical_mmx       (uint8_t *src, int stride);
 void ff_pred8x8_horizontal_mmx     (uint8_t *src, int stride);
-void ff_pred8x8_horizontal_mmxext  (uint8_t *src, int stride);
+void ff_pred8x8_horizontal_mmx2    (uint8_t *src, int stride);
 void ff_pred8x8_horizontal_ssse3   (uint8_t *src, int stride);
 void ff_pred8x8_plane_mmx          (uint8_t *src, int stride);
 void ff_pred8x8_plane_mmx2         (uint8_t *src, int stride);
 void ff_pred8x8_plane_sse2         (uint8_t *src, int stride);
 void ff_pred8x8_plane_ssse3        (uint8_t *src, int stride);
 void ff_pred8x8_tm_vp8_mmx         (uint8_t *src, int stride);
-void ff_pred8x8_tm_vp8_mmxext      (uint8_t *src, int stride);
+void ff_pred8x8_tm_vp8_mmx2        (uint8_t *src, int stride);
 void ff_pred8x8_tm_vp8_sse2        (uint8_t *src, int stride);
 void ff_pred8x8_tm_vp8_ssse3       (uint8_t *src, int stride);
 void ff_pred8x8l_top_dc_mmxext     (uint8_t *src, int has_topleft, int has_topright, int stride);
@@ -163,7 +163,7 @@ void ff_pred4x4_vertical_right_mmxext(uint8_t *src, const uint8_t *topright, int
 void ff_pred4x4_horizontal_up_mmxext(uint8_t *src, const uint8_t *topright, int stride);
 void ff_pred4x4_horizontal_down_mmxext(uint8_t *src, const uint8_t *topright, int stride);
 void ff_pred4x4_tm_vp8_mmx         (uint8_t *src, const uint8_t *topright, int stride);
-void ff_pred4x4_tm_vp8_mmxext      (uint8_t *src, const uint8_t *topright, int stride);
+void ff_pred4x4_tm_vp8_mmx2        (uint8_t *src, const uint8_t *topright, int stride);
 void ff_pred4x4_tm_vp8_ssse3       (uint8_t *src, const uint8_t *topright, int stride);
 void ff_pred4x4_vertical_vp8_mmxext(uint8_t *src, const uint8_t *topright, int stride);
 
@@ -199,10 +199,10 @@ void ff_h264_pred_init_x86(H264PredContext *h, int codec_id, const int bit_depth
         }
 
         if (mm_flags & AV_CPU_FLAG_MMX2) {
-            h->pred16x16[HOR_PRED8x8            ] = ff_pred16x16_horizontal_mmxext;
-            h->pred16x16[DC_PRED8x8             ] = ff_pred16x16_dc_mmxext;
+            h->pred16x16[HOR_PRED8x8            ] = ff_pred16x16_horizontal_mmx2;
+            h->pred16x16[DC_PRED8x8             ] = ff_pred16x16_dc_mmx2;
             if (chroma_format_idc == 1)
-                h->pred8x8[HOR_PRED8x8          ] = ff_pred8x8_horizontal_mmxext;
+                h->pred8x8[HOR_PRED8x8          ] = ff_pred8x8_horizontal_mmx2;
             h->pred8x8l [TOP_DC_PRED            ] = ff_pred8x8l_top_dc_mmxext;
             h->pred8x8l [DC_PRED                ] = ff_pred8x8l_dc_mmxext;
             h->pred8x8l [HOR_PRED               ] = ff_pred8x8l_horizontal_mmxext;
@@ -232,10 +232,10 @@ void ff_h264_pred_init_x86(H264PredContext *h, int codec_id, const int bit_depth
                 }
             }
             if (codec_id == CODEC_ID_VP8) {
-                h->pred16x16[PLANE_PRED8x8      ] = ff_pred16x16_tm_vp8_mmxext;
+                h->pred16x16[PLANE_PRED8x8      ] = ff_pred16x16_tm_vp8_mmx2;
                 h->pred8x8  [DC_PRED8x8         ] = ff_pred8x8_dc_rv40_mmxext;
-                h->pred8x8  [PLANE_PRED8x8      ] = ff_pred8x8_tm_vp8_mmxext;
-                h->pred4x4  [TM_VP8_PRED        ] = ff_pred4x4_tm_vp8_mmxext;
+                h->pred8x8  [PLANE_PRED8x8      ] = ff_pred8x8_tm_vp8_mmx2;
+                h->pred4x4  [TM_VP8_PRED        ] = ff_pred4x4_tm_vp8_mmx2;
                 h->pred4x4  [VERT_PRED          ] = ff_pred4x4_vertical_vp8_mmxext;
             } else {
                 if (chroma_format_idc == 1)

libavcodec/x86/vp8dsp-init.c

@@ -389,11 +389,13 @@ av_cold void ff_vp8dsp_init_x86(VP8DSPContext* c)
 
         c->vp8_v_loop_filter_simple = ff_vp8_v_loop_filter_simple_sse2;
 
+#if ARCH_X86_64 || HAVE_ALIGNED_STACK
         c->vp8_v_loop_filter16y_inner = ff_vp8_v_loop_filter16y_inner_sse2;
         c->vp8_v_loop_filter8uv_inner = ff_vp8_v_loop_filter8uv_inner_sse2;
 
         c->vp8_v_loop_filter16y       = ff_vp8_v_loop_filter16y_mbedge_sse2;
         c->vp8_v_loop_filter8uv       = ff_vp8_v_loop_filter8uv_mbedge_sse2;
+#endif
     }
 
     if (mm_flags & AV_CPU_FLAG_SSE2) {

libavcodec/x86/vp8dsp.asm

@@ -1465,27 +1465,6 @@ VP8_DC_WHT
 %endif
 %endmacro
 
-%macro SPLATB_REG 2-3
-%if cpuflag(ssse3)
-    movd           %1, %2d
-    pshufb         %1, %3
-%elif cpuflag(sse2)
-    movd           %1, %2d
-    punpcklbw      %1, %1
-    pshuflw        %1, %1, 0x0
-    punpcklqdq     %1, %1
-%elif cpuflag(mmx2)
-    movd           %1, %2d
-    punpcklbw      %1, %1
-    pshufw         %1, %1, 0x0
-%else
-    movd           %1, %2d
-    punpcklbw      %1, %1
-    punpcklwd      %1, %1
-    punpckldq      %1, %1
-%endif
-%endmacro
-
 %macro SIMPLE_LOOPFILTER 2
 cglobal vp8_%1_loop_filter_simple, 3, %2, 8, dst, stride, flim, cntr
 %if mmsize == 8 ; mmx/mmxext

libavformat/flvdec.c

@@ -26,6 +26,7 @@
 
 #include "libavutil/avstring.h"
 #include "libavutil/dict.h"
+#include "libavutil/opt.h"
 #include "libavutil/intfloat.h"
 #include "libavutil/mathematics.h"
 #include "libavcodec/bytestream.h"
@@ -38,6 +39,8 @@
 #define VALIDATE_INDEX_TS_THRESH 2500
 
 typedef struct {
+    const AVClass *class; ///< Class for private options.
+    int trust_metadata; ///< configure streams according onMetaData
     int wrong_dts; ///< wrong dts due to negative cts
     uint8_t *new_extradata[FLV_STREAM_TYPE_NB];
     int      new_extradata_size[FLV_STREAM_TYPE_NB];
@@ -327,6 +330,7 @@ finish:
 
 static int amf_parse_object(AVFormatContext *s, AVStream *astream, AVStream *vstream, const char *key, int64_t max_pos, int depth) {
     AVCodecContext *acodec, *vcodec;
+    FLVContext *flv = s->priv_data;
     AVIOContext *ioc;
     AMFDataType amf_type;
     char str_val[256];
@@ -406,6 +410,22 @@ static int amf_parse_object(AVFormatContext *s, AVStream *astream, AVStream *vst
                 if (!st)
                     return AVERROR(ENOMEM);
                 st->codec->codec_id = CODEC_ID_TEXT;
+            } else if (flv->trust_metadata) {
+                if (!strcmp(key, "videocodecid") && vcodec) {
+                    flv_set_video_codec(s, vstream, num_val);
+                } else
+                if (!strcmp(key, "audiocodecid") && acodec) {
+                    flv_set_audio_codec(s, astream, acodec, num_val);
+                } else
+                if (!strcmp(key, "audiosamplerate") && acodec) {
+                    acodec->sample_rate = num_val;
+                } else
+                if (!strcmp(key, "width") && vcodec) {
+                    vcodec->width = num_val;
+                } else
+                if (!strcmp(key, "height") && vcodec) {
+                    vcodec->height = num_val;
+                }
             }
         }
 
@@ -857,6 +877,20 @@ static int flv_read_seek(AVFormatContext *s, int stream_index,
     return avio_seek_time(s->pb, stream_index, ts, flags);
 }
 
+#define OFFSET(x) offsetof(FLVContext, x)
+#define VD AV_OPT_FLAG_VIDEO_PARAM | AV_OPT_FLAG_DECODING_PARAM
+static const AVOption options[] = {
+    { "flv_metadata", "Allocate streams according the onMetaData array",      OFFSET(trust_metadata), AV_OPT_TYPE_INT,    { 0 }, 0, 1, VD},
+    { NULL }
+};
+
+static const AVClass class = {
+    .class_name = "flvdec",
+    .item_name  = av_default_item_name,
+    .option     = options,
+    .version    = LIBAVUTIL_VERSION_INT,
+};
+
 AVInputFormat ff_flv_demuxer = {
     .name           = "flv",
     .long_name      = NULL_IF_CONFIG_SMALL("FLV format"),
@@ -867,4 +901,5 @@ AVInputFormat ff_flv_demuxer = {
     .read_seek      = flv_read_seek,
     .read_close     = flv_read_close,
     .extensions     = "flv",
+    .priv_class     = &class,
 };

libavutil/blowfish.c

@@ -381,8 +381,8 @@ void av_blowfish_crypt(AVBlowfish *ctx, uint8_t *dst, const uint8_t *src,
     uint32_t v0, v1;
     int i;
 
-    while (count > 0) {
-        if (decrypt) {
+    if (decrypt) {
+        while (count--) {
             v0 = AV_RB32(src);
             v1 = AV_RB32(src + 4);
 
@@ -396,7 +396,12 @@ void av_blowfish_crypt(AVBlowfish *ctx, uint8_t *dst, const uint8_t *src,
                     dst[i] = dst[i] ^ iv[i];
                 memcpy(iv, src, 8);
             }
-        } else {
+
+            src   += 8;
+            dst   += 8;
+        }
+    } else {
+        while (count--) {
             if (iv) {
                 for (i = 0; i < 8; i++)
                     dst[i] = src[i] ^ iv[i];
@@ -414,11 +419,10 @@ void av_blowfish_crypt(AVBlowfish *ctx, uint8_t *dst, const uint8_t *src,
 
             if (iv)
                 memcpy(iv, dst, 8);
-        }
 
-        src   += 8;
-        dst   += 8;
-        count -= 8;
+            src   += 8;
+            dst   += 8;
+        }
     }
 }
 

libavutil/x86/x86inc.asm

@@ -604,6 +604,7 @@ SECTION .note.GNU-stack noalloc noexec nowrite progbits
 ; All subsequent functions (up to the next INIT_CPUFLAGS) is built for the specified cpu.
 ; You shouldn't need to invoke this macro directly, it's a subroutine for INIT_MMX &co.
 %macro INIT_CPUFLAGS 0-2
+    CPU amdnop
     %if %0 >= 1
         %xdefine cpuname %1
         %assign cpuflags cpuflags_%1
@@ -625,6 +626,9 @@ SECTION .note.GNU-stack noalloc noexec nowrite progbits
         %elifidn %1, sse3
             %define movu lddqu
         %endif
+        %if notcpuflag(mmx2)
+            CPU basicnop
+        %endif
     %else
         %xdefine SUFFIX
         %undef cpuname

libavutil/x86/x86util.asm

@@ -256,15 +256,26 @@
 %define ABSB ABSB_MMX
 %define ABSB2 ABSB2_MMX
 
-%macro SPLATB_MMX 3
+%macro SPLATB_LOAD 3
+%if cpuflag(ssse3)
+    movd      %1, [%2-3]
+    pshufb    %1, %3
+%else
     movd      %1, [%2-3] ;to avoid crossing a cacheline
     punpcklbw %1, %1
     SPLATW    %1, %1, 3
+%endif
 %endmacro
 
-%macro SPLATB_SSSE3 3
-    movd      %1, [%2-3]
+%macro SPLATB_REG 3
+%if cpuflag(ssse3)
+    movd      %1, %2d
     pshufb    %1, %3
+%else
+    movd      %1, %2d
+    punpcklbw %1, %1
+    SPLATW    %1, %1, 0
+%endif
 %endmacro
 
 %macro PALIGNR_MMX 4-5 ; [dst,] src1, src2, imm, tmp
@@ -296,6 +307,14 @@
 %endif
 %endmacro
 
+%macro PSHUFLW 1+
+    %if mmsize == 8
+        pshufw %1
+    %else
+        pshuflw %1
+    %endif
+%endmacro
+
 %macro DEINTB 5 ; mask, reg1, mask, reg2, optional src to fill masks from
 %ifnum %5
     pand   m%3, m%5, m%4 ; src .. y6 .. y4
@@ -521,8 +540,22 @@
 %if mmsize == 16
     pshuflw    %1, %2, (%3)*0x55
     punpcklqdq %1, %1
-%else
+%elif cpuflag(mmx2)
     pshufw     %1, %2, (%3)*0x55
+%else
+    %ifnidn %1, %2
+        mova       %1, %2
+    %endif
+    %if %3 & 2
+        punpckhwd  %1, %1
+    %else
+        punpcklwd  %1, %1
+    %endif
+    %if %3 & 1
+        punpckhwd  %1, %1
+    %else
+        punpcklwd  %1, %1
+    %endif
 %endif
 %endmacro
 

libavutil/xtea.c

@@ -71,8 +71,8 @@ void av_xtea_crypt(AVXTEA *ctx, uint8_t *dst, const uint8_t *src, int count,
 {
     int i;
 
-    while (count > 0) {
-        if (decrypt) {
+    if (decrypt) {
+        while (count--) {
             xtea_crypt_ecb(ctx, dst, src, decrypt);
 
             if (iv) {
@@ -80,7 +80,12 @@ void av_xtea_crypt(AVXTEA *ctx, uint8_t *dst, const uint8_t *src, int count,
                     dst[i] = dst[i] ^ iv[i];
                 memcpy(iv, src, 8);
             }
-        } else {
+
+            src   += 8;
+            dst   += 8;
+        }
+    } else {
+        while (count--) {
             if (iv) {
                 for (i = 0; i < 8; i++)
                     dst[i] = src[i] ^ iv[i];
@@ -89,11 +94,9 @@ void av_xtea_crypt(AVXTEA *ctx, uint8_t *dst, const uint8_t *src, int count,
             } else {
                 xtea_crypt_ecb(ctx, dst, src, decrypt);
             }
+            src   += 8;
+            dst   += 8;
         }
-
-        src   += 8;
-        dst   += 8;
-        count -= 8;
     }
 }