Merge remote-tracking branch 'qatar/master'

* qatar/master:
  ppc: fix some pointer to integer casts
  ppc: fix 32-bit PIC build
  vmdaudio: fix decoding of 16-bit audio format.
  lavf: do not set codec_tag for rawvideo
  h264: check for out of bounds reads in ff_h264_decode_extradata().
  flvdec: Check for overflow before allocating arrays
  avconv: use correct output stream index when checking max_frames
  avconv: remove fake coded_frame on streamcopy hack

Conflicts:
	avconv.c
	libavcodec/h264.c
	libavcodec/ppc/asm.S
	libavcodec/vmdav.c
	libavformat/flvdec.c
	libavformat/utils.c

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

3 files changed, 99 insertions, 52 deletions

diff --git a/libavcodec/ppc/asm.S b/libavcodec/ppc/asm.S
index 2706d6b1d8..bbbf8a4a66 100644
--- a/libavcodec/ppc/asm.S
+++ b/libavcodec/ppc/asm.S
@@ -44,10 +44,13 @@ X(\name):
 L(\name):
 .endm
 
-.macro movrel rd, sym
+.macro movrel rd, sym, gp
     ld      \rd, \sym@got(r2)
 .endm
 
+.macro get_got rd
+.endm
+
 #else /* ARCH_PPC64 */
 
 #define PTR  .int
@@ -65,19 +68,25 @@ X(\name):
 \name:
 .endm
 
-.macro movrel rd, sym
+.macro movrel rd, sym, gp
 #if CONFIG_PIC
-    bcl             20, 31, lab_pic_\@
-lab_pic_\@:
-    mflr    \rd
-    addis   \rd, \rd, (\sym - lab_pic_\@)@ha
-    addi    \rd, \rd, (\sym - lab_pic_\@)@l
+    lwz     \rd, \sym@got(\gp)
 #else
     lis     \rd, \sym@ha
     la      \rd, \sym@l(\rd)
 #endif
 .endm
 
+.macro get_got rd
+#if CONFIG_PIC
+    bcl     20, 31, .Lgot\@
+.Lgot\@:
+    mflr    \rd
+    addis   \rd, \rd, _GLOBAL_OFFSET_TABLE_ - .Lgot\@@ha
+    addi    \rd, \rd, _GLOBAL_OFFSET_TABLE_ - .Lgot\@@l
+#endif
+.endm
+
 #endif /* ARCH_PPC64 */
 
 #if HAVE_IBM_ASM
diff --git a/libavcodec/ppc/fft_altivec_s.S b/libavcodec/ppc/fft_altivec_s.S
index 5d3c5406c3..16ce838c97 100644
--- a/libavcodec/ppc/fft_altivec_s.S
+++ b/libavcodec/ppc/fft_altivec_s.S
@@ -353,6 +353,7 @@ extfunc ff_fft_calc\interleave\()_altivec
     mflr    r0
     stp     r0, 2*PS(r1)
     stpu    r1, -(160+16*PS)(r1)
+    get_got r11
     addi    r6, r1, 16*PS
     stvm    r6, v20, v21, v22, v23, v24, v25, v26, v27, v28, v29
     mfvrsave r0
@@ -360,14 +361,14 @@ extfunc ff_fft_calc\interleave\()_altivec
     li      r6, 0xfffffffc
     mtvrsave r6
 
-    movrel  r6, fft_data
+    movrel  r6, fft_data, r11
     lvm     r6, v14, v15, v16, v17, v18, v19, v20, v21
     lvm     r6, v22, v23, v24, v25, v26, v27, v28, v29
 
     li      r9, 16
-    movrel  r12, X(ff_cos_tabs)
+    movrel  r12, X(ff_cos_tabs), r11
 
-    movrel  r6, fft_dispatch_tab\interleave\()_altivec
+    movrel  r6, fft_dispatch_tab\interleave\()_altivec, r11
     lwz     r3, 0(r3)
     subi    r3, r3, 2
     slwi    r3, r3, 2+ARCH_PPC64
diff --git a/libavcodec/vmdav.c b/libavcodec/vmdav.c
index 6729af6af7..63ac33db59 100644
--- a/libavcodec/vmdav.c
+++ b/libavcodec/vmdav.c
@@ -465,9 +465,8 @@ static av_cold int vmdvideo_decode_end(AVCodecContext *avctx)
 #define BLOCK_TYPE_SILENCE  3
 
 typedef struct VmdAudioContext {
-    AVCodecContext *avctx;
     int out_bps;
-    int predictors[2];
+    int chunk_size;
 } VmdAudioContext;
 
 static const uint16_t vmdaudio_table[128] = {
@@ -490,13 +489,23 @@ static av_cold int vmdaudio_decode_init(AVCodecContext *avctx)
 {
     VmdAudioContext *s = avctx->priv_data;
 
-    s->avctx = avctx;
+    if (avctx->channels < 1 || avctx->channels > 2) {
+        av_log(avctx, AV_LOG_ERROR, "invalid number of channels\n");
+        return AVERROR(EINVAL);
+    }
+    if (avctx->block_align < 1) {
+        av_log(avctx, AV_LOG_ERROR, "invalid block align\n");
+        return AVERROR(EINVAL);
+    }
+
     if (avctx->bits_per_coded_sample == 16)
         avctx->sample_fmt = AV_SAMPLE_FMT_S16;
     else
         avctx->sample_fmt = AV_SAMPLE_FMT_U8;
     s->out_bps = av_get_bytes_per_sample(avctx->sample_fmt);
 
+    s->chunk_size = avctx->block_align + avctx->channels * (s->out_bps == 2);
+
     av_log(avctx, AV_LOG_DEBUG, "%d channels, %d bits/sample, "
            "block align = %d, sample rate = %d\n",
            avctx->channels, avctx->bits_per_coded_sample, avctx->block_align,
@@ -505,41 +514,33 @@ static av_cold int vmdaudio_decode_init(AVCodecContext *avctx)
     return 0;
 }
 
-static void vmdaudio_decode_audio(VmdAudioContext *s, unsigned char *data,
-    const uint8_t *buf, int buf_size, int stereo)
+static void decode_audio_s16(int16_t *out, const uint8_t *buf, int buf_size,
+                             int channels)
 {
-    int i;
-    int chan = 0;
-    int16_t *out = (int16_t*)data;
-
-    for(i = 0; i < buf_size; i++) {
-        if(buf[i] & 0x80)
-            s->predictors[chan] -= vmdaudio_table[buf[i] & 0x7F];
-        else
-            s->predictors[chan] += vmdaudio_table[buf[i]];
-        s->predictors[chan] = av_clip_int16(s->predictors[chan]);
-        out[i] = s->predictors[chan];
-        chan ^= stereo;
+    int ch;
+    const uint8_t *buf_end = buf + buf_size;
+    int predictor[2];
+    int st = channels - 1;
+
+    /* decode initial raw sample */
+    for (ch = 0; ch < channels; ch++) {
+        predictor[ch] = (int16_t)AV_RL16(buf);
+        buf += 2;
+        *out++ = predictor[ch];
     }
-}
 
-static int vmdaudio_loadsound(VmdAudioContext *s, unsigned char *data,
-    const uint8_t *buf, int silent_chunks, int data_size)
-{
-    int silent_size = s->avctx->block_align * silent_chunks * s->out_bps;
-
-    if (silent_chunks) {
-        memset(data, s->out_bps == 2 ? 0x00 : 0x80, silent_size);
-        data += silent_size;
-    }
-    if (s->avctx->bits_per_coded_sample == 16)
-        vmdaudio_decode_audio(s, data, buf, data_size, s->avctx->channels == 2);
-    else {
-        /* just copy the data */
-        memcpy(data, buf, data_size);
+    /* decode DPCM samples */
+    ch = 0;
+    while (buf < buf_end) {
+        uint8_t b = *buf++;
+        if (b & 0x80)
+            predictor[ch] -= vmdaudio_table[b & 0x7F];
+        else
+            predictor[ch] += vmdaudio_table[b];
+        predictor[ch] = av_clip_int16(predictor[ch]);
+        *out++ = predictor[ch];
+        ch ^= st;
     }
-
-    return silent_size + data_size * s->out_bps;
 }
 
 static int vmdaudio_decode_frame(AVCodecContext *avctx,
@@ -547,10 +548,13 @@ static int vmdaudio_decode_frame(AVCodecContext *avctx,
                                  AVPacket *avpkt)
 {
     const uint8_t *buf = avpkt->data;
+    const uint8_t *buf_end;
     int buf_size = avpkt->size;
     VmdAudioContext *s = avctx->priv_data;
-    int block_type, silent_chunks;
-    unsigned char *output_samples = (unsigned char *)data;
+    int block_type, silent_chunks, audio_chunks;
+    int nb_samples, out_size;
+    uint8_t *output_samples_u8  = data;
+    int16_t *output_samples_s16 = data;
 
     if (buf_size < 16) {
         av_log(avctx, AV_LOG_WARNING, "skipping small junk packet\n");
@@ -566,13 +570,16 @@ static int vmdaudio_decode_frame(AVCodecContext *avctx,
     buf      += 16;
     buf_size -= 16;
 
+    /* get number of silent chunks */
     silent_chunks = 0;
     if (block_type == BLOCK_TYPE_INITIAL) {
         uint32_t flags;
-        if (buf_size < 4)
-            return -1;
-        flags = AV_RB32(buf);
-        silent_chunks  = av_popcount(flags);
+        if (buf_size < 4) {
+            av_log(avctx, AV_LOG_ERROR, "packet is too small\n");
+            return AVERROR(EINVAL);
+        }
+        flags         = AV_RB32(buf);
+        silent_chunks = av_popcount(flags);
         buf      += 4;
         buf_size -= 4;
     } else if (block_type == BLOCK_TYPE_SILENCE) {
@@ -581,11 +588,41 @@ static int vmdaudio_decode_frame(AVCodecContext *avctx,
     }
 
     /* ensure output buffer is large enough */
-    if (*data_size < (avctx->block_align*silent_chunks + buf_size) * s->out_bps)
+    audio_chunks = buf_size / s->chunk_size;
+    nb_samples   = ((silent_chunks + audio_chunks) * avctx->block_align) / avctx->channels;
+    out_size     = nb_samples * avctx->channels * s->out_bps;
+    if (*data_size < out_size)
         return -1;
 
-    *data_size = vmdaudio_loadsound(s, output_samples, buf, silent_chunks, buf_size);
+    /* decode silent chunks */
+    if (silent_chunks > 0) {
+        int silent_size = avctx->block_align * silent_chunks;
+        if (s->out_bps == 2) {
+            memset(output_samples_s16, 0x00, silent_size * 2);
+            output_samples_s16 += silent_size;
+        } else {
+            memset(output_samples_u8,  0x80, silent_size);
+            output_samples_u8 += silent_size;
+        }
+    }
+
+    /* decode audio chunks */
+    if (audio_chunks > 0) {
+        buf_end = buf + buf_size;
+        while (buf < buf_end) {
+            if (s->out_bps == 2) {
+                decode_audio_s16(output_samples_s16, buf, s->chunk_size,
+                                 avctx->channels);
+                output_samples_s16 += avctx->block_align;
+            } else {
+                memcpy(output_samples_u8, buf, s->chunk_size);
+                output_samples_u8  += avctx->block_align;
+            }
+            buf += s->chunk_size;
+        }
+    }
 
+    *data_size = out_size;
     return avpkt->size;
 }