Merge remote-tracking branch 'qatar/master'

* qatar/master:
  avplay: use libavresample for sample format conversion and channel mixing
  Fix compilation with YASM/NASM without AVX support.
  WMAL: do not output last frame again if nothing was decoded in current packet
  WMAL: do not start decoding if frame does not end in current packet
  adpcm-thp: fix invalid array indexing
  ppc: add const where needed in scalarproduct_int16_altivec()
  ppc: remove shift parameter from scalarproduct_int16_altivec()
  ppc: dsputil: do unaligned block accesses correctly
  dvenc: do not call dsputil functions with stride not a multiple of 16
  APIchanges: fill in some dates and commit hashes

Conflicts:
	doc/APIchanges
	ffplay.c
	libavcodec/adpcm.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>

doc/APIchanges

@@ -2,14 +2,14 @@ Never assume the API of libav* to be stable unless at least 1 month has passed
 since the last major version increase.
 
 The last version increases were:
-libavcodec:  2012-01-27
-libavdevice: 2011-04-18
-libavfilter: 2011-04-18
-libavformat: 2012-01-27
-libavresample: 2012-xx-xx
-libpostproc: 2011-04-18
-libswscale:  2011-06-20
-libavutil:   2011-04-18
+libavcodec:    2012-01-27
+libavdevice:   2011-04-18
+libavfilter:   2011-04-18
+libavformat:   2012-01-27
+libavresample: 2012-04-24
+libpostproc:   2011-04-18
+libswscale:    2011-06-20
+libavutil:     2011-04-18
 
 
 API changes, most recent first:
@@ -23,13 +23,13 @@ API changes, most recent first:
 2012-03-26 - a67d9cf - lavfi 2.66.100
   Add avfilter_fill_frame_from_{audio_,}buffer_ref() functions.
 
-2012-04-25 - xxxxxxx - lavu 51.29.0 - cpu.h
+2012-04-25 - 3527a73 - lavu 51.29.0 - cpu.h
   Add av_parse_cpu_flags()
 
-2012-xx-xx - xxxxxxx - lavr 0.0.0
+2012-04-24 - c8af852 - lavr 0.0.0
   Add libavresample audio conversion library
 
-2012-xx-xx - xxxxxxx - lavu 51.28.0 - audio_fifo.h
+2012-04-20 - 0c0d1bc - lavu 51.28.0 - audio_fifo.h
   Add audio FIFO functions:
     av_audio_fifo_free()
     av_audio_fifo_alloc()
@@ -41,10 +41,10 @@ API changes, most recent first:
     av_audio_fifo_size()
     av_audio_fifo_space()
 
-2012-xx-xx - xxxxxxx - lavfi 2.16.0 - avfiltergraph.h
-  Add avfilter_graph_parse2()
+2012-04-14 - lavfi 2.16.0 - avfiltergraph.h
+  d7bcc71 Add avfilter_graph_parse2().
 
-2012-xx-xx - xxxxxxx - lavu 51.27.0 - samplefmt.h
+2012-04-08 - 4d693b0 - lavu 51.27.0 - samplefmt.h
   Add av_get_packed_sample_fmt() and av_get_planar_sample_fmt()
 
 2012-03-21 - b75c67d - lavu 51.43.100

ffplay.c

@@ -40,7 +40,6 @@
 #include "libavformat/avformat.h"
 #include "libavdevice/avdevice.h"
 #include "libswscale/swscale.h"
-#include "libavcodec/audioconvert.h"
 #include "libavutil/opt.h"
 #include "libavcodec/avfft.h"
 #include "libswresample/swresample.h"
@@ -2130,7 +2129,8 @@ static int audio_decode_frame(VideoState *is, double *pts_ptr)
             /* if no pts, then compute it */
             pts = is->audio_clock;
             *pts_ptr = pts;
-            is->audio_clock += (double)data_size / (dec->channels * dec->sample_rate * av_get_bytes_per_sample(dec->sample_fmt));
+            is->audio_clock += (double)data_size /
+                (dec->channels * dec->sample_rate * av_get_bytes_per_sample(dec->sample_fmt));
 #ifdef DEBUG
             {
                 static double last_clock;
@@ -2373,9 +2373,9 @@ static void stream_component_close(VideoState *is, int stream_index)
         SDL_CloseAudio();
 
         packet_queue_end(&is->audioq);
+        av_free_packet(&is->audio_pkt);
         if (is->swr_ctx)
             swr_free(&is->swr_ctx);
-        av_free_packet(&is->audio_pkt);
         av_freep(&is->audio_buf1);
         is->audio_buf = NULL;
         av_freep(&is->frame);

libavcodec/adpcm.c

@@ -1212,12 +1212,14 @@ static int adpcm_decode_frame(AVCodecContext *avctx, void *data,
         int prev[2][2];
         int ch;
 
-        for (i = 0; i < 32; i++)
-            table[0][i] = sign_extend(bytestream2_get_be16u(&gb), 16);
+        for (i = 0; i < 2; i++)
+            for (n = 0; n < 16; n++)
+                table[i][n] = sign_extend(bytestream2_get_be16u(&gb), 16);
 
         /* Initialize the previous sample.  */
-        for (i = 0; i < 4; i++)
-            prev[i>>1][i&1] = sign_extend(bytestream2_get_be16u(&gb), 16);
+        for (i = 0; i < 2; i++)
+            for (n = 0; n < 2; n++)
+                prev[i][n] = sign_extend(bytestream2_get_be16u(&gb), 16);
 
         for (ch = 0; ch <= st; ch++) {
             samples = (short *)c->frame.data[0] + ch;

libavcodec/dv.c

@@ -668,7 +668,7 @@ static int dv_encode_video_segment(AVCodecContext *avctx, void *arg)
     int mb_x, mb_y, c_offset, linesize, y_stride;
     uint8_t*  y_ptr;
     uint8_t*  dif;
-    LOCAL_ALIGNED_8(uint8_t, scratch, [64]);
+    LOCAL_ALIGNED_8(uint8_t, scratch, [128]);
     EncBlockInfo  enc_blks[5*DV_MAX_BPM];
     PutBitContext pbs[5*DV_MAX_BPM];
     PutBitContext* pb;
@@ -723,10 +723,10 @@ static int dv_encode_video_segment(AVCodecContext *avctx, void *arg)
                     b[0] = c_ptr[0]; b[1] = c_ptr[1]; b[2] = c_ptr[2]; b[3] = c_ptr[3];
                     b[4] =     d[0]; b[5] =     d[1]; b[6] =     d[2]; b[7] =     d[3];
                     c_ptr += linesize;
-                    b += 8;
+                    b += 16;
                 }
                 c_ptr = scratch;
-                linesize = 8;
+                linesize = 16;
             }
 
             vs_bit_size += dv_init_enc_block(    enc_blk++, c_ptr           , linesize, s, 1);

libavcodec/ppc/int_altivec.c

@@ -79,28 +79,20 @@ static int ssd_int8_vs_int16_altivec(const int8_t *pix1, const int16_t *pix2,
     return u.score[3];
 }
 
-static int32_t scalarproduct_int16_altivec(int16_t *v1, const int16_t *v2,
-                                           int order, const int shift)
+static int32_t scalarproduct_int16_altivec(const int16_t *v1, const int16_t *v2,
+                                           int order)
 {
     int i;
     LOAD_ZERO;
-    register vec_s16 vec1, *pv;
+    const vec_s16 *pv;
+    register vec_s16 vec1;
     register vec_s32 res = vec_splat_s32(0), t;
-    register vec_u32 shifts;
     int32_t ires;
 
-    shifts = zero_u32v;
-    if(shift & 0x10) shifts = vec_add(shifts, vec_sl(vec_splat_u32(0x08), vec_splat_u32(0x1)));
-    if(shift & 0x08) shifts = vec_add(shifts, vec_splat_u32(0x08));
-    if(shift & 0x04) shifts = vec_add(shifts, vec_splat_u32(0x04));
-    if(shift & 0x02) shifts = vec_add(shifts, vec_splat_u32(0x02));
-    if(shift & 0x01) shifts = vec_add(shifts, vec_splat_u32(0x01));
-
     for(i = 0; i < order; i += 8){
-        pv = (vec_s16*)v1;
+        pv = (const vec_s16*)v1;
         vec1 = vec_perm(pv[0], pv[1], vec_lvsl(0, v1));
         t = vec_msum(vec1, vec_ld(0, v2), zero_s32v);
-        t = vec_sr(t, shifts);
         res = vec_sums(t, res);
         v1 += 8;
         v2 += 8;
@@ -114,31 +106,31 @@ static int32_t scalarproduct_and_madd_int16_altivec(int16_t *v1, const int16_t *
 {
     LOAD_ZERO;
     vec_s16 *pv1 = (vec_s16*)v1;
-    vec_s16 *pv2 = (vec_s16*)v2;
-    vec_s16 *pv3 = (vec_s16*)v3;
     register vec_s16 muls = {mul,mul,mul,mul,mul,mul,mul,mul};
-    register vec_s16 t0, t1, i0, i1;
-    register vec_s16 i2 = pv2[0], i3 = pv3[0];
+    register vec_s16 t0, t1, i0, i1, i4;
+    register vec_s16 i2 = vec_ld(0, v2), i3 = vec_ld(0, v3);
     register vec_s32 res = zero_s32v;
     register vec_u8 align = vec_lvsl(0, v2);
     int32_t ires;
     order >>= 4;
     do {
-        t0 = vec_perm(i2, pv2[1], align);
-        i2 = pv2[2];
-        t1 = vec_perm(pv2[1], i2, align);
+        i1 = vec_ld(16, v2);
+        t0 = vec_perm(i2, i1, align);
+        i2 = vec_ld(32, v2);
+        t1 = vec_perm(i1, i2, align);
         i0 = pv1[0];
         i1 = pv1[1];
         res = vec_msum(t0, i0, res);
         res = vec_msum(t1, i1, res);
-        t0 = vec_perm(i3, pv3[1], align);
-        i3 = pv3[2];
-        t1 = vec_perm(pv3[1], i3, align);
+        i4 = vec_ld(16, v3);
+        t0 = vec_perm(i3, i4, align);
+        i3 = vec_ld(32, v3);
+        t1 = vec_perm(i4, i3, align);
         pv1[0] = vec_mladd(t0, muls, i0);
         pv1[1] = vec_mladd(t1, muls, i1);
         pv1 += 2;
-        pv2 += 2;
-        pv3 += 2;
+        v2  += 8;
+        v3  += 8;
     } while(--order);
     res = vec_splat(vec_sums(res, zero_s32v), 3);
     vec_ste(res, 0, &ires);

libavcodec/wmalosslessdec.c

@@ -1216,8 +1216,8 @@ static int decode_packet(AVCodecContext *avctx, void *data, int *got_frame_ptr,
             save_bits(s, gb, num_bits_prev_frame, 1);
 
             /* decode the cross packet frame if it is valid */
-            if (!s->packet_loss)
-                decode_frame(s);
+            if (num_bits_prev_frame < remaining_packet_bits && !s->packet_loss)
+                    decode_frame(s);
         } else if (s->num_saved_bits - s->frame_offset) {
             av_dlog(avctx, "ignoring %x previously saved bits\n",
                     s->num_saved_bits - s->frame_offset);