Merge remote-tracking branch 'qatar/master'

* qatar/master:
  lavr: add x86-optimized functions for mixing 1-to-2 s16p with flt coeffs
  lavr: add x86-optimized functions for mixing 1-to-2 fltp with flt coeffs
  Add Dolby/DPLII downmix support to libavresample
  vorbisdec: replace div/mod in loop with a counter
  fate: vorbis: add 5.1 surround test
  rtpenc: Allow requesting H264 RTP packetization mode 0
  configure: Sort the library listings in the help text alphabetically
  dwt: remove variable-length arrays
  RTMPT protocol support
  http: Properly handle chunked transfer-encoding for replies to post data
  http: Fail reading if the connection has gone away
  amr: Mark an array const
  amr: More space cleanup
  rtpenc: Fix memory leaks in the muxer open function

Conflicts:
	Changelog
	configure
	doc/APIchanges
	libavformat/version.h

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

8 files changed, 91 insertions, 79 deletions

diff --git a/libavcodec/dwt.c b/libavcodec/dwt.c
index 6c1f1310a3..3980f065b8 100644
--- a/libavcodec/dwt.c
+++ b/libavcodec/dwt.c
@@ -245,9 +245,8 @@ static av_always_inline void inv_liftS(IDWTELEM *dst, IDWTELEM *src,
 }
 #endif /* ! liftS */
 
-static void horizontal_decompose53i(DWTELEM *b, int width)
+static void horizontal_decompose53i(DWTELEM *b, DWTELEM *temp, int width)
 {
-    DWTELEM temp[width];
     const int width2 = width >> 1;
     int x;
     const int w2 = (width + 1) >> 1;
@@ -313,8 +312,8 @@ static void vertical_decompose53iL0(DWTELEM *b0, DWTELEM *b1, DWTELEM *b2,
         b1[i] += (b0[i] + b2[i] + 2) >> 2;
 }
 
-static void spatial_decompose53i(DWTELEM *buffer, int width, int height,
-                                 int stride)
+static void spatial_decompose53i(DWTELEM *buffer, DWTELEM *temp,
+                                 int width, int height, int stride)
 {
     int y;
     DWTELEM *b0 = buffer + mirror(-2 - 1, height - 1) * stride;
@@ -325,9 +324,9 @@ static void spatial_decompose53i(DWTELEM *buffer, int width, int height,
         DWTELEM *b3 = buffer + mirror(y + 2, height - 1) * stride;
 
         if (y + 1 < (unsigned)height)
-            horizontal_decompose53i(b2, width);
+            horizontal_decompose53i(b2, temp, width);
         if (y + 2 < (unsigned)height)
-            horizontal_decompose53i(b3, width);
+            horizontal_decompose53i(b3, temp, width);
 
         if (y + 1 < (unsigned)height)
             vertical_decompose53iH0(b1, b2, b3, width);
@@ -339,9 +338,8 @@ static void spatial_decompose53i(DWTELEM *buffer, int width, int height,
     }
 }
 
-static void horizontal_decompose97i(DWTELEM *b, int width)
+static void horizontal_decompose97i(DWTELEM *b, DWTELEM *temp, int width)
 {
-    DWTELEM temp[width];
     const int w2 = (width + 1) >> 1;
 
     lift(temp + w2, b + 1, b,         1, 2, 2, width, W_AM, W_AO, W_AS, 1, 1);
@@ -391,8 +389,8 @@ static void vertical_decompose97iL1(DWTELEM *b0, DWTELEM *b1, DWTELEM *b2,
         b1[i] += (W_DM * (b0[i] + b2[i]) + W_DO) >> W_DS;
 }
 
-static void spatial_decompose97i(DWTELEM *buffer, int width, int height,
-                                 int stride)
+static void spatial_decompose97i(DWTELEM *buffer, DWTELEM *temp,
+                                 int width, int height, int stride)
 {
     int y;
     DWTELEM *b0 = buffer + mirror(-4 - 1, height - 1) * stride;
@@ -405,9 +403,9 @@ static void spatial_decompose97i(DWTELEM *buffer, int width, int height,
         DWTELEM *b5 = buffer + mirror(y + 4, height - 1) * stride;
 
         if (y + 3 < (unsigned)height)
-            horizontal_decompose97i(b4, width);
+            horizontal_decompose97i(b4, temp, width);
         if (y + 4 < (unsigned)height)
-            horizontal_decompose97i(b5, width);
+            horizontal_decompose97i(b5, temp, width);
 
         if (y + 3 < (unsigned)height)
             vertical_decompose97iH0(b3, b4, b5, width);
@@ -425,20 +423,20 @@ static void spatial_decompose97i(DWTELEM *buffer, int width, int height,
     }
 }
 
-void ff_spatial_dwt(DWTELEM *buffer, int width, int height, int stride,
-                    int type, int decomposition_count)
+void ff_spatial_dwt(DWTELEM *buffer, DWTELEM *temp, int width, int height,
+                    int stride, int type, int decomposition_count)
 {
     int level;
 
     for (level = 0; level < decomposition_count; level++) {
         switch (type) {
         case DWT_97:
-            spatial_decompose97i(buffer,
+            spatial_decompose97i(buffer, temp,
                                  width >> level, height >> level,
                                  stride << level);
             break;
         case DWT_53:
-            spatial_decompose53i(buffer,
+            spatial_decompose53i(buffer, temp,
                                  width >> level, height >> level,
                                  stride << level);
             break;
@@ -446,9 +444,8 @@ void ff_spatial_dwt(DWTELEM *buffer, int width, int height, int stride,
     }
 }
 
-static void horizontal_compose53i(IDWTELEM *b, int width)
+static void horizontal_compose53i(IDWTELEM *b, IDWTELEM *temp, int width)
 {
-    IDWTELEM temp[width];
     const int width2 = width >> 1;
     const int w2     = (width + 1) >> 1;
     int x;
@@ -508,6 +505,7 @@ static void spatial_compose53i_init(DWTCompose *cs, IDWTELEM *buffer,
 }
 
 static void spatial_compose53i_dy_buffered(DWTCompose *cs, slice_buffer *sb,
+                                           IDWTELEM *temp,
                                            int width, int height,
                                            int stride_line)
 {
@@ -537,17 +535,18 @@ static void spatial_compose53i_dy_buffered(DWTCompose *cs, slice_buffer *sb,
     }
 
     if (y - 1 < (unsigned)height)
-        horizontal_compose53i(b0, width);
+        horizontal_compose53i(b0, temp, width);
     if (y + 0 < (unsigned)height)
-        horizontal_compose53i(b1, width);
+        horizontal_compose53i(b1, temp, width);
 
     cs->b0  = b2;
     cs->b1  = b3;
     cs->y  += 2;
 }
 
-static void spatial_compose53i_dy(DWTCompose *cs, IDWTELEM *buffer, int width,
-                                  int height, int stride)
+static void spatial_compose53i_dy(DWTCompose *cs, IDWTELEM *buffer,
+                                  IDWTELEM *temp, int width, int height,
+                                  int stride)
 {
     int y        = cs->y;
     IDWTELEM *b0 = cs->b0;
@@ -561,27 +560,26 @@ static void spatial_compose53i_dy(DWTCompose *cs, IDWTELEM *buffer, int width,
         vertical_compose53iH0(b0, b1, b2, width);
 
     if (y - 1 < (unsigned)height)
-        horizontal_compose53i(b0, width);
+        horizontal_compose53i(b0, temp, width);
     if (y + 0 < (unsigned)height)
-        horizontal_compose53i(b1, width);
+        horizontal_compose53i(b1, temp, width);
 
     cs->b0  = b2;
     cs->b1  = b3;
     cs->y  += 2;
 }
 
-static void av_unused spatial_compose53i(IDWTELEM *buffer, int width,
-                                         int height, int stride)
+static void av_unused spatial_compose53i(IDWTELEM *buffer, IDWTELEM *temp,
+                                         int width, int height, int stride)
 {
     DWTCompose cs;
     spatial_compose53i_init(&cs, buffer, height, stride);
     while (cs.y <= height)
-        spatial_compose53i_dy(&cs, buffer, width, height, stride);
+        spatial_compose53i_dy(&cs, buffer, temp, width, height, stride);
 }
 
-void ff_snow_horizontal_compose97i(IDWTELEM *b, int width)
+void ff_snow_horizontal_compose97i(IDWTELEM *b, IDWTELEM *temp, int width)
 {
-    IDWTELEM temp[width];
     const int w2 = (width + 1) >> 1;
 
 #if 0 //maybe more understadable but slower
@@ -695,8 +693,9 @@ static void spatial_compose97i_init(DWTCompose *cs, IDWTELEM *buffer, int height
 }
 
 static void spatial_compose97i_dy_buffered(DWTContext *dsp, DWTCompose *cs,
-                                           slice_buffer *sb, int width,
-                                           int height, int stride_line)
+                                           slice_buffer * sb, IDWTELEM *temp,
+                                           int width, int height,
+                                           int stride_line)
 {
     int y = cs->y;
 
@@ -725,9 +724,9 @@ static void spatial_compose97i_dy_buffered(DWTContext *dsp, DWTCompose *cs,
     }
 
     if (y - 1 < (unsigned)height)
-        dsp->horizontal_compose97i(b0, width);
+        dsp->horizontal_compose97i(b0, temp, width);
     if (y + 0 < (unsigned)height)
-        dsp->horizontal_compose97i(b1, width);
+        dsp->horizontal_compose97i(b1, temp, width);
 
     cs->b0  = b2;
     cs->b1  = b3;
@@ -736,8 +735,9 @@ static void spatial_compose97i_dy_buffered(DWTContext *dsp, DWTCompose *cs,
     cs->y  += 2;
 }
 
-static void spatial_compose97i_dy(DWTCompose *cs, IDWTELEM *buffer, int width,
-                                  int height, int stride)
+static void spatial_compose97i_dy(DWTCompose *cs, IDWTELEM *buffer,
+                                  IDWTELEM *temp, int width, int height,
+                                  int stride)
 {
     int y        = cs->y;
     IDWTELEM *b0 = cs->b0;
@@ -757,9 +757,9 @@ static void spatial_compose97i_dy(DWTCompose *cs, IDWTELEM *buffer, int width,
         vertical_compose97iH0(b0, b1, b2, width);
 
     if (y - 1 < (unsigned)height)
-        ff_snow_horizontal_compose97i(b0, width);
+        ff_snow_horizontal_compose97i(b0, temp, width);
     if (y + 0 < (unsigned)height)
-        ff_snow_horizontal_compose97i(b1, width);
+        ff_snow_horizontal_compose97i(b1, temp, width);
 
     cs->b0  = b2;
     cs->b1  = b3;
@@ -768,13 +768,13 @@ static void spatial_compose97i_dy(DWTCompose *cs, IDWTELEM *buffer, int width,
     cs->y  += 2;
 }
 
-static void av_unused spatial_compose97i(IDWTELEM *buffer, int width,
-                                         int height, int stride)
+static void av_unused spatial_compose97i(IDWTELEM *buffer, IDWTELEM *temp,
+                                         int width, int height, int stride)
 {
     DWTCompose cs;
     spatial_compose97i_init(&cs, buffer, height, stride);
     while (cs.y <= height)
-        spatial_compose97i_dy(&cs, buffer, width, height, stride);
+        spatial_compose97i_dy(&cs, buffer, temp, width, height, stride);
 }
 
 void ff_spatial_idwt_buffered_init(DWTCompose *cs, slice_buffer *sb, int width,
@@ -797,9 +797,9 @@ void ff_spatial_idwt_buffered_init(DWTCompose *cs, slice_buffer *sb, int width,
 }
 
 void ff_spatial_idwt_buffered_slice(DWTContext *dsp, DWTCompose *cs,
-                                    slice_buffer *slice_buf, int width,
-                                    int height, int stride_line, int type,
-                                    int decomposition_count, int y)
+                                    slice_buffer *slice_buf, IDWTELEM *temp,
+                                    int width, int height, int stride_line,
+                                    int type, int decomposition_count, int y)
 {
     const int support = type == 1 ? 3 : 5;
     int level;
@@ -810,13 +810,13 @@ void ff_spatial_idwt_buffered_slice(DWTContext *dsp, DWTCompose *cs,
         while (cs[level].y <= FFMIN((y >> level) + support, height >> level)) {
             switch (type) {
             case DWT_97:
-                spatial_compose97i_dy_buffered(dsp, cs + level, slice_buf,
+                spatial_compose97i_dy_buffered(dsp, cs + level, slice_buf, temp,
                                                width >> level,
                                                height >> level,
                                                stride_line << level);
                 break;
             case DWT_53:
-                spatial_compose53i_dy_buffered(cs + level, slice_buf,
+                spatial_compose53i_dy_buffered(cs + level, slice_buf, temp,
                                                width >> level,
                                                height >> level,
                                                stride_line << level);
@@ -844,8 +844,9 @@ static void ff_spatial_idwt_init(DWTCompose *cs, IDWTELEM *buffer, int width,
     }
 }
 
-static void ff_spatial_idwt_slice(DWTCompose *cs, IDWTELEM *buffer, int width,
-                                  int height, int stride, int type,
+static void ff_spatial_idwt_slice(DWTCompose *cs, IDWTELEM *buffer,
+                                  IDWTELEM *temp, int width, int height,
+                                  int stride, int type,
                                   int decomposition_count, int y)
 {
     const int support = type == 1 ? 3 : 5;
@@ -857,26 +858,26 @@ static void ff_spatial_idwt_slice(DWTCompose *cs, IDWTELEM *buffer, int width,
         while (cs[level].y <= FFMIN((y >> level) + support, height >> level)) {
             switch (type) {
             case DWT_97:
-                spatial_compose97i_dy(cs + level, buffer, width >> level,
+                spatial_compose97i_dy(cs + level, buffer, temp, width >> level,
                                       height >> level, stride << level);
                 break;
             case DWT_53:
-                spatial_compose53i_dy(cs + level, buffer, width >> level,
+                spatial_compose53i_dy(cs + level, buffer, temp, width >> level,
                                       height >> level, stride << level);
                 break;
             }
         }
 }
 
-void ff_spatial_idwt(IDWTELEM *buffer, int width, int height, int stride,
-                     int type, int decomposition_count)
+void ff_spatial_idwt(IDWTELEM *buffer, IDWTELEM *temp, int width, int height,
+                     int stride, int type, int decomposition_count)
 {
     DWTCompose cs[MAX_DECOMPOSITIONS];
     int y;
     ff_spatial_idwt_init(cs, buffer, width, height, stride, type,
                          decomposition_count);
     for (y = 0; y < height; y += 4)
-        ff_spatial_idwt_slice(cs, buffer, width, height, stride, type,
+        ff_spatial_idwt_slice(cs, buffer, temp, width, height, stride, type,
                               decomposition_count, y);
 }
 
@@ -885,7 +886,7 @@ static inline int w_c(void *v, uint8_t *pix1, uint8_t *pix2, int line_size,
 {
     int s, i, j;
     const int dec_count = w == 8 ? 3 : 4;
-    int tmp[32 * 32];
+    int tmp[32 * 32], tmp2[32];
     int level, ori;
     static const int scale[2][2][4][4] = {
         {
@@ -927,7 +928,7 @@ static inline int w_c(void *v, uint8_t *pix1, uint8_t *pix2, int line_size,
         pix2 += line_size;
     }
 
-    ff_spatial_dwt(tmp, w, h, 32, type, dec_count);
+    ff_spatial_dwt(tmp, tmp2, w, h, 32, type, dec_count);
 
     s = 0;
     assert(w == h);
diff --git a/libavcodec/dwt.h b/libavcodec/dwt.h
index aa7d98dc28..53e1aaab06 100644
--- a/libavcodec/dwt.h
+++ b/libavcodec/dwt.h
@@ -79,7 +79,7 @@ typedef struct DWTContext {
     void (*vertical_compose97i)(IDWTELEM *b0, IDWTELEM *b1, IDWTELEM *b2,
                                 IDWTELEM *b3, IDWTELEM *b4, IDWTELEM *b5,
                                 int width);
-    void (*horizontal_compose97i)(IDWTELEM *b, int width);
+    void (*horizontal_compose97i)(IDWTELEM *b, IDWTELEM *temp, int width);
     void (*inner_add_yblock)(const uint8_t *obmc, const int obmc_stride,
                              uint8_t **block, int b_w, int b_h, int src_x,
                              int src_y, int src_stride, slice_buffer *sb,
@@ -239,7 +239,7 @@ IDWTELEM *ff_slice_buffer_load_line(slice_buffer *buf, int line);
 void ff_snow_vertical_compose97i(IDWTELEM *b0, IDWTELEM *b1, IDWTELEM *b2,
                                  IDWTELEM *b3, IDWTELEM *b4, IDWTELEM *b5,
                                  int width);
-void ff_snow_horizontal_compose97i(IDWTELEM *b, int width);
+void ff_snow_horizontal_compose97i(IDWTELEM *b, IDWTELEM *temp, int width);
 void ff_snow_inner_add_yblock(const uint8_t *obmc, const int obmc_stride,
                               uint8_t **block, int b_w, int b_h, int src_x,
                               int src_y, int src_stride, slice_buffer *sb,
@@ -248,18 +248,18 @@ void ff_snow_inner_add_yblock(const uint8_t *obmc, const int obmc_stride,
 int ff_w53_32_c(void *v, uint8_t *pix1, uint8_t *pix2, int line_size, int h);
 int ff_w97_32_c(void *v, uint8_t *pix1, uint8_t *pix2, int line_size, int h);
 
-void ff_spatial_dwt(int *buffer, int width, int height, int stride, int type,
-                    int decomposition_count);
+void ff_spatial_dwt(int *buffer, int *temp, int width, int height, int stride,
+                    int type, int decomposition_count);
 
 void ff_spatial_idwt_buffered_init(DWTCompose *cs, slice_buffer *sb, int width,
                                    int height, int stride_line, int type,
                                    int decomposition_count);
 void ff_spatial_idwt_buffered_slice(DWTContext *dsp, DWTCompose *cs,
-                                    slice_buffer *slice_buf, int width,
-                                    int height, int stride_line, int type,
-                                    int decomposition_count, int y);
-void ff_spatial_idwt(IDWTELEM *buffer, int width, int height, int stride,
-                     int type, int decomposition_count);
+                                    slice_buffer *slice_buf, IDWTELEM *temp,
+                                    int width, int height, int stride_line,
+                                    int type, int decomposition_count, int y);
+void ff_spatial_idwt(IDWTELEM *buffer, IDWTELEM *temp, int width, int height,
+                     int stride, int type, int decomposition_count);
 
 void ff_dwt_init(DWTContext *c);
 void ff_dwt_init_x86(DWTContext *c);
diff --git a/libavcodec/snow.c b/libavcodec/snow.c
index cd0d9feddd..186bc8bd4d 100644
--- a/libavcodec/snow.c
+++ b/libavcodec/snow.c
@@ -440,6 +440,8 @@ av_cold int ff_snow_common_init(AVCodecContext *avctx){
 
     s->spatial_idwt_buffer= av_mallocz(width*height*sizeof(IDWTELEM));
     s->spatial_dwt_buffer= av_mallocz(width*height*sizeof(DWTELEM)); //FIXME this does not belong here
+    s->temp_dwt_buffer = av_mallocz(width * sizeof(DWTELEM));
+    s->temp_idwt_buffer = av_mallocz(width * sizeof(IDWTELEM));
 
     for(i=0; i<MAX_REF_FRAMES; i++)
         for(j=0; j<MAX_REF_FRAMES; j++)
@@ -618,7 +620,9 @@ av_cold void ff_snow_common_end(SnowContext *s)
     int plane_index, level, orientation, i;
 
     av_freep(&s->spatial_dwt_buffer);
+    av_freep(&s->temp_dwt_buffer);
     av_freep(&s->spatial_idwt_buffer);
+    av_freep(&s->temp_idwt_buffer);
 
     s->m.me.temp= NULL;
     av_freep(&s->m.me.scratchpad);
diff --git a/libavcodec/snow.h b/libavcodec/snow.h
index 32f116d34d..c3280154dd 100644
--- a/libavcodec/snow.h
+++ b/libavcodec/snow.h
@@ -132,7 +132,9 @@ typedef struct SnowContext{
     int16_t (*ref_mvs[MAX_REF_FRAMES])[2];
     uint32_t *ref_scores[MAX_REF_FRAMES];
     DWTELEM *spatial_dwt_buffer;
+    DWTELEM *temp_dwt_buffer;
     IDWTELEM *spatial_idwt_buffer;
+    IDWTELEM *temp_idwt_buffer;
     int colorspace_type;
     int chroma_h_shift;
     int chroma_v_shift;
diff --git a/libavcodec/snowdec.c b/libavcodec/snowdec.c
index f622252c69..a4d8d4e0d4 100644
--- a/libavcodec/snowdec.c
+++ b/libavcodec/snowdec.c
@@ -502,7 +502,7 @@ static int decode_frame(AVCodecContext *avctx, void *data, int *data_size, AVPac
             }
 
             for(; yd<slice_h; yd+=4){
-                ff_spatial_idwt_buffered_slice(&s->dwt, cs, &s->sb, w, h, 1, s->spatial_decomposition_type, s->spatial_decomposition_count, yd);
+                ff_spatial_idwt_buffered_slice(&s->dwt, cs, &s->sb, s->temp_idwt_buffer, w, h, 1, s->spatial_decomposition_type, s->spatial_decomposition_count, yd);
             }
 
             if(s->qlog == LOSSLESS_QLOG){
diff --git a/libavcodec/snowenc.c b/libavcodec/snowenc.c
index 70919603a6..c853a34c6c 100644
--- a/libavcodec/snowenc.c
+++ b/libavcodec/snowenc.c
@@ -94,7 +94,7 @@ static void dwt_quantize(SnowContext *s, Plane *p, DWTELEM *buffer, int width, i
     //FIXME pass the copy cleanly ?
 
 //    memcpy(dwt_buffer, buffer, height * stride * sizeof(DWTELEM));
-    ff_spatial_dwt(buffer, width, height, stride, type, s->spatial_decomposition_count);
+    ff_spatial_dwt(buffer, s->temp_dwt_buffer, width, height, stride, type, s->spatial_decomposition_count);
 
     for(level=0; level<s->spatial_decomposition_count; level++){
         for(orientation=level ? 1 : 0; orientation<4; orientation++){
@@ -119,7 +119,7 @@ static void dwt_quantize(SnowContext *s, Plane *p, DWTELEM *buffer, int width, i
                     for(xs= 0; xs<Q2_STEP; xs++){
                         memcpy(idwt2_buffer, best_dequant, height * stride * sizeof(IDWTELEM));
                         dequantize_all(s, p, idwt2_buffer, width, height);
-                        ff_spatial_idwt(idwt2_buffer, width, height, stride, type, s->spatial_decomposition_count);
+                        ff_spatial_idwt(idwt2_buffer, s->temp_idwt_buffer, width, height, stride, type, s->spatial_decomposition_count);
                         find_sse(s, p, best_score, score_stride, idwt2_buffer, s->spatial_idwt_buffer, level, orientation);
                         memcpy(idwt2_buffer, best_dequant, height * stride * sizeof(IDWTELEM));
                         for(y=ys; y<b->height; y+= Q2_STEP){
@@ -130,7 +130,7 @@ static void dwt_quantize(SnowContext *s, Plane *p, DWTELEM *buffer, int width, i
                             }
                         }
                         dequantize_all(s, p, idwt2_buffer, width, height);
-                        ff_spatial_idwt(idwt2_buffer, width, height, stride, type, s->spatial_decomposition_count);
+                        ff_spatial_idwt(idwt2_buffer, s->temp_idwt_buffer, width, height, stride, type, s->spatial_decomposition_count);
                         find_sse(s, p, score, score_stride, idwt2_buffer, s->spatial_idwt_buffer, level, orientation);
                         for(y=ys; y<b->height; y+= Q2_STEP){
                             for(x=xs; x<b->width; x+= Q2_STEP){
@@ -1588,7 +1588,7 @@ static void calculate_visual_weight(SnowContext *s, Plane *p){
 
             memset(s->spatial_idwt_buffer, 0, sizeof(*s->spatial_idwt_buffer)*width*height);
             ibuf[b->width/2 + b->height/2*b->stride]= 256*16;
-            ff_spatial_idwt(s->spatial_idwt_buffer, width, height, width, s->spatial_decomposition_type, s->spatial_decomposition_count);
+            ff_spatial_idwt(s->spatial_idwt_buffer, s->temp_idwt_buffer, width, height, width, s->spatial_decomposition_type, s->spatial_decomposition_count);
             for(y=0; y<height; y++){
                 for(x=0; x<width; x++){
                     int64_t d= s->spatial_idwt_buffer[x + y*width]*16;
@@ -1778,7 +1778,7 @@ redo_frame:
             /*  if(QUANTIZE2)
                 dwt_quantize(s, p, s->spatial_dwt_buffer, w, h, w, s->spatial_decomposition_type);
             else*/
-                ff_spatial_dwt(s->spatial_dwt_buffer, w, h, w, s->spatial_decomposition_type, s->spatial_decomposition_count);
+                ff_spatial_dwt(s->spatial_dwt_buffer, s->temp_dwt_buffer, w, h, w, s->spatial_decomposition_type, s->spatial_decomposition_count);
 
             if(s->pass1_rc && plane_index==0){
                 int delta_qlog = ratecontrol_1pass(s, pic);
@@ -1818,7 +1818,7 @@ redo_frame:
                 }
             }
 
-            ff_spatial_idwt(s->spatial_idwt_buffer, w, h, w, s->spatial_decomposition_type, s->spatial_decomposition_count);
+            ff_spatial_idwt(s->spatial_idwt_buffer, s->temp_idwt_buffer, w, h, w, s->spatial_decomposition_type, s->spatial_decomposition_count);
             if(s->qlog == LOSSLESS_QLOG){
                 for(y=0; y<h; y++){
                     for(x=0; x<w; x++){
diff --git a/libavcodec/vorbisdec.c b/libavcodec/vorbisdec.c
index 65bfb218ea..133f14bd1a 100644
--- a/libavcodec/vorbisdec.c
+++ b/libavcodec/vorbisdec.c
@@ -1413,17 +1413,24 @@ static av_always_inline int vorbis_residue_decode_internal(vorbis_context *vc,
                                 }
 
                             } else if (vr_type == 2) {
-                                voffs = voffset;
+                                unsigned voffs_div = FASTDIV(voffset, ch);
+                                unsigned voffs_mod = voffset - voffs_div * ch;
 
                                 for (k = 0; k < step; ++k) {
                                     coffs = get_vlc2(gb, codebook.vlc.table, codebook.nb_bits, 3) * dim;
-                                    for (l = 0; l < dim; ++l, ++voffs) {
-                                        vec[voffs / ch + (voffs % ch) * vlen] += codebook.codevectors[coffs + l];  // FPMATH FIXME use if and counter instead of / and %
+                                    for (l = 0; l < dim; ++l) {
+                                        vec[voffs_div + voffs_mod * vlen] +=
+                                            codebook.codevectors[coffs + l];
 
                                         av_dlog(NULL, " pass %d offs: %d curr: %f change: %f cv offs.: %d+%d  \n",
-                                                pass, voffset / ch + (voffs % ch) * vlen,
-                                                vec[voffset / ch + (voffs % ch) * vlen],
+                                                pass, voffs_div + voffs_mod * vlen,
+                                                vec[voffs_div + voffs_mod * vlen],
                                                 codebook.codevectors[coffs + l], coffs, l);
+
+                                        if (++voffs_mod == ch) {
+                                            voffs_div++;
+                                            voffs_mod = 0;
+                                        }
                                     }
                                 }
                             }
diff --git a/libavcodec/x86/snowdsp_mmx.c b/libavcodec/x86/snowdsp_mmx.c
index f107d55e87..f340a868be 100644
--- a/libavcodec/x86/snowdsp_mmx.c
+++ b/libavcodec/x86/snowdsp_mmx.c
@@ -26,9 +26,8 @@
 #include "libavcodec/dwt.h"
 #include "dsputil_mmx.h"
 
-static void ff_snow_horizontal_compose97i_sse2(IDWTELEM *b, int width){
+static void ff_snow_horizontal_compose97i_sse2(IDWTELEM *b, IDWTELEM *temp, int width){
     const int w2= (width+1)>>1;
-    DECLARE_ALIGNED(16, IDWTELEM, temp)[width>>1];
     const int w_l= (width>>1);
     const int w_r= w2 - 1;
     int i;
@@ -215,9 +214,8 @@ static void ff_snow_horizontal_compose97i_sse2(IDWTELEM *b, int width){
     }
 }
 
-static void ff_snow_horizontal_compose97i_mmx(IDWTELEM *b, int width){
+static void ff_snow_horizontal_compose97i_mmx(IDWTELEM *b, IDWTELEM *temp, int width){
     const int w2= (width+1)>>1;
-    IDWTELEM temp[width >> 1];
     const int w_l= (width>>1);
     const int w_r= w2 - 1;
     int i;