Merge remote-tracking branch 'qatar/master'

* qatar/master:
  h264: Factorize declaration of mb_sizes array.
  vsrc_buffer: when no frame is available, return an error instead of segfaulting.
  configure: add dl to frei0r extralibs.
  dsputil x86: use SSE float instruction instead of SSE2 integer equivalent
  dsputil x86: remove deprecated parameter from scalarproduct_int16 prototype
  vp8dsp x86: perform rounding shift with a single instruction
  fate: add BMP tests.
  swscale: handle complete dimensions for monoblack/white.
  aacenc: Mark deinterleave_input_samples argument as const.
  vf_unsharp: Mark readonly variable as const.
  h264: fix 4:2:2 PCM-macroblocks decoding

Conflicts:
	configure
	libavcodec/h264.h
	libavcodec/x86/dsputil_mmx.c
	libavfilter/vf_unsharp.c

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

7 files changed, 27 insertions, 22 deletions

diff --git a/libavcodec/aacenc.c b/libavcodec/aacenc.c
index fd72dafd17..3ac2ad890d 100644
--- a/libavcodec/aacenc.c
+++ b/libavcodec/aacenc.c
@@ -477,7 +477,7 @@ static void put_bitstream_info(AVCodecContext *avctx, AACEncContext *s,
  * Deinterleave input samples.
  * Channels are reordered from libavcodec's default order to AAC order.
  */
-static void deinterleave_input_samples(AACEncContext *s, AVFrame *frame)
+static void deinterleave_input_samples(AACEncContext *s, const AVFrame *frame)
 {
     int ch, i;
     const int sinc = s->channels;
diff --git a/libavcodec/h264.c b/libavcodec/h264.c
index fc612905ab..b909600b3e 100644
--- a/libavcodec/h264.c
+++ b/libavcodec/h264.c
@@ -48,6 +48,8 @@
 // #undef NDEBUG
 #include <assert.h>
 
+const uint16_t ff_h264_mb_sizes[4] = { 256, 384, 512, 768 };
+
 static const uint8_t rem6[QP_MAX_NUM + 1] = {
     0, 1, 2, 3, 4, 5, 0, 1, 2, 3, 4, 5, 0, 1, 2, 3, 4, 5, 0, 1, 2,
     3, 4, 5, 0, 1, 2, 3, 4, 5, 0, 1, 2, 3, 4, 5, 0, 1, 2, 3, 4, 5,
@@ -2126,7 +2128,8 @@ static av_always_inline void hl_decode_mb_internal(H264Context *h, int simple,
         if (pixel_shift) {
             int j;
             GetBitContext gb;
-            init_get_bits(&gb, (uint8_t *)h->mb, 384 * bit_depth);
+            init_get_bits(&gb, (uint8_t *)h->mb,
+                          ff_h264_mb_sizes[h->sps.chroma_format_idc] * bit_depth);
 
             for (i = 0; i < 16; i++) {
                 uint16_t *tmp_y = (uint16_t *)(dest_y + i * linesize);
@@ -2157,7 +2160,7 @@ static av_always_inline void hl_decode_mb_internal(H264Context *h, int simple,
             }
         } else {
             for (i = 0; i < 16; i++)
-                memcpy(dest_y + i * linesize, h->mb + i * 8, 16);
+                memcpy(dest_y + i * linesize, (uint8_t *)h->mb + i * 16, 16);
             if (simple || !CONFIG_GRAY || !(s->flags & CODEC_FLAG_GRAY)) {
                 if (!h->sps.chroma_format_idc) {
                     for (i = 0; i < 8; i++) {
@@ -2165,9 +2168,11 @@ static av_always_inline void hl_decode_mb_internal(H264Context *h, int simple,
                         memset(dest_cr + i*uvlinesize, 1 << (bit_depth - 1), 8);
                     }
                 } else {
+                    uint8_t *src_cb = (uint8_t *)h->mb + 256;
+                    uint8_t *src_cr = (uint8_t *)h->mb + 256 + block_h * 8;
                     for (i = 0; i < block_h; i++) {
-                        memcpy(dest_cb + i * uvlinesize, h->mb + 128 + i * 4, 8);
-                        memcpy(dest_cr + i * uvlinesize, h->mb + 160 + i * 4, 8);
+                        memcpy(dest_cb + i * uvlinesize, src_cb + i * 8, 8);
+                        memcpy(dest_cr + i * uvlinesize, src_cr + i * 8, 8);
                     }
                 }
             }
@@ -2358,7 +2363,8 @@ static av_always_inline void hl_decode_mb_444_internal(H264Context *h,
         } else {
             for (p = 0; p < plane_count; p++)
                 for (i = 0; i < 16; i++)
-                    memcpy(dest[p] + i * linesize, h->mb + p * 128 + i * 8, 16);
+                    memcpy(dest[p] + i * linesize,
+                           (uint8_t *)h->mb + p * 256 + i * 16, 16);
         }
     } else {
         if (IS_INTRA(mb_type)) {
diff --git a/libavcodec/h264.h b/libavcodec/h264.h
index 153d83c695..f5542bfb1d 100644
--- a/libavcodec/h264.h
+++ b/libavcodec/h264.h
@@ -602,6 +602,7 @@ typedef struct H264Context{
 
 
 extern const uint8_t ff_h264_chroma_qp[5][QP_MAX_NUM+1]; ///< One chroma qp table for each possible bit depth (8-12).
+extern const uint16_t ff_h264_mb_sizes[4];
 
 /**
  * Decode SEI
diff --git a/libavcodec/h264_cabac.c b/libavcodec/h264_cabac.c
index 2c83348a5f..65f2cb4dc8 100644
--- a/libavcodec/h264_cabac.c
+++ b/libavcodec/h264_cabac.c
@@ -1984,8 +1984,8 @@ decode_intra_mb:
     h->slice_table[ mb_xy ]= h->slice_num;
 
     if(IS_INTRA_PCM(mb_type)) {
-        static const uint16_t mb_sizes[4] = {256,384,512,768};
-        const int mb_size = mb_sizes[h->sps.chroma_format_idc]*h->sps.bit_depth_luma >> 3;
+        const int mb_size = ff_h264_mb_sizes[h->sps.chroma_format_idc] *
+                            h->sps.bit_depth_luma >> 3;
         const uint8_t *ptr;
 
         // We assume these blocks are very rare so we do not optimize it.
diff --git a/libavcodec/h264_cavlc.c b/libavcodec/h264_cavlc.c
index baddd36800..a4debade2e 100644
--- a/libavcodec/h264_cavlc.c
+++ b/libavcodec/h264_cavlc.c
@@ -765,8 +765,8 @@ decode_intra_mb:
 
     if(IS_INTRA_PCM(mb_type)){
         unsigned int x;
-        static const uint16_t mb_sizes[4] = {256,384,512,768};
-        const int mb_size = mb_sizes[h->sps.chroma_format_idc]*h->sps.bit_depth_luma >> 3;
+        const int mb_size = ff_h264_mb_sizes[h->sps.chroma_format_idc] *
+                            h->sps.bit_depth_luma >> 3;
 
         // We assume these blocks are very rare so we do not optimize it.
         align_get_bits(&s->gb);
diff --git a/libavcodec/x86/dsputil_mmx.c b/libavcodec/x86/dsputil_mmx.c
index 3d85bcf132..e17a8005a7 100644
--- a/libavcodec/x86/dsputil_mmx.c
+++ b/libavcodec/x86/dsputil_mmx.c
@@ -2655,9 +2655,9 @@ void ff_vp3_idct_put_sse2(uint8_t *dest, int line_size, DCTELEM *block);
 void ff_vp3_idct_add_sse2(uint8_t *dest, int line_size, DCTELEM *block);
 
 int32_t ff_scalarproduct_int16_mmx2(const int16_t *v1, const int16_t *v2,
-                                    int len);
+                                    int order);
 int32_t ff_scalarproduct_int16_sse2(const int16_t *v1, const int16_t *v2,
-                                    int len);
+                                    int order);
 int32_t ff_scalarproduct_and_madd_int16_mmx2(int16_t *v1, const int16_t *v2,
                                              const int16_t *v3,
                                              int order, int mul);
diff --git a/libavcodec/x86/vp8dsp.asm b/libavcodec/x86/vp8dsp.asm
index 87e05c4ddf..42bb479b9b 100644
--- a/libavcodec/x86/vp8dsp.asm
+++ b/libavcodec/x86/vp8dsp.asm
@@ -144,6 +144,8 @@ filter_h6_shuf1: db 0, 5, 1, 6, 2, 7, 3, 8, 4, 9, 5, 10, 6, 11,  7, 12
 filter_h6_shuf2: db 1, 2, 2, 3, 3, 4, 4, 5, 5, 6, 6,  7, 7,  8,  8,  9
 filter_h6_shuf3: db 3, 4, 4, 5, 5, 6, 6, 7, 7, 8, 8,  9, 9, 10, 10, 11
 
+pw_256:  times 8 dw 256
+
 pw_20091: times 4 dw 20091
 pw_17734: times 4 dw 17734
 
@@ -205,8 +207,7 @@ cglobal put_vp8_epel%1_h6, 6, 6 + npicregs, 8, dst, dststride, src, srcstride, h
     pmaddubsw m2, m7
     paddsw    m0, m1
     paddsw    m0, m2
-    paddsw    m0, [pw_64]
-    psraw     m0, 7
+    pmulhrsw  m0, [pw_256]
     packuswb  m0, m0
     movh  [dstq], m0        ; store
 
@@ -219,7 +220,7 @@ cglobal put_vp8_epel%1_h6, 6, 6 + npicregs, 8, dst, dststride, src, srcstride, h
 
 cglobal put_vp8_epel%1_h4, 6, 6 + npicregs, 7, dst, dststride, src, srcstride, height, mx, picreg
     shl      mxd, 4
-    mova      m2, [pw_64]
+    mova      m2, [pw_256]
     mova      m3, [filter_h2_shuf]
     mova      m4, [filter_h4_shuf]
 %ifdef PIC
@@ -235,9 +236,8 @@ cglobal put_vp8_epel%1_h4, 6, 6 + npicregs, 7, dst, dststride, src, srcstride, h
     pshufb    m1, m4
     pmaddubsw m0, m5
     pmaddubsw m1, m6
-    paddsw    m0, m2
     paddsw    m0, m1
-    psraw     m0, 7
+    pmulhrsw  m0, m2
     packuswb  m0, m0
     movh  [dstq], m0        ; store
 
@@ -255,7 +255,7 @@ cglobal put_vp8_epel%1_v4, 7, 7, 8, dst, dststride, src, srcstride, height, picr
 %endif
     mova      m5, [fourtap_filter_hb+myq-16]
     mova      m6, [fourtap_filter_hb+myq]
-    mova      m7, [pw_64]
+    mova      m7, [pw_256]
 
     ; read 3 lines
     sub     srcq, srcstrideq
@@ -275,8 +275,7 @@ cglobal put_vp8_epel%1_v4, 7, 7, 8, dst, dststride, src, srcstride, height, picr
     pmaddubsw m2, m6
     paddsw    m4, m2
     mova      m2, m3
-    paddsw    m4, m7
-    psraw     m4, 7
+    pmulhrsw  m4, m7
     packuswb  m4, m4
     movh  [dstq], m4
 
@@ -319,9 +318,8 @@ cglobal put_vp8_epel%1_v6, 7, 7, 8, dst, dststride, src, srcstride, height, picr
     paddsw    m6, m1
     paddsw    m6, m7
     mova      m1, m2
-    paddsw    m6, [pw_64]
     mova      m2, m3
-    psraw     m6, 7
+    pmulhrsw  m6, [pw_256]
     mova      m3, m4
     packuswb  m6, m6
     mova      m4, m5