Merge remote-tracking branch 'qatar/master'

* qatar/master:
  SBR DSP: fix SSE code to not use SSE2 instructions.
  cpu: initialize mask to -1, so that by default, optimizations are used.
  error_resilience: initialize s->block_index[].
  svq3: protect against negative quantizers.
  Don't use ff_cropTbl[] for IDCT.
  swscale: make filterPos 32bit.
  FATE: add CPUFLAGS variable, mapping to -cpuflags avconv option.
  avconv: add -cpuflags option for setting supported cpuflags.
  cpu: add av_set_cpu_flags_mask().
  libx264: Allow overriding the sliced threads option
  avconv: fix counting encoded video size.

Conflicts:
	doc/APIchanges
	doc/fate.texi
	doc/ffmpeg.texi
	ffmpeg.c
	libavcodec/h264idct_template.c
	libavcodec/svq3.c
	libavutil/avutil.h
	libavutil/cpu.c
	libavutil/cpu.h
	libswscale/swscale.c
	tests/Makefile
	tests/fate-run.sh
	tests/regression-funcs.sh

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

7 files changed, 41 insertions, 38 deletions

diff --git a/libswscale/ppc/swscale_altivec.c b/libswscale/ppc/swscale_altivec.c
index b6aed04..df8afb0 100644
--- a/libswscale/ppc/swscale_altivec.c
+++ b/libswscale/ppc/swscale_altivec.c
@@ -144,7 +144,7 @@ static void yuv2planeX_altivec(const int16_t *filter, int filterSize,
 
 static void hScale_altivec_real(SwsContext *c, int16_t *dst, int dstW,
                                 const uint8_t *src, const int16_t *filter,
-                                const int16_t *filterPos, int filterSize)
+                                const int32_t *filterPos, int filterSize)
 {
     register int i;
     DECLARE_ALIGNED(16, int, tempo)[4];
diff --git a/libswscale/swscale.c b/libswscale/swscale.c
index 6dc02b4..38856d9 100644
--- a/libswscale/swscale.c
+++ b/libswscale/swscale.c
@@ -63,7 +63,7 @@ static av_always_inline void fillPlane(uint8_t* plane, int stride,
 
 static void hScale16To19_c(SwsContext *c, int16_t *_dst, int dstW, const uint8_t *_src,
                            const int16_t *filter,
-                           const int16_t *filterPos, int filterSize)
+                           const int32_t *filterPos, int filterSize)
 {
     int i;
     int32_t *dst = (int32_t *) _dst;
@@ -89,7 +89,7 @@ static void hScale16To19_c(SwsContext *c, int16_t *_dst, int dstW, const uint8_t
 
 static void hScale16To15_c(SwsContext *c, int16_t *dst, int dstW, const uint8_t *_src,
                            const int16_t *filter,
-                           const int16_t *filterPos, int filterSize)
+                           const int32_t *filterPos, int filterSize)
 {
     int i;
     const uint16_t *src = (const uint16_t *) _src;
@@ -113,7 +113,7 @@ static void hScale16To15_c(SwsContext *c, int16_t *dst, int dstW, const uint8_t
 
 // bilinear / bicubic scaling
 static void hScale8To15_c(SwsContext *c, int16_t *dst, int dstW, const uint8_t *src,
-                          const int16_t *filter, const int16_t *filterPos,
+                          const int16_t *filter, const int32_t *filterPos,
                           int filterSize)
 {
     int i;
@@ -131,7 +131,7 @@ static void hScale8To15_c(SwsContext *c, int16_t *dst, int dstW, const uint8_t *
 }
 
 static void hScale8To19_c(SwsContext *c, int16_t *_dst, int dstW, const uint8_t *src,
-                          const int16_t *filter, const int16_t *filterPos,
+                          const int16_t *filter, const int32_t *filterPos,
                           int filterSize)
 {
     int i;
@@ -234,7 +234,7 @@ static void hyscale_fast_c(SwsContext *c, int16_t *dst, int dstWidth,
 static av_always_inline void hyscale(SwsContext *c, int16_t *dst, int dstWidth,
                                      const uint8_t *src_in[4], int srcW, int xInc,
                                      const int16_t *hLumFilter,
-                                     const int16_t *hLumFilterPos, int hLumFilterSize,
+                                     const int32_t *hLumFilterPos, int hLumFilterSize,
                                      uint8_t *formatConvBuffer,
                                      uint32_t *pal, int isAlpha)
 {
@@ -282,7 +282,7 @@ static void hcscale_fast_c(SwsContext *c, int16_t *dst1, int16_t *dst2,
 static av_always_inline void hcscale(SwsContext *c, int16_t *dst1, int16_t *dst2, int dstWidth,
                                      const uint8_t *src_in[4],
                                      int srcW, int xInc, const int16_t *hChrFilter,
-                                     const int16_t *hChrFilterPos, int hChrFilterSize,
+                                     const int32_t *hChrFilterPos, int hChrFilterSize,
                                      uint8_t *formatConvBuffer, uint32_t *pal)
 {
     const uint8_t *src1 = src_in[1], *src2 = src_in[2];
@@ -326,10 +326,10 @@ static int swScale(SwsContext *c, const uint8_t* src[],
     const int chrXInc= c->chrXInc;
     const enum PixelFormat dstFormat= c->dstFormat;
     const int flags= c->flags;
-    int16_t *vLumFilterPos= c->vLumFilterPos;
-    int16_t *vChrFilterPos= c->vChrFilterPos;
-    int16_t *hLumFilterPos= c->hLumFilterPos;
-    int16_t *hChrFilterPos= c->hChrFilterPos;
+    int32_t *vLumFilterPos= c->vLumFilterPos;
+    int32_t *vChrFilterPos= c->vChrFilterPos;
+    int32_t *hLumFilterPos= c->hLumFilterPos;
+    int32_t *hChrFilterPos= c->hChrFilterPos;
     int16_t *hLumFilter= c->hLumFilter;
     int16_t *hChrFilter= c->hChrFilter;
     int32_t *lumMmxFilter= c->lumMmxFilter;
diff --git a/libswscale/swscale_internal.h b/libswscale/swscale_internal.h
index 9304356..e73806e 100644
--- a/libswscale/swscale_internal.h
+++ b/libswscale/swscale_internal.h
@@ -299,10 +299,10 @@ typedef struct SwsContext {
     int16_t *hChrFilter;          ///< Array of horizontal filter coefficients for chroma     planes.
     int16_t *vLumFilter;          ///< Array of vertical   filter coefficients for luma/alpha planes.
     int16_t *vChrFilter;          ///< Array of vertical   filter coefficients for chroma     planes.
-    int16_t *hLumFilterPos;       ///< Array of horizontal filter starting positions for each dst[i] for luma/alpha planes.
-    int16_t *hChrFilterPos;       ///< Array of horizontal filter starting positions for each dst[i] for chroma     planes.
-    int16_t *vLumFilterPos;       ///< Array of vertical   filter starting positions for each dst[i] for luma/alpha planes.
-    int16_t *vChrFilterPos;       ///< Array of vertical   filter starting positions for each dst[i] for chroma     planes.
+    int32_t *hLumFilterPos;       ///< Array of horizontal filter starting positions for each dst[i] for luma/alpha planes.
+    int32_t *hChrFilterPos;       ///< Array of horizontal filter starting positions for each dst[i] for chroma     planes.
+    int32_t *vLumFilterPos;       ///< Array of vertical   filter starting positions for each dst[i] for luma/alpha planes.
+    int32_t *vChrFilterPos;       ///< Array of vertical   filter starting positions for each dst[i] for chroma     planes.
     int hLumFilterSize;           ///< Horizontal filter size for luma/alpha pixels.
     int hChrFilterSize;           ///< Horizontal filter size for chroma     pixels.
     int vLumFilterSize;           ///< Vertical   filter size for luma/alpha pixels.
@@ -515,10 +515,10 @@ typedef struct SwsContext {
     /** @{ */
     void (*hyScale)(struct SwsContext *c, int16_t *dst, int dstW,
                     const uint8_t *src, const int16_t *filter,
-                    const int16_t *filterPos, int filterSize);
+                    const int32_t *filterPos, int filterSize);
     void (*hcScale)(struct SwsContext *c, int16_t *dst, int dstW,
                     const uint8_t *src, const int16_t *filter,
-                    const int16_t *filterPos, int filterSize);
+                    const int32_t *filterPos, int filterSize);
     /** @} */
 
     /// Color range conversion function for luma plane if needed.
diff --git a/libswscale/utils.c b/libswscale/utils.c
index 6391d51..7e8285d 100644
--- a/libswscale/utils.c
+++ b/libswscale/utils.c
@@ -191,7 +191,7 @@ static double getSplineCoeff(double a, double b, double c, double d, double dist
                                          dist-1.0);
 }
 
-static int initFilter(int16_t **outFilter, int16_t **filterPos, int *outFilterSize, int xInc,
+static int initFilter(int16_t **outFilter, int32_t **filterPos, int *outFilterSize, int xInc,
                       int srcW, int dstW, int filterAlign, int one, int flags, int cpu_flags,
                       SwsVector *srcFilter, SwsVector *dstFilter, double param[2])
 {
@@ -207,7 +207,7 @@ static int initFilter(int16_t **outFilter, int16_t **filterPos, int *outFilterSi
     emms_c(); //FIXME this should not be required but it IS (even for non-MMX versions)
 
     // NOTE: the +3 is for the MMX(+1)/SSE(+3) scaler which reads over the end
-    FF_ALLOC_OR_GOTO(NULL, *filterPos, (dstW+3)*sizeof(int16_t), fail);
+    FF_ALLOC_OR_GOTO(NULL, *filterPos, (dstW+3)*sizeof(**filterPos), fail);
 
     if (FFABS(xInc - 0x10000) <10) { // unscaled
         int i;
diff --git a/libswscale/x86/scale.asm b/libswscale/x86/scale.asm
index 3c8b6fa..b8cea49 100644
--- a/libswscale/x86/scale.asm
+++ b/libswscale/x86/scale.asm
@@ -38,7 +38,7 @@ SECTION .text
 ;                               (SwsContext *c, int{16,32}_t *dst,
 ;                                int dstW, const uint{8,16}_t *src,
 ;                                const int16_t *filter,
-;                                const int16_t *filterPos, int filterSize);
+;                                const int32_t *filterPos, int filterSize);
 ;
 ; Scale one horizontal line. Input is either 8-bits width or 16-bits width
 ; ($source_width can be either 8, 9, 10 or 16, difference is whether we have to
@@ -53,6 +53,9 @@ SECTION .text
 cglobal hscale%1to%2_%4_%5, %6, 7, %7
 %if ARCH_X86_64
     movsxd        r2, r2d
+%define mov32 movsxd
+%else ; x86-32
+%define mov32 mov
 %endif ; x86-64
 %if %2 == 19
 %if mmsize == 8 ; mmx
@@ -95,14 +98,14 @@ cglobal hscale%1to%2_%4_%5, %6, 7, %7
 %else ; %2 == 19
     lea           r1, [r1+r2*(4>>r2shr)]
 %endif ; %2 == 15/19
-    lea           r5, [r5+r2*(2>>r2shr)]
+    lea           r5, [r5+r2*(4>>r2shr)]
     neg           r2
 
 .loop:
 %if %3 == 4 ; filterSize == 4 scaling
     ; load 2x4 or 4x4 source pixels into m0/m1
-    movsx         r0, word [r5+r2*2+0]   ; filterPos[0]
-    movsx         r6, word [r5+r2*2+2]   ; filterPos[1]
+    mov32         r0, dword [r5+r2*4+0]  ; filterPos[0]
+    mov32         r6, dword [r5+r2*4+4]  ; filterPos[1]
     movlh         m0, [r3+r0*srcmul]     ; src[filterPos[0] + {0,1,2,3}]
 %if mmsize == 8
     movlh         m1, [r3+r6*srcmul]     ; src[filterPos[1] + {0,1,2,3}]
@@ -112,8 +115,8 @@ cglobal hscale%1to%2_%4_%5, %6, 7, %7
 %else ; %1 == 8
     movd          m4, [r3+r6*srcmul]     ; src[filterPos[1] + {0,1,2,3}]
 %endif
-    movsx         r0, word [r5+r2*2+4]   ; filterPos[2]
-    movsx         r6, word [r5+r2*2+6]   ; filterPos[3]
+    mov32         r0, dword [r5+r2*4+8]  ; filterPos[2]
+    mov32         r6, dword [r5+r2*4+12] ; filterPos[3]
     movlh         m1, [r3+r0*srcmul]     ; src[filterPos[2] + {0,1,2,3}]
 %if %1 > 8
     movhps        m1, [r3+r6*srcmul]     ; src[filterPos[3] + {0,1,2,3}]
@@ -156,8 +159,8 @@ cglobal hscale%1to%2_%4_%5, %6, 7, %7
 %endif ; mmx/sse2/ssse3/sse4
 %else ; %3 == 8, i.e. filterSize == 8 scaling
     ; load 2x8 or 4x8 source pixels into m0, m1, m4 and m5
-    movsx         r0, word [r5+r2*1+0]   ; filterPos[0]
-    movsx         r6, word [r5+r2*1+2]   ; filterPos[1]
+    mov32         r0, dword [r5+r2*2+0]  ; filterPos[0]
+    mov32         r6, dword [r5+r2*2+4]  ; filterPos[1]
     movbh         m0, [r3+ r0   *srcmul] ; src[filterPos[0] + {0,1,2,3,4,5,6,7}]
 %if mmsize == 8
     movbh         m1, [r3+(r0+4)*srcmul] ; src[filterPos[0] + {4,5,6,7}]
@@ -165,8 +168,8 @@ cglobal hscale%1to%2_%4_%5, %6, 7, %7
     movbh         m5, [r3+(r6+4)*srcmul] ; src[filterPos[1] + {4,5,6,7}]
 %else ; mmsize == 16
     movbh         m1, [r3+ r6   *srcmul] ; src[filterPos[1] + {0,1,2,3,4,5,6,7}]
-    movsx         r0, word [r5+r2*1+4]   ; filterPos[2]
-    movsx         r6, word [r5+r2*1+6]   ; filterPos[3]
+    mov32         r0, dword [r5+r2*2+8]  ; filterPos[2]
+    mov32         r6, dword [r5+r2*2+12] ; filterPos[3]
     movbh         m4, [r3+ r0   *srcmul] ; src[filterPos[2] + {0,1,2,3,4,5,6,7}]
     movbh         m5, [r3+ r6   *srcmul] ; src[filterPos[3] + {0,1,2,3,4,5,6,7}]
 %endif ; mmsize == 8/16
@@ -251,7 +254,7 @@ cglobal hscale%1to%2_%4_%5, %6, 7, %7
 %define r1x     r1
 %define filter2 r6m
 %endif ; x86-32/64
-    lea           r5, [r5+r2*2]
+    lea           r5, [r5+r2*4]
 %if %2 == 15
     lea           r1, [r1+r2*2]
 %else ; %2 == 19
@@ -261,8 +264,8 @@ cglobal hscale%1to%2_%4_%5, %6, 7, %7
     neg           r2
 
 .loop:
-    movsx         r0, word [r5+r2*2+0]   ; filterPos[0]
-    movsx        r1x, word [r5+r2*2+2]   ; filterPos[1]
+    mov32         r0, dword [r5+r2*4+0]  ; filterPos[0]
+    mov32        r1x, dword [r5+r2*4+4]  ; filterPos[1]
     ; FIXME maybe do 4px/iteration on x86-64 (x86-32 wouldn't have enough regs)?
     pxor          m4, m4
     pxor          m5, m5
@@ -293,7 +296,7 @@ cglobal hscale%1to%2_%4_%5, %6, 7, %7
     jl .innerloop
 
 %ifidn %4, X4
-    movsx        r1x, word [r5+r2*2+2]   ; filterPos[1]
+    mov32        r1x, dword [r5+r2*4+4]  ; filterPos[1]
     movlh         m0, [src_reg+r0 *srcmul] ; split last 4 srcpx of dstpx[0]
     sub          r1x, r6                   ; and first 4 srcpx of dstpx[1]
 %if %1 > 8
diff --git a/libswscale/x86/swscale_mmx.c b/libswscale/x86/swscale_mmx.c
index 28fc004..0c8732c 100644
--- a/libswscale/x86/swscale_mmx.c
+++ b/libswscale/x86/swscale_mmx.c
@@ -94,8 +94,8 @@ void updateMMXDitherTables(SwsContext *c, int dstY, int lumBufIndex, int chrBufI
     int16_t **alpPixBuf= c->alpPixBuf;
     const int vLumBufSize= c->vLumBufSize;
     const int vChrBufSize= c->vChrBufSize;
-    int16_t *vLumFilterPos= c->vLumFilterPos;
-    int16_t *vChrFilterPos= c->vChrFilterPos;
+    int32_t *vLumFilterPos= c->vLumFilterPos;
+    int32_t *vChrFilterPos= c->vChrFilterPos;
     int16_t *vLumFilter= c->vLumFilter;
     int16_t *vChrFilter= c->vChrFilter;
     int32_t *lumMmxFilter= c->lumMmxFilter;
@@ -266,7 +266,7 @@ extern void ff_hscale ## from_bpc ## to ## to_bpc ## _ ## filter_n ## _ ## opt(
                                                 SwsContext *c, int16_t *data, \
                                                 int dstW, const uint8_t *src, \
                                                 const int16_t *filter, \
-                                                const int16_t *filterPos, int filterSize)
+                                                const int32_t *filterPos, int filterSize)
 
 #define SCALE_FUNCS(filter_n, opt) \
     SCALE_FUNC(filter_n,  8, 15, opt); \
diff --git a/libswscale/x86/swscale_template.c b/libswscale/x86/swscale_template.c
index 4eee894..cf8e802 100644
--- a/libswscale/x86/swscale_template.c
+++ b/libswscale/x86/swscale_template.c
@@ -1450,7 +1450,7 @@ static void RENAME(hyscale_fast)(SwsContext *c, int16_t *dst,
                                  int dstWidth, const uint8_t *src,
                                  int srcW, int xInc)
 {
-    int16_t *filterPos = c->hLumFilterPos;
+    int32_t *filterPos = c->hLumFilterPos;
     int16_t *filter    = c->hLumFilter;
     void    *mmx2FilterCode= c->lumMmx2FilterCode;
     int i;
@@ -1546,7 +1546,7 @@ static void RENAME(hcscale_fast)(SwsContext *c, int16_t *dst1, int16_t *dst2,
                                  int dstWidth, const uint8_t *src1,
                                  const uint8_t *src2, int srcW, int xInc)
 {
-    int16_t *filterPos = c->hChrFilterPos;
+    int32_t *filterPos = c->hChrFilterPos;
     int16_t *filter    = c->hChrFilter;
     void    *mmx2FilterCode= c->chrMmx2FilterCode;
     int i;