Merge commit '9c12c6ff'

* commit '9c12c6ff':
  motion_est: convert stride to ptrdiff_t

Conflicts:
	libavcodec/me_cmp.c
	libavcodec/ppc/me_cmp.c
	libavcodec/x86/me_cmp_init.c

See: 9c669672Merged-by: Michael Niedermayer <michaelni@gmx.at>

libavcodec/arm/me_cmp_init_arm.c

@@ -26,17 +26,17 @@
 #include "libavcodec/mpegvideo.h"
 
 int ff_pix_abs16_armv6(MpegEncContext *s, uint8_t *blk1, uint8_t *blk2,
-                       int line_size, int h);
+                       ptrdiff_t stride, int h);
 int ff_pix_abs16_x2_armv6(MpegEncContext *s, uint8_t *blk1, uint8_t *blk2,
-                          int line_size, int h);
+                          ptrdiff_t stride, int h);
 int ff_pix_abs16_y2_armv6(MpegEncContext *s, uint8_t *blk1, uint8_t *blk2,
-                          int line_size, int h);
+                          ptrdiff_t stride, int h);
 
 int ff_pix_abs8_armv6(MpegEncContext *s, uint8_t *blk1, uint8_t *blk2,
-                      int line_size, int h);
+                      ptrdiff_t stride, int h);
 
 int ff_sse16_armv6(MpegEncContext *s, uint8_t *blk1, uint8_t *blk2,
-                   int line_size, int h);
+                   ptrdiff_t stride, int h);
 
 av_cold void ff_me_cmp_init_arm(MECmpContext *c, AVCodecContext *avctx)
 {

libavcodec/me_cmp.h

@@ -47,7 +47,8 @@ struct MpegEncContext;
  * width < 8 are neither used nor implemented. */
 typedef int (*me_cmp_func)(struct MpegEncContext *c,
                            uint8_t *blk1 /* align width (8 or 16) */,
-                           uint8_t *blk2 /* align 1 */, int line_size, int h);
+                           uint8_t *blk2 /* align 1 */, ptrdiff_t stride,
+                           int h);
 
 typedef struct MECmpContext {
     int (*sum_abs_dctelem)(int16_t *block /* align 16 */);

libavcodec/motion_est.c

@@ -290,7 +290,7 @@ static int cmp_qpel(MpegEncContext *s, const int x, const int y, const int subx,
 #include "motion_est_template.c"
 
 static int zero_cmp(MpegEncContext *s, uint8_t *a, uint8_t *b,
-                    int stride, int h)
+                    ptrdiff_t stride, int h)
 {
     return 0;
 }

libavcodec/ppc/me_cmp.c

@@ -55,7 +55,7 @@
 }
 #endif
 static int sad16_x2_altivec(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
-                            int line_size, int h)
+                            ptrdiff_t stride, int h)
 {
     int i;
     int __attribute__((aligned(16))) s = 0;
@@ -83,8 +83,8 @@ static int sad16_x2_altivec(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
         /* Add each 4 pixel group together and put 4 results into sad. */
         sad = vec_sum4s(t5, sad);
 
-        pix1 += line_size;
-        pix2 += line_size;
+        pix1 += stride;
+        pix2 += stride;
     }
     /* Sum up the four partial sums, and put the result into s. */
     sumdiffs = vec_sums((vector signed int) sad, (vector signed int) zero);
@@ -95,7 +95,7 @@ static int sad16_x2_altivec(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
 }
 
 static int sad16_y2_altivec(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
-                            int line_size, int h)
+                            ptrdiff_t stride, int h)
 {
     int i;
     int  __attribute__((aligned(16))) s = 0;
@@ -105,9 +105,9 @@ static int sad16_y2_altivec(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
     vector unsigned int sad = (vector unsigned int) vec_splat_u32(0);
     vector signed int sumdiffs;
 
-    uint8_t *pix3 = pix2 + line_size;
+    uint8_t *pix3 = pix2 + stride;
 
-    /* Due to the fact that pix3 = pix2 + line_size, the pix3 of one
+    /* Due to the fact that pix3 = pix2 + stride, the pix3 of one
      * iteration becomes pix2 in the next iteration. We can use this
      * fact to avoid a potentially expensive unaligned read, each
      * time around the loop.
@@ -132,9 +132,9 @@ static int sad16_y2_altivec(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
         /* Add each 4 pixel group together and put 4 results into sad. */
         sad = vec_sum4s(t5, sad);
 
-        pix1 += line_size;
+        pix1 += stride;
         pix2v = pix3v;
-        pix3 += line_size;
+        pix3 += stride;
     }
 
     /* Sum up the four partial sums, and put the result into s. */
@@ -145,11 +145,11 @@ static int sad16_y2_altivec(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
 }
 
 static int sad16_xy2_altivec(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
-                             int line_size, int h)
+                             ptrdiff_t stride, int h)
 {
     int i;
     int  __attribute__((aligned(16))) s = 0;
-    uint8_t *pix3 = pix2 + line_size;
+    uint8_t *pix3 = pix2 + stride;
     const vector unsigned char zero =
         (const vector unsigned char) vec_splat_u8(0);
     const vector unsigned short two =
@@ -163,7 +163,7 @@ static int sad16_xy2_altivec(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
     vector unsigned char perm1, perm2, pix2v, pix2iv;
     GET_PERM(perm1, perm2, pix2);
 
-    /* Due to the fact that pix3 = pix2 + line_size, the pix3 of one
+    /* Due to the fact that pix3 = pix2 + stride, the pix3 of one
      * iteration becomes pix2 in the next iteration. We can use this
      * fact to avoid a potentially expensive unaligned read, as well
      * as some splitting, and vector addition each time around the loop.
@@ -219,8 +219,8 @@ static int sad16_xy2_altivec(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
         /* Add each 4 pixel group together and put 4 results into sad. */
         sad = vec_sum4s(t5, sad);
 
-        pix1 += line_size;
-        pix3 += line_size;
+        pix1 += stride;
+        pix3 += stride;
         /* Transfer the calculated values for pix3 into pix2. */
         t1 = t3;
         t2 = t4;
@@ -234,7 +234,7 @@ static int sad16_xy2_altivec(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
 }
 
 static int sad16_altivec(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
-                         int line_size, int h)
+                         ptrdiff_t stride, int h)
 {
     int i;
     int  __attribute__((aligned(16))) s;
@@ -256,8 +256,8 @@ static int sad16_altivec(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
         /* Add each 4 pixel group together and put 4 results into sad. */
         sad = vec_sum4s(t5, sad);
 
-        pix1 += line_size;
-        pix2 += line_size;
+        pix1 += stride;
+        pix2 += stride;
     }
 
     /* Sum up the four partial sums, and put the result into s. */
@@ -269,7 +269,7 @@ static int sad16_altivec(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
 }
 
 static int sad8_altivec(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
-                        int line_size, int h)
+                        ptrdiff_t stride, int h)
 {
     int i;
     int  __attribute__((aligned(16))) s;
@@ -298,8 +298,8 @@ static int sad8_altivec(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
         /* Add each 4 pixel group together and put 4 results into sad. */
         sad = vec_sum4s(t5, sad);
 
-        pix1 += line_size;
-        pix2 += line_size;
+        pix1 += stride;
+        pix2 += stride;
     }
 
     /* Sum up the four partial sums, and put the result into s. */
@@ -313,7 +313,7 @@ static int sad8_altivec(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
 /* Sum of Squared Errors for an 8x8 block, AltiVec-enhanced.
  * It's the sad8_altivec code above w/ squaring added. */
 static int sse8_altivec(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
-                        int line_size, int h)
+                        ptrdiff_t stride, int h)
 {
     int i;
     int  __attribute__((aligned(16))) s;
@@ -343,8 +343,8 @@ static int sse8_altivec(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
         /* Square the values and add them to our sum. */
         sum = vec_msum(t5, t5, sum);
 
-        pix1 += line_size;
-        pix2 += line_size;
+        pix1 += stride;
+        pix2 += stride;
     }
 
     /* Sum up the four partial sums, and put the result into s. */
@@ -358,7 +358,7 @@ static int sse8_altivec(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
 /* Sum of Squared Errors for a 16x16 block, AltiVec-enhanced.
  * It's the sad16_altivec code above w/ squaring added. */
 static int sse16_altivec(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
-                         int line_size, int h)
+                         ptrdiff_t stride, int h)
 {
     int i;
     int  __attribute__((aligned(16))) s;
@@ -383,8 +383,8 @@ static int sse16_altivec(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
         /* Square the values and add them to our sum. */
         sum = vec_msum(t5, t5, sum);
 
-        pix1 += line_size;
-        pix2 += line_size;
+        pix1 += stride;
+        pix2 += stride;
     }
 
     /* Sum up the four partial sums, and put the result into s. */
@@ -396,7 +396,7 @@ static int sse16_altivec(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
 }
 
 static int hadamard8_diff8x8_altivec(MpegEncContext *s, uint8_t *dst,
-                                     uint8_t *src, int stride, int h)
+                                     uint8_t *src, ptrdiff_t stride, int h)
 {
     int __attribute__((aligned(16))) sum;
     register const vector unsigned char vzero =
@@ -522,7 +522,7 @@ static int hadamard8_diff8x8_altivec(MpegEncContext *s, uint8_t *dst,
  * but xlc goes to around 660 on the regular C code...
  */
 static int hadamard8_diff16x8_altivec(MpegEncContext *s, uint8_t *dst,
-                                      uint8_t *src, int stride, int h)
+                                      uint8_t *src, ptrdiff_t stride, int h)
 {
     int __attribute__((aligned(16))) sum;
     register vector signed short
@@ -713,7 +713,7 @@ static int hadamard8_diff16x8_altivec(MpegEncContext *s, uint8_t *dst,
 }
 
 static int hadamard8_diff16_altivec(MpegEncContext *s, uint8_t *dst,
-                                    uint8_t *src, int stride, int h)
+                                    uint8_t *src, ptrdiff_t stride, int h)
 {
     int score = hadamard8_diff16x8_altivec(s, dst, src, stride, 8);
 

libavcodec/snow_dwt.c

@@ -745,7 +745,7 @@ void ff_spatial_idwt(IDWTELEM *buffer, IDWTELEM *temp, int width, int height,
                               decomposition_count, y);
 }
 
-static inline int w_c(struct MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, int line_size,
+static inline int w_c(struct MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, ptrdiff_t line_size,
                       int w, int h, int type)
 {
     int s, i, j;
@@ -814,32 +814,32 @@ static inline int w_c(struct MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, in
     return s >> 9;
 }
 
-static int w53_8_c(struct MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, int line_size, int h)
+static int w53_8_c(struct MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, ptrdiff_t line_size, int h)
 {
     return w_c(v, pix1, pix2, line_size, 8, h, 1);
 }
 
-static int w97_8_c(struct MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, int line_size, int h)
+static int w97_8_c(struct MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, ptrdiff_t line_size, int h)
 {
     return w_c(v, pix1, pix2, line_size, 8, h, 0);
 }
 
-static int w53_16_c(struct MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, int line_size, int h)
+static int w53_16_c(struct MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, ptrdiff_t line_size, int h)
 {
     return w_c(v, pix1, pix2, line_size, 16, h, 1);
 }
 
-static int w97_16_c(struct MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, int line_size, int h)
+static int w97_16_c(struct MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, ptrdiff_t line_size, int h)
 {
     return w_c(v, pix1, pix2, line_size, 16, h, 0);
 }
 
-int ff_w53_32_c(struct MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, int line_size, int h)
+int ff_w53_32_c(struct MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, ptrdiff_t line_size, int h)
 {
     return w_c(v, pix1, pix2, line_size, 32, h, 1);
 }
 
-int ff_w97_32_c(struct MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, int line_size, int h)
+int ff_w97_32_c(struct MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, ptrdiff_t line_size, int h)
 {
     return w_c(v, pix1, pix2, line_size, 32, h, 0);
 }

libavcodec/snow_dwt.h

@@ -105,8 +105,8 @@ void ff_snow_inner_add_yblock(const uint8_t *obmc, const int obmc_stride,
                               int src_y, int src_stride, slice_buffer *sb,
                               int add, uint8_t *dst8);
 
-int ff_w53_32_c(struct MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, int line_size, int h);
-int ff_w97_32_c(struct MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, int line_size, int h);
+int ff_w53_32_c(struct MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, ptrdiff_t line_size, int h);
+int ff_w97_32_c(struct MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, ptrdiff_t line_size, int h);
 
 void ff_spatial_dwt(int *buffer, int *temp, int width, int height, int stride,
                     int type, int decomposition_count);

libavcodec/x86/me_cmp.asm

@@ -215,7 +215,7 @@ hadamard8_16_wrapper %1, 3
 %elif cpuflag(mmx)
 ALIGN 16
 ; int ff_hadamard8_diff_ ## cpu(MpegEncContext *s, uint8_t *src1,
-;                               uint8_t *src2, int stride, int h)
+;                               uint8_t *src2, ptrdiff_t stride, int h)
 ; r0 = void *s = unused, int h = unused (always 8)
 ; note how r1, r2 and r3 are not clobbered in this function, so 16x16
 ; can simply call this 2x2x (and that's why we access rsp+gprsize
@@ -280,7 +280,7 @@ INIT_XMM ssse3
 HADAMARD8_DIFF 9
 
 ; int ff_sse*_*(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
-;               int line_size, int h)
+;               ptrdiff_t line_size, int h)
 
 %macro SUM_SQUARED_ERRORS 1
 cglobal sse%1, 5,5,8, v, pix1, pix2, lsize, h
@@ -395,7 +395,7 @@ INIT_XMM ssse3
 SUM_ABS_DCTELEM 6, 2
 
 ;------------------------------------------------------------------------------
-; int ff_hf_noise*_mmx(uint8_t *pix1, int lsize, int h)
+; int ff_hf_noise*_mmx(uint8_t *pix1, ptrdiff_t lsize, int h)
 ;------------------------------------------------------------------------------
 ; %1 = 8/16. %2-5=m#
 %macro HF_NOISE_PART1 5
@@ -437,7 +437,6 @@ SUM_ABS_DCTELEM 6, 2
 ; %1 = 8/16
 %macro HF_NOISE 1
 cglobal hf_noise%1, 3,3,0, pix1, lsize, h
-    movsxdifnidn lsizeq, lsized
     sub        hd, 2
     pxor       m7, m7
     pxor       m6, m6
@@ -472,7 +471,7 @@ HF_NOISE 8
 HF_NOISE 16
 
 ;---------------------------------------------------------------------------------------
-;int ff_sad_<opt>(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, int stride, int h);
+;int ff_sad_<opt>(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, ptrdiff_t stride, int h);
 ;---------------------------------------------------------------------------------------
 ;%1 = 8/16
 %macro SAD 1
@@ -527,7 +526,7 @@ INIT_XMM sse2
 SAD 16
 
 ;------------------------------------------------------------------------------------------
-;int ff_sad_x2_<opt>(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, int stride, int h);
+;int ff_sad_x2_<opt>(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, ptrdiff_t stride, int h);
 ;------------------------------------------------------------------------------------------
 ;%1 = 8/16
 %macro SAD_X2 1
@@ -604,7 +603,7 @@ INIT_XMM sse2
 SAD_X2 16
 
 ;------------------------------------------------------------------------------------------
-;int ff_sad_y2_<opt>(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, int stride, int h);
+;int ff_sad_y2_<opt>(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, ptrdiff_t stride, int h);
 ;------------------------------------------------------------------------------------------
 ;%1 = 8/16
 %macro SAD_Y2 1
@@ -674,7 +673,7 @@ INIT_XMM sse2
 SAD_Y2 16
 
 ;-------------------------------------------------------------------------------------------
-;int ff_sad_approx_xy2_<opt>(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, int stride, int h);
+;int ff_sad_approx_xy2_<opt>(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, ptrdiff_t stride, int h);
 ;-------------------------------------------------------------------------------------------
 ;%1 = 8/16
 %macro SAD_APPROX_XY2 1
@@ -776,7 +775,7 @@ SAD_APPROX_XY2 16
 
 ;--------------------------------------------------------------------
 ;int ff_vsad_intra(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
-;                  int line_size, int h);
+;                  ptrdiff_t line_size, int h);
 ;--------------------------------------------------------------------
 ; %1 = 8/16
 %macro VSAD_INTRA 1
@@ -837,7 +836,7 @@ VSAD_INTRA 16
 
 ;---------------------------------------------------------------------
 ;int ff_vsad_approx(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
-;                   int line_size, int h);
+;                   ptrdiff_t line_size, int h);
 ;---------------------------------------------------------------------
 ; %1 = 8/16
 %macro VSAD_APPROX 1