From: Loren Merritt <pengvado@videolan.org>
Date: Wed, 9 Jan 2008 11:25:09 +0000 (+0000)
Subject: shut up some valgrind warnings
X-Git-Url: https://granicus.if.org/sourcecode?a=commitdiff_plain;h=85a9209e7b85e7a38fa3770f963dc9535fe4a19c;p=libx264

shut up some valgrind warnings


git-svn-id: svn://svn.videolan.org/x264/trunk@719 df754926-b1dd-0310-bc7b-ec298dee348c
---

diff --git a/common/amd64/deblock-a.asm b/common/amd64/deblock-a.asm
index 3e9d5938..70aad4c6 100644
--- a/common/amd64/deblock-a.asm
+++ b/common/amd64/deblock-a.asm
@@ -318,7 +318,8 @@ cglobal x264_deblock_h_luma_sse2
     lea    r11, [r10+r10*2]
     lea    rax, [rdi-4]
     lea    r9,  [rdi-4+r11]
-    %define pix_tmp rsp-104 ; 16x6 for the buffer + 8 for x264_deblock_v_luma_sse2's return address
+    sub    rsp, 0x68
+    %define pix_tmp rsp
 
     ; transpose 6x16 -> tmp space
     TRANSPOSE6x8_MEM  PASS8ROWS(rax, r9, r10, r11), pix_tmp
@@ -352,6 +353,7 @@ cglobal x264_deblock_h_luma_sse2
     movq   mm3, [pix_tmp+0x40]
     TRANSPOSE8x4_STORE  PASS8ROWS(rax, r9, r10, r11)
 
+    add    rsp, 0x68
     ret
 
 
diff --git a/common/amd64/pixel-a.asm b/common/amd64/pixel-a.asm
index 6de7ffeb..0cebcecf 100644
--- a/common/amd64/pixel-a.asm
+++ b/common/amd64/pixel-a.asm
@@ -805,9 +805,10 @@ cglobal x264_intra_satd_x3_4x4_mmxext
 ;  void x264_intra_satd_x3_16x16_mmxext( uint8_t *fenc, uint8_t *fdec, int *res )
 ;-----------------------------------------------------------------------------
 cglobal x264_intra_satd_x3_16x16_mmxext
-%define  sums    rsp-32 ; +24
-%define  top_1d  rsp-64 ; +32
-%define  left_1d rsp-96 ; +32
+    sub          rsp, 96
+%define  sums    rsp+64 ; size 24
+%define  top_1d  rsp+32 ; size 32
+%define  left_1d rsp    ; size 32
 
     mov   qword [sums+0], 0
     mov   qword [sums+8], 0
@@ -913,15 +914,17 @@ cglobal x264_intra_satd_x3_16x16_mmxext
     movd        [parm3q+8], mm2 ; i16x16_dc satd
     movd        [parm3q+4], mm1 ; i16x16_h satd
     movd        [parm3q+0], mm0 ; i16x16_v satd
+    add         rsp, 96
     ret
 
 ;-----------------------------------------------------------------------------
 ;  void x264_intra_satd_x3_8x8c_mmxext( uint8_t *fenc, uint8_t *fdec, int *res )
 ;-----------------------------------------------------------------------------
 cglobal x264_intra_satd_x3_8x8c_mmxext
-%define  sums    rsp-32 ; +24
-%define  top_1d  rsp-48 ; +16
-%define  left_1d rsp-64 ; +16
+    sub          rsp, 64
+%define  sums    rsp+32 ; size 24
+%define  top_1d  rsp+16 ; size 16
+%define  left_1d rsp    ; size 16
 
     mov   qword [sums+0], 0
     mov   qword [sums+8], 0
@@ -1041,6 +1044,7 @@ cglobal x264_intra_satd_x3_8x8c_mmxext
     movd        [parm3q+0], mm0 ; i8x8c_dc satd
     movd        [parm3q+4], mm1 ; i8x8c_h satd
     movd        [parm3q+8], mm2 ; i8x8c_v satd
+    add         rsp, 64
     ret
 
 
diff --git a/common/amd64/pixel-sse2.asm b/common/amd64/pixel-sse2.asm
index fa8668e3..a906d3a1 100644
--- a/common/amd64/pixel-sse2.asm
+++ b/common/amd64/pixel-sse2.asm
@@ -1000,9 +1000,9 @@ cglobal x264_pixel_ssim_4x4x2_core_sse2
     pshufd    xmm6, xmm4, 0xB1
     packssdw  xmm1, xmm2
     paddd     xmm3, xmm5
-    pmaddwd   xmm1, xmm8
-    paddd     xmm4, xmm6
     pshufd    xmm1, xmm1, 0xD8
+    paddd     xmm4, xmm6
+    pmaddwd   xmm1, xmm8
     movdqa    xmm5, xmm3
     punpckldq xmm3, xmm4
     punpckhdq xmm5, xmm4
diff --git a/common/i386/pixel-sse2.asm b/common/i386/pixel-sse2.asm
index 45e2af73..9ab23c61 100644
--- a/common/i386/pixel-sse2.asm
+++ b/common/i386/pixel-sse2.asm
@@ -973,9 +973,9 @@ cglobal x264_pixel_ssim_4x4x2_core_sse2
     pshufd    xmm6, xmm4, 0xB1
     packssdw  xmm1, xmm2
     paddd     xmm3, xmm5
-    pmaddwd   xmm1, xmm7
-    paddd     xmm4, xmm6
     pshufd    xmm1, xmm1, 0xD8
+    paddd     xmm4, xmm6
+    pmaddwd   xmm1, xmm7
     movdqa    xmm5, xmm3
     punpckldq xmm3, xmm4
     punpckhdq xmm5, xmm4
diff --git a/common/mc.c b/common/mc.c
index 00547adb..6ce44e34 100644
--- a/common/mc.c
+++ b/common/mc.c
@@ -430,7 +430,7 @@ void x264_frame_filter( x264_t *h, x264_frame_t *frame, int mb_y, int b_end )
             uint8_t  *ref  = frame->plane[0] + y * stride - PADH;
             uint16_t *line = frame->integral + (y+1) * stride - PADH + 1;
             uint16_t v = line[0] = 0;
-            for( x = 0; x < stride-1; x++ )
+            for( x = 1; x < stride-1; x++ )
                 line[x] = v += ref[x] + line[x-stride] - line[x-stride-1];
             line -= 8*stride;
             if( y >= 9-PADV )