[xiph-commits] r11352 - branches/theora-mmx/lib/x86_64

[dlenski at svn.xiph.org]

Author: dlenski
Date: 2006-05-06 12:37:48 -0700 (Sat, 06 May 2006)
New Revision: 11352

Modified:
   branches/theora-mmx/lib/x86_64/dsp_mmx.c
Log:
A few very small changes to lib/x86_64: basically I am trying to make all integer function arguments 64-bit where possible, but maintaining the return values as 32-bit quantities in hope of preserving the current rounding behavior.


Modified: branches/theora-mmx/lib/x86_64/dsp_mmx.c
===================================================================
--- branches/theora-mmx/lib/x86_64/dsp_mmx.c	2006-05-06 19:36:24 UTC (rev 11351)
+++ branches/theora-mmx/lib/x86_64/dsp_mmx.c	2006-05-06 19:37:48 UTC (rev 11352)
@@ -26,7 +26,7 @@
 #define DSP_OP_ABS_DIFF(a,b) abs((((int)(a)) - ((int)(b))))
 
 static void sub8x8__mmx (unsigned char *FiltPtr, unsigned char *ReconPtr,
-                  ogg_int16_t *DctInputPtr, ogg_uint32_t PixelsPerLine,
+                  ogg_int16_t *DctInputPtr, ogg_uint64_t PixelsPerLine,
                   ogg_uint64_t ReconPixelsPerLine) 
 {
   __asm__ __volatile__ (
@@ -50,7 +50,7 @@
     "  movq        %%mm0,  (%2)     \n\t" /* write answer out */
     "  movq        %%mm2, 8(%2)     \n\t" /* write answer out */
     /* Increment pointers */
-    "  add         $16, %2           \n\t"
+    "  add         $16, %2          \n\t"
     "  add         %3, %0           \n\t"
     "  add         %4, %1           \n\t"
     ".endr                          \n\t"
@@ -65,7 +65,7 @@
 }
 
 static void sub8x8_128__mmx (unsigned char *FiltPtr, ogg_int16_t *DctInputPtr,
-                      ogg_uint32_t PixelsPerLine) 
+                      ogg_uint64_t PixelsPerLine) 
 {
   __asm__ __volatile__ (
     "  .balign 16                   \n\t"
@@ -99,7 +99,7 @@
 
 static void sub8x8avg2__mmx (unsigned char *FiltPtr, unsigned char *ReconPtr1,
                      unsigned char *ReconPtr2, ogg_int16_t *DctInputPtr,
-                     ogg_uint32_t PixelsPerLine,
+                     ogg_uint64_t PixelsPerLine,
                      ogg_uint64_t ReconPixelsPerLine) 
 {
   __asm__ __volatile__ (
@@ -188,7 +188,7 @@
     "  paddw       %%mm0, %%mm5     \n\t"
     "  movd        %%mm5, %%rdi     \n\t"
     "  movsx       %%di, %%rdi      \n\t"
-    "  mov        %%rdi, %0        \n\t"
+    "  mov         %%rdi, %0        \n\t"
 
     "  movq        %%mm7, %%mm0     \n\t"
     "  psrlq       $32, %%mm7       \n\t"
@@ -206,11 +206,11 @@
   return (( (XXSum<<6) - XSum*XSum ) );
 }
 
-static ogg_uint32_t inter8x8_err__mmx (unsigned char *SrcData, ogg_uint32_t SrcStride,
-		                 unsigned char *RefDataPtr, ogg_uint32_t RefStride)
+static ogg_uint32_t inter8x8_err__mmx (unsigned char *SrcData, ogg_uint64_t SrcStride,
+		                 unsigned char *RefDataPtr, ogg_uint64_t RefStride)
 {
-  ogg_uint32_t  XSum;
-  ogg_uint32_t  XXSum;
+  ogg_uint64_t  XSum;
+  ogg_uint64_t  XXSum;
 
   __asm__ __volatile__ (
     "  .balign 16                   \n\t"
@@ -256,7 +256,7 @@
     "  paddw       %%mm0, %%mm5     \n\t"
     "  movd        %%mm5, %%rdi     \n\t"
     "  movsx       %%di, %%rdi      \n\t"
-    "  mov        %%rdi, %0        \n\t"
+    "  mov         %%rdi, %0        \n\t"
 
     "  movq        %%mm7, %%mm0     \n\t"
     "  psrlq       $32, %%mm7       \n\t"