[xiph-commits] r11683 - trunk/theora-exp/lib/x86

[j at svn.xiph.org]

Author: j
Date: 2006-07-01 14:33:56 -0700 (Sat, 01 Jul 2006)
New Revision: 11683

Modified:
   trunk/theora-exp/lib/x86/mmxfrag.c
   trunk/theora-exp/lib/x86/mmxidct.c
Log:
asm fixed to compile with Apple's broken gas on Mac OS X Intel.
 * replace .balign 16 with .p2align 4
 * reference OC_IDCT_CONSTS as it is done for x86_64.
   (this also works on i386, so no need for the ifdefs.)



Modified: trunk/theora-exp/lib/x86/mmxfrag.c
===================================================================
--- trunk/theora-exp/lib/x86/mmxfrag.c	2006-07-01 14:53:19 UTC (rev 11682)
+++ trunk/theora-exp/lib/x86/mmxfrag.c	2006-07-01 21:33:56 UTC (rev 11683)
@@ -21,7 +21,7 @@
  const ogg_int16_t *_residue){
   __asm__ __volatile__(
    "  mov          $0x7, %%ecx  \n\t" /* 8x loop */
-   "  .balign 16                \n\t"
+   "  .p2align 4                \n\t"
    "1:movq           %3, %%mm0  \n\t" /* Set mm0 to 0x0080008000800080 */
    "  movq         (%1), %%mm2  \n\t" /* First four input values */
    "  movq        %%mm0, %%mm1  \n\t" /* Set mm1 == mm0 */
@@ -48,7 +48,7 @@
   __asm__ __volatile__(
    "  movl         $0x7,   %%eax   \n\t" /* 8x loop */
    "  pxor         %%mm0,  %%mm0   \n\t" /* zero mm0  */
-   "  .balign 16                   \n\t"
+   "  .p2align 4                   \n\t"
    "1: movq        (%4),   %%mm2   \n\t" /* load mm2 with _src */
    "  movq         %%mm2,  %%mm3   \n\t" /* copy mm2 to mm3 */
    "  punpckhbw    %%mm0,  %%mm2   \n\t" /* expand high part of _src to 16 bits */
@@ -81,7 +81,7 @@
    "  movl         $0x7,   %%eax   \n\t" /* 8x loop */
    "  pxor         %%mm0,  %%mm0   \n\t" /* zero mm0 */
    "  movq         (%4),   %%mm2   \n\t" /* load mm2 with _src1 */
-   "  .balign 16                   \n\t"
+   "  .p2align 4                   \n\t"
    "1:movq         (%6),   %%mm4   \n\t" /* packed SRC2 */ 
    "  movq         %%mm2,  %%mm3   \n\t" /* copy to mm3 */
    "  movq         %%mm4,  %%mm5   \n\t" /* copy packed src2 to mm5 */
@@ -124,7 +124,7 @@
    "  movl         $0x7,   %7      \n\t" /* 8x loop */
    "  pxor         %%mm0,  %%mm0   \n\t" /* zero mm0 */
    "  movq         (%4),   %%mm2   \n\t" /* load mm2 with _src1 */
-   "  .balign 16                   \n\t"
+   "  .p2align 4                   \n\t"
    "1: movq        (%6),   %%mm4   \n\t" /* packed SRC2 */ 
    "  movq         %%mm2,  %%mm3   \n\t" /* copy to mm3 */
    "  movq         %%mm4,  %%mm5   \n\t" /* copy packed src2 to mm5 */

Modified: trunk/theora-exp/lib/x86/mmxidct.c
===================================================================
--- trunk/theora-exp/lib/x86/mmxidct.c	2006-07-01 14:53:19 UTC (rev 11682)
+++ trunk/theora-exp/lib/x86/mmxidct.c	2006-07-01 21:33:56 UTC (rev 11683)
@@ -269,21 +269,12 @@
 void oc_idct8x8_mmx(ogg_int16_t _y[64]){
 /*This routine accepts an 8x8 matrix, but in transposed form.
   Every 4x4 submatrix is transposed.*/
-#if (defined(__amd64__) || defined(__x86_64__))
   __asm__ __volatile__(
    ""
    :
    :"d" (_y),
     "c" (OC_IDCT_CONSTS)
   );
-#else
-  __asm__ __volatile__(
-   "  movl $OC_IDCT_CONSTS,%%ecx\n\t"
-   :
-   :"d" (_y)
-   :"ecx"
-  );
-#endif
 #define OC_I(_k)      OC_M2STR((_k*16))"("OC_Y_REG")"
 #define OC_J(_k)      OC_M2STR(((_k-4)*16)+8)"("OC_Y_REG")"
   OC_ROW_IDCT;
@@ -432,21 +423,12 @@
 )
 
 void oc_idct8x8_10_mmx(ogg_int16_t _y[64],const ogg_int16_t _x[64]){
-#if (defined(__amd64__) ||  defined(__x86_64__))
   __asm__ __volatile__(
    ""
    :
    :"d" (_y),
    "c" (OC_IDCT_CONSTS)
   );
-#else
-  __asm__ __volatile__(
-   "  mov    $OC_IDCT_CONSTS,%%ecx\n\t"
-   :
-   :"d" (_y)
-   :"ecx"
-  );
-#endif
 #define OC_I(_k) OC_M2STR((_k*16))"("OC_Y_REG")"
 #define OC_J(_k) OC_M2STR(((_k-4)*16)+8)"("OC_Y_REG")"
   /*Done with dequant, descramble, and partial transpose.