Revert r1506 (we actually are generating columns with the IDCT, so the naming makes sense in retrospect); further de-confusification in the forward DCT

git-svn-id: svn://svn.code.sf.net/p/libjpeg-turbo/code/trunk@1507 632fc199-4ca6-4c93-a231-07263d6284db

3 files changed, 72 insertions, 74 deletions

diff --git a/simd/jfdctint-altivec.c b/simd/jfdctint-altivec.c
index 1ddf261..c13850a 100644
--- a/simd/jfdctint-altivec.c
+++ b/simd/jfdctint-altivec.c
@@ -145,7 +145,7 @@
   out3 = vec_pack(out3l, out3h);  \
 }
 
-#define DO_FDCT_ROWS()  \
+#define DO_FDCT_PASS1()  \
 {  \
   /* Even part */  \
   \
@@ -162,7 +162,7 @@
   DO_FDCT_COMMON(1);  \
 }
 
-#define DO_FDCT_COLS()  \
+#define DO_FDCT_PASS2()  \
 {  \
   /* Even part */  \
   \
@@ -234,7 +234,7 @@ jsimd_fdct_islow_altivec (DCTELEM *data)
   tmp3 = vec_add(col3, col4);
   tmp4 = vec_sub(col3, col4);
 
-  DO_FDCT_ROWS();
+  DO_FDCT_PASS1();
 
   /* Pass 2: process columns */
 
@@ -249,7 +249,7 @@ jsimd_fdct_islow_altivec (DCTELEM *data)
   tmp3 = vec_add(row3, row4);
   tmp4 = vec_sub(row3, row4);
 
-  DO_FDCT_COLS();
+  DO_FDCT_PASS2();
 
   vec_st(out0, 0, data);
   vec_st(out1, 16, data);
diff --git a/simd/jidctfst-altivec.c b/simd/jidctfst-altivec.c
index 9779a6a..fd7a2a3 100644
--- a/simd/jidctfst-altivec.c
+++ b/simd/jidctfst-altivec.c
@@ -119,12 +119,11 @@ jsimd_idct_ifast_altivec (void * dct_table_, JCOEFPTR coef_block,
 
   __vector short row0, row1, row2, row3, row4, row5, row6, row7,
     col0, col1, col2, col3, col4, col5, col6, col7,
-    final0, final1, final2, final3, final4, final5, final6, final7,
     quant0, quant1, quant2, quant3, quant4, quant5, quant6, quant7,
     tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7, tmp10, tmp11, tmp12, tmp13,
     z5, z10, z10s, z11, z12s, z13,
     out0, out1, out2, out3, out4, out5, out6, out7;
-  __vector signed char final;
+  __vector signed char outb;
 
   /* Constants */
   __vector short pw_zero = { __8X(0) },
@@ -206,53 +205,53 @@ jsimd_idct_ifast_altivec (void * dct_table_, JCOEFPTR coef_block,
   out6 = vec_sra(out6, pass1_bits3);
   out7 = vec_sra(out7, pass1_bits3);
 
-  TRANSPOSE(out, final);
+  TRANSPOSE(out, col);
 
-  final = vec_packs(final0, final0);
-  final = vec_add(final, pb_centerjsamp);
+  outb = vec_packs(col0, col0);
+  outb = vec_add(outb, pb_centerjsamp);
   outptr = (int *)(output_buf[0] + output_col);
-  vec_ste((__vector int)final, 0, outptr);
-  vec_ste((__vector int)final, 4, outptr);
+  vec_ste((__vector int)outb, 0, outptr);
+  vec_ste((__vector int)outb, 4, outptr);
 
-  final = vec_packs(final1, final1);
-  final = vec_add(final, pb_centerjsamp);
+  outb = vec_packs(col1, col1);
+  outb = vec_add(outb, pb_centerjsamp);
   outptr = (int *)(output_buf[1] + output_col);
-  vec_ste((__vector int)final, 0, outptr);
-  vec_ste((__vector int)final, 4, outptr);
+  vec_ste((__vector int)outb, 0, outptr);
+  vec_ste((__vector int)outb, 4, outptr);
 
-  final = vec_packs(final2, final2);
-  final = vec_add(final, pb_centerjsamp);
+  outb = vec_packs(col2, col2);
+  outb = vec_add(outb, pb_centerjsamp);
   outptr = (int *)(output_buf[2] + output_col);
-  vec_ste((__vector int)final, 0, outptr);
-  vec_ste((__vector int)final, 4, outptr);
+  vec_ste((__vector int)outb, 0, outptr);
+  vec_ste((__vector int)outb, 4, outptr);
 
-  final = vec_packs(final3, final3);
-  final = vec_add(final, pb_centerjsamp);
+  outb = vec_packs(col3, col3);
+  outb = vec_add(outb, pb_centerjsamp);
   outptr = (int *)(output_buf[3] + output_col);
-  vec_ste((__vector int)final, 0, outptr);
-  vec_ste((__vector int)final, 4, outptr);
+  vec_ste((__vector int)outb, 0, outptr);
+  vec_ste((__vector int)outb, 4, outptr);
 
-  final = vec_packs(final4, final4);
-  final = vec_add(final, pb_centerjsamp);
+  outb = vec_packs(col4, col4);
+  outb = vec_add(outb, pb_centerjsamp);
   outptr = (int *)(output_buf[4] + output_col);
-  vec_ste((__vector int)final, 0, outptr);
-  vec_ste((__vector int)final, 4, outptr);
+  vec_ste((__vector int)outb, 0, outptr);
+  vec_ste((__vector int)outb, 4, outptr);
 
-  final = vec_packs(final5, final5);
-  final = vec_add(final, pb_centerjsamp);
+  outb = vec_packs(col5, col5);
+  outb = vec_add(outb, pb_centerjsamp);
   outptr = (int *)(output_buf[5] + output_col);
-  vec_ste((__vector int)final, 0, outptr);
-  vec_ste((__vector int)final, 4, outptr);
+  vec_ste((__vector int)outb, 0, outptr);
+  vec_ste((__vector int)outb, 4, outptr);
 
-  final = vec_packs(final6, final6);
-  final = vec_add(final, pb_centerjsamp);
+  outb = vec_packs(col6, col6);
+  outb = vec_add(outb, pb_centerjsamp);
   outptr = (int *)(output_buf[6] + output_col);
-  vec_ste((__vector int)final, 0, outptr);
-  vec_ste((__vector int)final, 4, outptr);
+  vec_ste((__vector int)outb, 0, outptr);
+  vec_ste((__vector int)outb, 4, outptr);
 
-  final = vec_packs(final7, final7);
-  final = vec_add(final, pb_centerjsamp);
+  outb = vec_packs(col7, col7);
+  outb = vec_add(outb, pb_centerjsamp);
   outptr = (int *)(output_buf[7] + output_col);
-  vec_ste((__vector int)final, 0, outptr);
-  vec_ste((__vector int)final, 4, outptr);
+  vec_ste((__vector int)outb, 0, outptr);
+  vec_ste((__vector int)outb, 4, outptr);
 }
diff --git a/simd/jidctint-altivec.c b/simd/jidctint-altivec.c
index a7b6f00..7f0f8d0 100644
--- a/simd/jidctint-altivec.c
+++ b/simd/jidctint-altivec.c
@@ -214,7 +214,6 @@ jsimd_idct_islow_altivec (void * dct_table_, JCOEFPTR coef_block,
 
   __vector short row0, row1, row2, row3, row4, row5, row6, row7,
     col0, col1, col2, col3, col4, col5, col6, col7,
-    final0, final1, final2, final3, final4, final5, final6, final7,
     quant0, quant1, quant2, quant3, quant4, quant5, quant6, quant7,
     tmp0, tmp1, tmp2, tmp3, z3, z4,
     z34l, z34h, col71l, col71h, col26l, col26h, col53l, col53h,
@@ -225,7 +224,7 @@ jsimd_idct_islow_altivec (void * dct_table_, JCOEFPTR coef_block,
     z3l, z3h, z4l, z4h,
     out0l, out0h, out1l, out1h, out2l, out2h, out3l, out3h, out4l, out4h,
     out5l, out5h, out6l, out6h, out7l, out7h;
-  __vector signed char final;
+  __vector signed char outb;
 
   /* Constants */
   __vector short pw_zero = { __8X(0) },
@@ -308,53 +307,53 @@ jsimd_idct_islow_altivec (void * dct_table_, JCOEFPTR coef_block,
 
   DO_IDCT(row, 2);
 
-  TRANSPOSE(out, final);
+  TRANSPOSE(out, col);
 
-  final = vec_packs(final0, final0);
-  final = vec_add(final, pb_centerjsamp);
+  outb = vec_packs(col0, col0);
+  outb = vec_add(outb, pb_centerjsamp);
   outptr = (int *)(output_buf[0] + output_col);
-  vec_ste((__vector int)final, 0, outptr);
-  vec_ste((__vector int)final, 4, outptr);
+  vec_ste((__vector int)outb, 0, outptr);
+  vec_ste((__vector int)outb, 4, outptr);
 
-  final = vec_packs(final1, final1);
-  final = vec_add(final, pb_centerjsamp);
+  outb = vec_packs(col1, col1);
+  outb = vec_add(outb, pb_centerjsamp);
   outptr = (int *)(output_buf[1] + output_col);
-  vec_ste((__vector int)final, 0, outptr);
-  vec_ste((__vector int)final, 4, outptr);
+  vec_ste((__vector int)outb, 0, outptr);
+  vec_ste((__vector int)outb, 4, outptr);
 
-  final = vec_packs(final2, final2);
-  final = vec_add(final, pb_centerjsamp);
+  outb = vec_packs(col2, col2);
+  outb = vec_add(outb, pb_centerjsamp);
   outptr = (int *)(output_buf[2] + output_col);
-  vec_ste((__vector int)final, 0, outptr);
-  vec_ste((__vector int)final, 4, outptr);
+  vec_ste((__vector int)outb, 0, outptr);
+  vec_ste((__vector int)outb, 4, outptr);
 
-  final = vec_packs(final3, final3);
-  final = vec_add(final, pb_centerjsamp);
+  outb = vec_packs(col3, col3);
+  outb = vec_add(outb, pb_centerjsamp);
   outptr = (int *)(output_buf[3] + output_col);
-  vec_ste((__vector int)final, 0, outptr);
-  vec_ste((__vector int)final, 4, outptr);
+  vec_ste((__vector int)outb, 0, outptr);
+  vec_ste((__vector int)outb, 4, outptr);
 
-  final = vec_packs(final4, final4);
-  final = vec_add(final, pb_centerjsamp);
+  outb = vec_packs(col4, col4);
+  outb = vec_add(outb, pb_centerjsamp);
   outptr = (int *)(output_buf[4] + output_col);
-  vec_ste((__vector int)final, 0, outptr);
-  vec_ste((__vector int)final, 4, outptr);
+  vec_ste((__vector int)outb, 0, outptr);
+  vec_ste((__vector int)outb, 4, outptr);
 
-  final = vec_packs(final5, final5);
-  final = vec_add(final, pb_centerjsamp);
+  outb = vec_packs(col5, col5);
+  outb = vec_add(outb, pb_centerjsamp);
   outptr = (int *)(output_buf[5] + output_col);
-  vec_ste((__vector int)final, 0, outptr);
-  vec_ste((__vector int)final, 4, outptr);
+  vec_ste((__vector int)outb, 0, outptr);
+  vec_ste((__vector int)outb, 4, outptr);
 
-  final = vec_packs(final6, final6);
-  final = vec_add(final, pb_centerjsamp);
+  outb = vec_packs(col6, col6);
+  outb = vec_add(outb, pb_centerjsamp);
   outptr = (int *)(output_buf[6] + output_col);
-  vec_ste((__vector int)final, 0, outptr);
-  vec_ste((__vector int)final, 4, outptr);
+  vec_ste((__vector int)outb, 0, outptr);
+  vec_ste((__vector int)outb, 4, outptr);
 
-  final = vec_packs(final7, final7);
-  final = vec_add(final, pb_centerjsamp);
+  outb = vec_packs(col7, col7);
+  outb = vec_add(outb, pb_centerjsamp);
   outptr = (int *)(output_buf[7] + output_col);
-  vec_ste((__vector int)final, 0, outptr);
-  vec_ste((__vector int)final, 4, outptr);
+  vec_ste((__vector int)outb, 0, outptr);
+  vec_ste((__vector int)outb, 4, outptr);
 }