Regroup the 32 and 64bit sections and restore 64bit CAXPY

kernel/power/KERNEL.POWER8

@@ -1,3 +1,4 @@
+# Big-endian 32bit (AIX) is supported through the POWER6 GEMM kernels, no separate TRMM
 ifeq ($(__BYTE_ORDER__)$(BINARY32),__ORDER_BIG_ENDIAN__1)
 SGEMMKERNEL    =  gemm_kernel_power6.S
 SGEMMINCOPY    =
@@ -35,12 +36,6 @@ ZGEMMINCOPYOBJ =  zgemm_incopy$(TSUFFIX).$(SUFFIX)
 ZGEMMITCOPYOBJ =  zgemm_itcopy$(TSUFFIX).$(SUFFIX)
 ZGEMMONCOPYOBJ =  zgemm_oncopy$(TSUFFIX).$(SUFFIX)
 ZGEMMOTCOPYOBJ =  zgemm_otcopy$(TSUFFIX).$(SUFFIX)
-DTRSMKERNEL_LN  =  trsm_kernel_power6_LN.S
-DTRSMKERNEL_LT  =  trsm_kernel_power6_LT.S
-DTRSMKERNEL_RN  =  trsm_kernel_power6_LT.S
-DTRSMKERNEL_RT  =  trsm_kernel_power6_RT.S
-
-CAXPYKERNEL  = zaxpy.S
 
 else
 
@@ -93,10 +88,6 @@ ZGEMMONCOPYOBJ =  zgemm_oncopy$(TSUFFIX).$(SUFFIX)
 ZGEMMOTCOPYOBJ =  zgemm_otcopy$(TSUFFIX).$(SUFFIX)
 ZGEMMINCOPYOBJ =  zgemm_incopy$(TSUFFIX).$(SUFFIX)
 ZGEMMITCOPYOBJ =  zgemm_itcopy$(TSUFFIX).$(SUFFIX)
-DTRSMKERNEL_LN	= ../generic/trsm_kernel_LN.c
-DTRSMKERNEL_LT	= dtrsm_kernel_LT_16x4_power8.S
-DTRSMKERNEL_RN	= ../generic/trsm_kernel_RN.c
-DTRSMKERNEL_RT	= ../generic/trsm_kernel_RT.c
 endif
 
 STRSMKERNEL_LN	=  ../generic/trsm_kernel_LN.c
@@ -104,10 +95,17 @@ STRSMKERNEL_LT	=  ../generic/trsm_kernel_LT.c
 STRSMKERNEL_RN	=  ../generic/trsm_kernel_RN.c
 STRSMKERNEL_RT	=  ../generic/trsm_kernel_RT.c
 
-#DTRSMKERNEL_LN	= ../generic/trsm_kernel_LN.c
-#DTRSMKERNEL_LT	= dtrsm_kernel_LT_16x4_power8.S
-#DTRSMKERNEL_RN	= ../generic/trsm_kernel_RN.c
-#DTRSMKERNEL_RT	= ../generic/trsm_kernel_RT.c
+ifeq ($(__BYTE_ORDER__)$(BINARY32),__ORDER_BIG_ENDIAN__1)
+DTRSMKERNEL_LN  =  trsm_kernel_power6_LN.S
+DTRSMKERNEL_LT  =  trsm_kernel_power6_LT.S
+DTRSMKERNEL_RN  =  trsm_kernel_power6_LT.S
+DTRSMKERNEL_RT  =  trsm_kernel_power6_RT.S
+else
+DTRSMKERNEL_LN	= ../generic/trsm_kernel_LN.c
+DTRSMKERNEL_LT	= dtrsm_kernel_LT_16x4_power8.S
+DTRSMKERNEL_RN	= ../generic/trsm_kernel_RN.c
+DTRSMKERNEL_RT	= ../generic/trsm_kernel_RT.c
+endif
 
 CTRSMKERNEL_LN	= ../generic/trsm_kernel_LN.c
 CTRSMKERNEL_LT	= ../generic/trsm_kernel_LT.c
@@ -204,15 +202,20 @@ ZASUMKERNEL  = zasum.c
 #
 SAXPYKERNEL  = saxpy.c
 DAXPYKERNEL  = daxpy.c
-#ifneq ($(__BYTE_ORDER__),__ORDER_BIG_ENDIAN__)
-#ifneq ($(GCCVERSIONGTEQ9),1)
-#CAXPYKERNEL  = caxpy_power8.S
-#else
-#CAXPYKERNEL  = caxpy.c
-#endif
-#else
-#CAXPYKERNEL  = caxpy.c
-#endif
+ä
+ifeq ($(__BYTE_ORDER__)$(BINARY32),__ORDER_BIG_ENDIAN__1)
+CAXPYKERNEL  = zaxpy.S
+else
+ifneq ($(__BYTE_ORDER__),__ORDER_BIG_ENDIAN__)
+ifneq ($(GCCVERSIONGTEQ9),1)
+CAXPYKERNEL  = caxpy_power8.S
+else
+CAXPYKERNEL  = caxpy.c
+endif
+else
+CAXPYKERNEL  = caxpy.c
+endif
+endif
 #
 ZAXPYKERNEL  = zaxpy.c
 #