===================================================================
@@ -0,0 +1,96 @@
+/* { dg-options "-maltivec=be -mabi=altivec -std=gnu99 -mno-vsx" } */
+
+#include "harness.h"
+
+static void test()
+{
+ /* Input vectors. */
+ vector unsigned char vuca = {0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15};
+ vector unsigned char vucb
+ = {16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31};
+ vector signed char vsca
+ = {-16,-15,-14,-13,-12,-11,-10,-9,-8,-7,-6,-5,-4,-3,-2,-1};
+ vector signed char vscb = {0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15};
+ vector unsigned short vusa = {0,1,2,3,4,5,6,7};
+ vector unsigned short vusb = {8,9,10,11,12,13,14,15};
+ vector signed short vssa = {-8,-7,-6,-5,-4,-3,-2,-1};
+ vector signed short vssb = {0,1,2,3,4,5,6,7};
+ vector unsigned int vuia = {0,1,2,3};
+ vector unsigned int vuib = {4,5,6,7};
+ vector signed int vsia = {-4,-3,-2,-1};
+ vector signed int vsib = {0,1,2,3};
+ vector float vfa = {-4.0,-3.0,-2.0,-1.0};
+ vector float vfb = {0.0,1.0,2.0,3.0};
+
+ /* Result vectors. */
+ vector unsigned char vuch, vucl;
+ vector signed char vsch, vscl;
+ vector unsigned short vush, vusl;
+ vector signed short vssh, vssl;
+ vector unsigned int vuih, vuil;
+ vector signed int vsih, vsil;
+ vector float vfh, vfl;
+
+ /* Expected result vectors. */
+#if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
+ vector unsigned char vucrh = {24,8,25,9,26,10,27,11,28,12,29,13,30,14,31,15};
+ vector unsigned char vucrl = {16,0,17,1,18,2,19,3,20,4,21,5,22,6,23,7};
+ vector signed char vscrh = {8,-8,9,-7,10,-6,11,-5,12,-4,13,-3,14,-2,15,-1};
+ vector signed char vscrl = {0,-16,1,-15,2,-14,3,-13,4,-12,5,-11,6,-10,7,-9};
+ vector unsigned short vusrh = {12,4,13,5,14,6,15,7};
+ vector unsigned short vusrl = {8,0,9,1,10,2,11,3};
+ vector signed short vssrh = {4,-4,5,-3,6,-2,7,-1};
+ vector signed short vssrl = {0,-8,1,-7,2,-6,3,-5};
+ vector unsigned int vuirh = {6,2,7,3};
+ vector unsigned int vuirl = {4,0,5,1};
+ vector signed int vsirh = {2,-2,3,-1};
+ vector signed int vsirl = {0,-4,1,-3};
+ vector float vfrh = {2.0,-2.0,3.0,-1.0};
+ vector float vfrl = {0.0,-4.0,1.0,-3.0};
+#else
+ vector unsigned char vucrh = {0,16,1,17,2,18,3,19,4,20,5,21,6,22,7,23};
+ vector unsigned char vucrl = {8,24,9,25,10,26,11,27,12,28,13,29,14,30,15,31};
+ vector signed char vscrh = {-16,0,-15,1,-14,2,-13,3,-12,4,-11,5,-10,6,-9,7};
+ vector signed char vscrl = {-8,8,-7,9,-6,10,-5,11,-4,12,-3,13,-2,14,-1,15};
+ vector unsigned short vusrh = {0,8,1,9,2,10,3,11};
+ vector unsigned short vusrl = {4,12,5,13,6,14,7,15};
+ vector signed short vssrh = {-8,0,-7,1,-6,2,-5,3};
+ vector signed short vssrl = {-4,4,-3,5,-2,6,-1,7};
+ vector unsigned int vuirh = {0,4,1,5};
+ vector unsigned int vuirl = {2,6,3,7};
+ vector signed int vsirh = {-4,0,-3,1};
+ vector signed int vsirl = {-2,2,-1,3};
+ vector float vfrh = {-4.0,0.0,-3.0,1.0};
+ vector float vfrl = {-2.0,2.0,-1.0,3.0};
+#endif
+
+ vuch = vec_mergeh (vuca, vucb);
+ vucl = vec_mergel (vuca, vucb);
+ vsch = vec_mergeh (vsca, vscb);
+ vscl = vec_mergel (vsca, vscb);
+ vush = vec_mergeh (vusa, vusb);
+ vusl = vec_mergel (vusa, vusb);
+ vssh = vec_mergeh (vssa, vssb);
+ vssl = vec_mergel (vssa, vssb);
+ vuih = vec_mergeh (vuia, vuib);
+ vuil = vec_mergel (vuia, vuib);
+ vsih = vec_mergeh (vsia, vsib);
+ vsil = vec_mergel (vsia, vsib);
+ vfh = vec_mergeh (vfa, vfb );
+ vfl = vec_mergel (vfa, vfb );
+
+ check (vec_all_eq (vuch, vucrh), "vuch");
+ check (vec_all_eq (vucl, vucrl), "vucl");
+ check (vec_all_eq (vsch, vscrh), "vsch");
+ check (vec_all_eq (vscl, vscrl), "vscl");
+ check (vec_all_eq (vush, vusrh), "vush");
+ check (vec_all_eq (vusl, vusrl), "vusl");
+ check (vec_all_eq (vssh, vssrh), "vssh");
+ check (vec_all_eq (vssl, vssrl), "vssl");
+ check (vec_all_eq (vuih, vuirh), "vuih");
+ check (vec_all_eq (vuil, vuirl), "vuil");
+ check (vec_all_eq (vsih, vsirh), "vsih");
+ check (vec_all_eq (vsil, vsirl), "vsil");
+ check (vec_all_eq (vfh, vfrh), "vfh");
+ check (vec_all_eq (vfl, vfrl), "vfl");
+}
===================================================================
@@ -0,0 +1,77 @@
+#include "harness.h"
+
+static void test()
+{
+ /* Input vectors. */
+ vector unsigned char vuca = {0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15};
+ vector unsigned char vucb
+ = {16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31};
+ vector signed char vsca
+ = {-16,-15,-14,-13,-12,-11,-10,-9,-8,-7,-6,-5,-4,-3,-2,-1};
+ vector signed char vscb = {0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15};
+ vector unsigned short vusa = {0,1,2,3,4,5,6,7};
+ vector unsigned short vusb = {8,9,10,11,12,13,14,15};
+ vector signed short vssa = {-8,-7,-6,-5,-4,-3,-2,-1};
+ vector signed short vssb = {0,1,2,3,4,5,6,7};
+ vector unsigned int vuia = {0,1,2,3};
+ vector unsigned int vuib = {4,5,6,7};
+ vector signed int vsia = {-4,-3,-2,-1};
+ vector signed int vsib = {0,1,2,3};
+ vector float vfa = {-4.0,-3.0,-2.0,-1.0};
+ vector float vfb = {0.0,1.0,2.0,3.0};
+
+ /* Result vectors. */
+ vector unsigned char vuch, vucl;
+ vector signed char vsch, vscl;
+ vector unsigned short vush, vusl;
+ vector signed short vssh, vssl;
+ vector unsigned int vuih, vuil;
+ vector signed int vsih, vsil;
+ vector float vfh, vfl;
+
+ /* Expected result vectors. */
+ vector unsigned char vucrh = {0,16,1,17,2,18,3,19,4,20,5,21,6,22,7,23};
+ vector unsigned char vucrl = {8,24,9,25,10,26,11,27,12,28,13,29,14,30,15,31};
+ vector signed char vscrh = {-16,0,-15,1,-14,2,-13,3,-12,4,-11,5,-10,6,-9,7};
+ vector signed char vscrl = {-8,8,-7,9,-6,10,-5,11,-4,12,-3,13,-2,14,-1,15};
+ vector unsigned short vusrh = {0,8,1,9,2,10,3,11};
+ vector unsigned short vusrl = {4,12,5,13,6,14,7,15};
+ vector signed short vssrh = {-8,0,-7,1,-6,2,-5,3};
+ vector signed short vssrl = {-4,4,-3,5,-2,6,-1,7};
+ vector unsigned int vuirh = {0,4,1,5};
+ vector unsigned int vuirl = {2,6,3,7};
+ vector signed int vsirh = {-4,0,-3,1};
+ vector signed int vsirl = {-2,2,-1,3};
+ vector float vfrh = {-4.0,0.0,-3.0,1.0};
+ vector float vfrl = {-2.0,2.0,-1.0,3.0};
+
+ vuch = vec_mergeh (vuca, vucb);
+ vucl = vec_mergel (vuca, vucb);
+ vsch = vec_mergeh (vsca, vscb);
+ vscl = vec_mergel (vsca, vscb);
+ vush = vec_mergeh (vusa, vusb);
+ vusl = vec_mergel (vusa, vusb);
+ vssh = vec_mergeh (vssa, vssb);
+ vssl = vec_mergel (vssa, vssb);
+ vuih = vec_mergeh (vuia, vuib);
+ vuil = vec_mergel (vuia, vuib);
+ vsih = vec_mergeh (vsia, vsib);
+ vsil = vec_mergel (vsia, vsib);
+ vfh = vec_mergeh (vfa, vfb );
+ vfl = vec_mergel (vfa, vfb );
+
+ check (vec_all_eq (vuch, vucrh), "vuch");
+ check (vec_all_eq (vucl, vucrl), "vucl");
+ check (vec_all_eq (vsch, vscrh), "vsch");
+ check (vec_all_eq (vscl, vscrl), "vscl");
+ check (vec_all_eq (vush, vusrh), "vush");
+ check (vec_all_eq (vusl, vusrl), "vusl");
+ check (vec_all_eq (vssh, vssrh), "vssh");
+ check (vec_all_eq (vssl, vssrl), "vssl");
+ check (vec_all_eq (vuih, vuirh), "vuih");
+ check (vec_all_eq (vuil, vuirl), "vuil");
+ check (vec_all_eq (vsih, vsirh), "vsih");
+ check (vec_all_eq (vsil, vsirl), "vsil");
+ check (vec_all_eq (vfh, vfrh), "vfh");
+ check (vec_all_eq (vfl, vfrl), "vfl");
+}
===================================================================
@@ -0,0 +1,46 @@
+/* { dg-skip-if "" { powerpc*-*-darwin* } { "*" } { "" } } */
+/* { dg-require-effective-target powerpc_vsx_ok } */
+/* { dg-options "-maltivec=be -mabi=altivec -std=gnu99 -mvsx" } */
+
+#include "harness.h"
+
+static int vec_long_eq (vector long x, vector long y)
+{
+ return (x[0] == y[0] && x[1] == y[1]);
+}
+
+static void test()
+{
+ /* Input vectors. */
+ vector long vla = {-2,-1};
+ vector long vlb = {0,1};
+ vector double vda = {-2.0,-1.0};
+ vector double vdb = {0.0,1.0};
+
+ /* Result vectors. */
+ vector long vlh, vll;
+ vector double vdh, vdl;
+
+ /* Expected result vectors. */
+#if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
+ vector long vlrh = {1,-1};
+ vector long vlrl = {0,-2};
+ vector double vdrh = {1.0,-1.0};
+ vector double vdrl = {0.0,-2.0};
+#else
+ vector long vlrh = {-2,0};
+ vector long vlrl = {-1,1};
+ vector double vdrh = {-2.0,0.0};
+ vector double vdrl = {-1.0,1.0};
+#endif
+
+ vlh = vec_mergeh (vla, vlb);
+ vll = vec_mergel (vla, vlb);
+ vdh = vec_mergeh (vda, vdb);
+ vdl = vec_mergel (vda, vdb);
+
+ check (vec_long_eq (vlh, vlrh), "vlh");
+ check (vec_long_eq (vll, vlrl), "vll");
+ check (vec_all_eq (vdh, vdrh), "vdh" );
+ check (vec_all_eq (vdl, vdrl), "vdl" );
+}
===================================================================
@@ -0,0 +1,39 @@
+/* { dg-skip-if "" { powerpc*-*-darwin* } { "*" } { "" } } */
+/* { dg-require-effective-target powerpc_vsx_ok } */
+/* { dg-options "-maltivec -mabi=altivec -std=gnu99 -mvsx" } */
+
+#include "harness.h"
+
+static int vec_long_eq (vector long x, vector long y)
+{
+ return (x[0] == y[0] && x[1] == y[1]);
+}
+
+static void test()
+{
+ /* Input vectors. */
+ vector long vla = {-2,-1};
+ vector long vlb = {0,1};
+ vector double vda = {-2.0,-1.0};
+ vector double vdb = {0.0,1.0};
+
+ /* Result vectors. */
+ vector long vlh, vll;
+ vector double vdh, vdl;
+
+ /* Expected result vectors. */
+ vector long vlrh = {-2,0};
+ vector long vlrl = {-1,1};
+ vector double vdrh = {-2.0,0.0};
+ vector double vdrl = {-1.0,1.0};
+
+ vlh = vec_mergeh (vla, vlb);
+ vll = vec_mergel (vla, vlb);
+ vdh = vec_mergeh (vda, vdb);
+ vdl = vec_mergel (vda, vdb);
+
+ check (vec_long_eq (vlh, vlrh), "vlh");
+ check (vec_long_eq (vll, vlrl), "vll");
+ check (vec_all_eq (vdh, vdrh), "vdh" );
+ check (vec_all_eq (vdl, vdrl), "vdl" );
+}
===================================================================
@@ -29895,22 +29895,28 @@ altivec_expand_vec_perm_const (rtx operands[4])
{ OPTION_MASK_ALTIVEC, CODE_FOR_altivec_vpkuwum,
{ 2, 3, 6, 7, 10, 11, 14, 15, 18, 19, 22, 23, 26, 27, 30, 31 } },
{ OPTION_MASK_ALTIVEC,
- BYTES_BIG_ENDIAN ? CODE_FOR_altivec_vmrghb : CODE_FOR_altivec_vmrglb,
+ (BYTES_BIG_ENDIAN ? CODE_FOR_altivec_vmrghb_direct
+ : CODE_FOR_altivec_vmrglb_direct),
{ 0, 16, 1, 17, 2, 18, 3, 19, 4, 20, 5, 21, 6, 22, 7, 23 } },
{ OPTION_MASK_ALTIVEC,
- BYTES_BIG_ENDIAN ? CODE_FOR_altivec_vmrghh : CODE_FOR_altivec_vmrglh,
+ (BYTES_BIG_ENDIAN ? CODE_FOR_altivec_vmrghh_direct
+ : CODE_FOR_altivec_vmrglh_direct),
{ 0, 1, 16, 17, 2, 3, 18, 19, 4, 5, 20, 21, 6, 7, 22, 23 } },
{ OPTION_MASK_ALTIVEC,
- BYTES_BIG_ENDIAN ? CODE_FOR_altivec_vmrghw : CODE_FOR_altivec_vmrglw,
+ (BYTES_BIG_ENDIAN ? CODE_FOR_altivec_vmrghw_direct
+ : CODE_FOR_altivec_vmrglw_direct),
{ 0, 1, 2, 3, 16, 17, 18, 19, 4, 5, 6, 7, 20, 21, 22, 23 } },
{ OPTION_MASK_ALTIVEC,
- BYTES_BIG_ENDIAN ? CODE_FOR_altivec_vmrglb : CODE_FOR_altivec_vmrghb,
+ (BYTES_BIG_ENDIAN ? CODE_FOR_altivec_vmrglb_direct
+ : CODE_FOR_altivec_vmrghb_direct),
{ 8, 24, 9, 25, 10, 26, 11, 27, 12, 28, 13, 29, 14, 30, 15, 31 } },
{ OPTION_MASK_ALTIVEC,
- BYTES_BIG_ENDIAN ? CODE_FOR_altivec_vmrglh : CODE_FOR_altivec_vmrghh,
+ (BYTES_BIG_ENDIAN ? CODE_FOR_altivec_vmrglh_direct
+ : CODE_FOR_altivec_vmrghh_direct),
{ 8, 9, 24, 25, 10, 11, 26, 27, 12, 13, 28, 29, 14, 15, 30, 31 } },
{ OPTION_MASK_ALTIVEC,
- BYTES_BIG_ENDIAN ? CODE_FOR_altivec_vmrglw : CODE_FOR_altivec_vmrghw,
+ (BYTES_BIG_ENDIAN ? CODE_FOR_altivec_vmrglw_direct
+ : CODE_FOR_altivec_vmrghw_direct),
{ 8, 9, 10, 11, 24, 25, 26, 27, 12, 13, 14, 15, 28, 29, 30, 31 } },
{ OPTION_MASK_P8_VECTOR, CODE_FOR_p8_vmrgew,
{ 0, 1, 2, 3, 16, 17, 18, 19, 8, 9, 10, 11, 24, 25, 26, 27 } },
===================================================================
@@ -1678,25 +1678,55 @@
;; Expanders for builtins
(define_expand "vsx_mergel_<mode>"
- [(set (match_operand:VSX_D 0 "vsx_register_operand" "")
- (vec_select:VSX_D
- (vec_concat:<VS_double>
- (match_operand:VSX_D 1 "vsx_register_operand" "")
- (match_operand:VSX_D 2 "vsx_register_operand" ""))
- (parallel [(const_int 1) (const_int 3)])))]
+ [(use (match_operand:VSX_D 0 "vsx_register_operand" ""))
+ (use (match_operand:VSX_D 1 "vsx_register_operand" ""))
+ (use (match_operand:VSX_D 2 "vsx_register_operand" ""))]
"VECTOR_MEM_VSX_P (<MODE>mode)"
- "")
+{
+ rtvec v;
+ rtx x;
+ /* Special handling for LE with -maltivec=be. */
+ if (!BYTES_BIG_ENDIAN && VECTOR_ELT_ORDER_BIG)
+ {
+ v = gen_rtvec (2, GEN_INT (0), GEN_INT (2));
+ x = gen_rtx_VEC_CONCAT (<VS_double>mode, operands[2], operands[1]);
+ }
+ else
+ {
+ v = gen_rtvec (2, GEN_INT (1), GEN_INT (3));
+ x = gen_rtx_VEC_CONCAT (<VS_double>mode, operands[1], operands[2]);
+ }
+
+ x = gen_rtx_VEC_SELECT (<MODE>mode, x, gen_rtx_PARALLEL (VOIDmode, v));
+ emit_insn (gen_rtx_SET (VOIDmode, operands[0], x));
+})
+
(define_expand "vsx_mergeh_<mode>"
- [(set (match_operand:VSX_D 0 "vsx_register_operand" "")
- (vec_select:VSX_D
- (vec_concat:<VS_double>
- (match_operand:VSX_D 1 "vsx_register_operand" "")
- (match_operand:VSX_D 2 "vsx_register_operand" ""))
- (parallel [(const_int 0) (const_int 2)])))]
+ [(use (match_operand:VSX_D 0 "vsx_register_operand" ""))
+ (use (match_operand:VSX_D 1 "vsx_register_operand" ""))
+ (use (match_operand:VSX_D 2 "vsx_register_operand" ""))]
"VECTOR_MEM_VSX_P (<MODE>mode)"
- "")
+{
+ rtvec v;
+ rtx x;
+ /* Special handling for LE with -maltivec=be. */
+ if (!BYTES_BIG_ENDIAN && VECTOR_ELT_ORDER_BIG)
+ {
+ v = gen_rtvec (2, GEN_INT (1), GEN_INT (3));
+ x = gen_rtx_VEC_CONCAT (<VS_double>mode, operands[2], operands[1]);
+ }
+ else
+ {
+ v = gen_rtvec (2, GEN_INT (0), GEN_INT (2));
+ x = gen_rtx_VEC_CONCAT (<VS_double>mode, operands[1], operands[2]);
+ }
+
+ x = gen_rtx_VEC_SELECT (<MODE>mode, x, gen_rtx_PARALLEL (VOIDmode, v));
+ emit_insn (gen_rtx_SET (VOIDmode, operands[0], x));
+})
+
;; V2DF/V2DI splat
(define_insn "vsx_splat_<mode>"
[(set (match_operand:VSX_D 0 "vsx_register_operand" "=wd,wd,wd,?wa,?wa,?wa")
===================================================================
@@ -129,6 +129,9 @@
UNSPEC_VUPKHU_V4SF
UNSPEC_VUPKLU_V4SF
UNSPEC_VGBBD
+ UNSPEC_VMRGH_DIRECT
+ UNSPEC_VMRGL_DIRECT
+ UNSPEC_VSPLT_DIRECT
])
(define_c_enum "unspecv"
@@ -677,16 +680,16 @@
{
emit_insn (gen_altivec_vmulesh (even, operands[1], operands[2]));
emit_insn (gen_altivec_vmulosh (odd, operands[1], operands[2]));
- emit_insn (gen_altivec_vmrghw (high, even, odd));
- emit_insn (gen_altivec_vmrglw (low, even, odd));
+ emit_insn (gen_altivec_vmrghw_direct (high, even, odd));
+ emit_insn (gen_altivec_vmrglw_direct (low, even, odd));
emit_insn (gen_altivec_vpkuwum (operands[0], high, low));
}
else
{
emit_insn (gen_altivec_vmulosh (even, operands[1], operands[2]));
emit_insn (gen_altivec_vmulesh (odd, operands[1], operands[2]));
- emit_insn (gen_altivec_vmrghw (high, odd, even));
- emit_insn (gen_altivec_vmrglw (low, odd, even));
+ emit_insn (gen_altivec_vmrghw_direct (high, odd, even));
+ emit_insn (gen_altivec_vmrglw_direct (low, odd, even));
emit_insn (gen_altivec_vpkuwum (operands[0], low, high));
}
@@ -839,9 +842,40 @@
"vmladduhm %0,%1,%2,%3"
[(set_attr "type" "veccomplex")])
-(define_insn "altivec_vmrghb"
+(define_expand "altivec_vmrghb"
+ [(use (match_operand:V16QI 0 "register_operand" ""))
+ (use (match_operand:V16QI 1 "register_operand" ""))
+ (use (match_operand:V16QI 2 "register_operand" ""))]
+ "TARGET_ALTIVEC"
+{
+ rtvec v;
+ rtx x;
+
+ /* Special handling for LE with -maltivec=be. */
+ if (!BYTES_BIG_ENDIAN && VECTOR_ELT_ORDER_BIG)
+ {
+ v = gen_rtvec (16, GEN_INT (8), GEN_INT (24), GEN_INT (9), GEN_INT (25),
+ GEN_INT (10), GEN_INT (26), GEN_INT (11), GEN_INT (27),
+ GEN_INT (12), GEN_INT (28), GEN_INT (13), GEN_INT (29),
+ GEN_INT (14), GEN_INT (30), GEN_INT (15), GEN_INT (31));
+ x = gen_rtx_VEC_CONCAT (V32QImode, operands[2], operands[1]);
+ }
+ else
+ {
+ v = gen_rtvec (16, GEN_INT (0), GEN_INT (16), GEN_INT (1), GEN_INT (17),
+ GEN_INT (2), GEN_INT (18), GEN_INT (3), GEN_INT (19),
+ GEN_INT (4), GEN_INT (20), GEN_INT (5), GEN_INT (21),
+ GEN_INT (6), GEN_INT (22), GEN_INT (7), GEN_INT (23));
+ x = gen_rtx_VEC_CONCAT (V32QImode, operands[1], operands[2]);
+ }
+
+ x = gen_rtx_VEC_SELECT (V16QImode, x, gen_rtx_PARALLEL (VOIDmode, v));
+ emit_insn (gen_rtx_SET (VOIDmode, operands[0], x));
+})
+
+(define_insn "*altivec_vmrghb_internal"
[(set (match_operand:V16QI 0 "register_operand" "=v")
- (vec_select:V16QI
+ (vec_select:V16QI
(vec_concat:V32QI
(match_operand:V16QI 1 "register_operand" "v")
(match_operand:V16QI 2 "register_operand" "v"))
@@ -854,12 +888,53 @@
(const_int 6) (const_int 22)
(const_int 7) (const_int 23)])))]
"TARGET_ALTIVEC"
+{
+ if (BYTES_BIG_ENDIAN)
+ return "vmrghb %0,%1,%2";
+ else
+ return "vmrglb %0,%2,%1";
+}
+ [(set_attr "type" "vecperm")])
+
+(define_insn "altivec_vmrghb_direct"
+ [(set (match_operand:V16QI 0 "register_operand" "=v")
+ (unspec:V16QI [(match_operand:V16QI 1 "register_operand" "v")
+ (match_operand:V16QI 2 "register_operand" "v")]
+ UNSPEC_VMRGH_DIRECT))]
+ "TARGET_ALTIVEC"
"vmrghb %0,%1,%2"
[(set_attr "type" "vecperm")])
-(define_insn "altivec_vmrghh"
+(define_expand "altivec_vmrghh"
+ [(use (match_operand:V8HI 0 "register_operand" ""))
+ (use (match_operand:V8HI 1 "register_operand" ""))
+ (use (match_operand:V8HI 2 "register_operand" ""))]
+ "TARGET_ALTIVEC"
+{
+ rtvec v;
+ rtx x;
+
+ /* Special handling for LE with -maltivec=be. */
+ if (!BYTES_BIG_ENDIAN && VECTOR_ELT_ORDER_BIG)
+ {
+ v = gen_rtvec (8, GEN_INT (4), GEN_INT (12), GEN_INT (5), GEN_INT (13),
+ GEN_INT (6), GEN_INT (14), GEN_INT (7), GEN_INT (15));
+ x = gen_rtx_VEC_CONCAT (V16HImode, operands[2], operands[1]);
+ }
+ else
+ {
+ v = gen_rtvec (8, GEN_INT (0), GEN_INT (8), GEN_INT (1), GEN_INT (9),
+ GEN_INT (2), GEN_INT (10), GEN_INT (3), GEN_INT (11));
+ x = gen_rtx_VEC_CONCAT (V16HImode, operands[1], operands[2]);
+ }
+
+ x = gen_rtx_VEC_SELECT (V8HImode, x, gen_rtx_PARALLEL (VOIDmode, v));
+ emit_insn (gen_rtx_SET (VOIDmode, operands[0], x));
+})
+
+(define_insn "*altivec_vmrghh_internal"
[(set (match_operand:V8HI 0 "register_operand" "=v")
- (vec_select:V8HI
+ (vec_select:V8HI
(vec_concat:V16HI
(match_operand:V8HI 1 "register_operand" "v")
(match_operand:V8HI 2 "register_operand" "v"))
@@ -868,10 +943,49 @@
(const_int 2) (const_int 10)
(const_int 3) (const_int 11)])))]
"TARGET_ALTIVEC"
+{
+ if (BYTES_BIG_ENDIAN)
+ return "vmrghh %0,%1,%2";
+ else
+ return "vmrglh %0,%2,%1";
+}
+ [(set_attr "type" "vecperm")])
+
+(define_insn "altivec_vmrghh_direct"
+ [(set (match_operand:V8HI 0 "register_operand" "=v")
+ (unspec:V8HI [(match_operand:V8HI 1 "register_operand" "v")
+ (match_operand:V8HI 2 "register_operand" "v")]
+ UNSPEC_VMRGH_DIRECT))]
+ "TARGET_ALTIVEC"
"vmrghh %0,%1,%2"
[(set_attr "type" "vecperm")])
-(define_insn "altivec_vmrghw"
+(define_expand "altivec_vmrghw"
+ [(use (match_operand:V4SI 0 "register_operand" ""))
+ (use (match_operand:V4SI 1 "register_operand" ""))
+ (use (match_operand:V4SI 2 "register_operand" ""))]
+ "VECTOR_MEM_ALTIVEC_P (V4SImode)"
+{
+ rtvec v;
+ rtx x;
+
+ /* Special handling for LE with -maltivec=be. */
+ if (!BYTES_BIG_ENDIAN && VECTOR_ELT_ORDER_BIG)
+ {
+ v = gen_rtvec (4, GEN_INT (2), GEN_INT (6), GEN_INT (3), GEN_INT (7));
+ x = gen_rtx_VEC_CONCAT (V8SImode, operands[2], operands[1]);
+ }
+ else
+ {
+ v = gen_rtvec (4, GEN_INT (0), GEN_INT (4), GEN_INT (1), GEN_INT (5));
+ x = gen_rtx_VEC_CONCAT (V8SImode, operands[1], operands[2]);
+ }
+
+ x = gen_rtx_VEC_SELECT (V4SImode, x, gen_rtx_PARALLEL (VOIDmode, v));
+ emit_insn (gen_rtx_SET (VOIDmode, operands[0], x));
+})
+
+(define_insn "*altivec_vmrghw_internal"
[(set (match_operand:V4SI 0 "register_operand" "=v")
(vec_select:V4SI
(vec_concat:V8SI
@@ -880,6 +994,20 @@
(parallel [(const_int 0) (const_int 4)
(const_int 1) (const_int 5)])))]
"VECTOR_MEM_ALTIVEC_P (V4SImode)"
+{
+ if (BYTES_BIG_ENDIAN)
+ return "vmrghw %0,%1,%2";
+ else
+ return "vmrglw %0,%2,%1";
+}
+ [(set_attr "type" "vecperm")])
+
+(define_insn "altivec_vmrghw_direct"
+ [(set (match_operand:V4SI 0 "register_operand" "=v")
+ (unspec:V4SI [(match_operand:V4SI 1 "register_operand" "v")
+ (match_operand:V4SI 2 "register_operand" "v")]
+ UNSPEC_VMRGH_DIRECT))]
+ "TARGET_ALTIVEC"
"vmrghw %0,%1,%2"
[(set_attr "type" "vecperm")])
@@ -892,10 +1020,46 @@
(parallel [(const_int 0) (const_int 4)
(const_int 1) (const_int 5)])))]
"VECTOR_MEM_ALTIVEC_P (V4SFmode)"
- "vmrghw %0,%1,%2"
+{
+ if (BYTES_BIG_ENDIAN)
+ return "vmrghw %0,%1,%2";
+ else
+ return "vmrglw %0,%2,%1";
+}
[(set_attr "type" "vecperm")])
-(define_insn "altivec_vmrglb"
+(define_expand "altivec_vmrglb"
+ [(use (match_operand:V16QI 0 "register_operand" ""))
+ (use (match_operand:V16QI 1 "register_operand" ""))
+ (use (match_operand:V16QI 2 "register_operand" ""))]
+ "TARGET_ALTIVEC"
+{
+ rtvec v;
+ rtx x;
+
+ /* Special handling for LE with -maltivec=be. */
+ if (!BYTES_BIG_ENDIAN && VECTOR_ELT_ORDER_BIG)
+ {
+ v = gen_rtvec (16, GEN_INT (0), GEN_INT (16), GEN_INT (1), GEN_INT (17),
+ GEN_INT (2), GEN_INT (18), GEN_INT (3), GEN_INT (19),
+ GEN_INT (4), GEN_INT (20), GEN_INT (5), GEN_INT (21),
+ GEN_INT (6), GEN_INT (22), GEN_INT (7), GEN_INT (23));
+ x = gen_rtx_VEC_CONCAT (V32QImode, operands[2], operands[1]);
+ }
+ else
+ {
+ v = gen_rtvec (16, GEN_INT (8), GEN_INT (24), GEN_INT (9), GEN_INT (25),
+ GEN_INT (10), GEN_INT (26), GEN_INT (11), GEN_INT (27),
+ GEN_INT (12), GEN_INT (28), GEN_INT (13), GEN_INT (29),
+ GEN_INT (14), GEN_INT (30), GEN_INT (15), GEN_INT (31));
+ x = gen_rtx_VEC_CONCAT (V32QImode, operands[1], operands[2]);
+ }
+
+ x = gen_rtx_VEC_SELECT (V16QImode, x, gen_rtx_PARALLEL (VOIDmode, v));
+ emit_insn (gen_rtx_SET (VOIDmode, operands[0], x));
+})
+
+(define_insn "*altivec_vmrglb_internal"
[(set (match_operand:V16QI 0 "register_operand" "=v")
(vec_select:V16QI
(vec_concat:V32QI
@@ -910,10 +1074,51 @@
(const_int 14) (const_int 30)
(const_int 15) (const_int 31)])))]
"TARGET_ALTIVEC"
+{
+ if (BYTES_BIG_ENDIAN)
+ return "vmrglb %0,%1,%2";
+ else
+ return "vmrghb %0,%2,%1";
+}
+ [(set_attr "type" "vecperm")])
+
+(define_insn "altivec_vmrglb_direct"
+ [(set (match_operand:V16QI 0 "register_operand" "=v")
+ (unspec:V16QI [(match_operand:V16QI 1 "register_operand" "v")
+ (match_operand:V16QI 2 "register_operand" "v")]
+ UNSPEC_VMRGL_DIRECT))]
+ "TARGET_ALTIVEC"
"vmrglb %0,%1,%2"
[(set_attr "type" "vecperm")])
-(define_insn "altivec_vmrglh"
+(define_expand "altivec_vmrglh"
+ [(use (match_operand:V8HI 0 "register_operand" ""))
+ (use (match_operand:V8HI 1 "register_operand" ""))
+ (use (match_operand:V8HI 2 "register_operand" ""))]
+ "TARGET_ALTIVEC"
+{
+ rtvec v;
+ rtx x;
+
+ /* Special handling for LE with -maltivec=be. */
+ if (!BYTES_BIG_ENDIAN && VECTOR_ELT_ORDER_BIG)
+ {
+ v = gen_rtvec (8, GEN_INT (0), GEN_INT (8), GEN_INT (1), GEN_INT (9),
+ GEN_INT (2), GEN_INT (10), GEN_INT (3), GEN_INT (11));
+ x = gen_rtx_VEC_CONCAT (V16HImode, operands[2], operands[1]);
+ }
+ else
+ {
+ v = gen_rtvec (8, GEN_INT (4), GEN_INT (12), GEN_INT (5), GEN_INT (13),
+ GEN_INT (6), GEN_INT (14), GEN_INT (7), GEN_INT (15));
+ x = gen_rtx_VEC_CONCAT (V16HImode, operands[1], operands[2]);
+ }
+
+ x = gen_rtx_VEC_SELECT (V8HImode, x, gen_rtx_PARALLEL (VOIDmode, v));
+ emit_insn (gen_rtx_SET (VOIDmode, operands[0], x));
+})
+
+(define_insn "*altivec_vmrglh_internal"
[(set (match_operand:V8HI 0 "register_operand" "=v")
(vec_select:V8HI
(vec_concat:V16HI
@@ -924,10 +1129,49 @@
(const_int 6) (const_int 14)
(const_int 7) (const_int 15)])))]
"TARGET_ALTIVEC"
+{
+ if (BYTES_BIG_ENDIAN)
+ return "vmrglh %0,%1,%2";
+ else
+ return "vmrghh %0,%2,%1";
+}
+ [(set_attr "type" "vecperm")])
+
+(define_insn "altivec_vmrglh_direct"
+ [(set (match_operand:V8HI 0 "register_operand" "=v")
+ (unspec:V8HI [(match_operand:V8HI 1 "register_operand" "v")
+ (match_operand:V8HI 2 "register_operand" "v")]
+ UNSPEC_VMRGL_DIRECT))]
+ "TARGET_ALTIVEC"
"vmrglh %0,%1,%2"
[(set_attr "type" "vecperm")])
-(define_insn "altivec_vmrglw"
+(define_expand "altivec_vmrglw"
+ [(use (match_operand:V4SI 0 "register_operand" ""))
+ (use (match_operand:V4SI 1 "register_operand" ""))
+ (use (match_operand:V4SI 2 "register_operand" ""))]
+ "VECTOR_MEM_ALTIVEC_P (V4SImode)"
+{
+ rtvec v;
+ rtx x;
+
+ /* Special handling for LE with -maltivec=be. */
+ if (!BYTES_BIG_ENDIAN && VECTOR_ELT_ORDER_BIG)
+ {
+ v = gen_rtvec (4, GEN_INT (0), GEN_INT (4), GEN_INT (1), GEN_INT (5));
+ x = gen_rtx_VEC_CONCAT (V8SImode, operands[2], operands[1]);
+ }
+ else
+ {
+ v = gen_rtvec (4, GEN_INT (2), GEN_INT (6), GEN_INT (3), GEN_INT (7));
+ x = gen_rtx_VEC_CONCAT (V8SImode, operands[1], operands[2]);
+ }
+
+ x = gen_rtx_VEC_SELECT (V4SImode, x, gen_rtx_PARALLEL (VOIDmode, v));
+ emit_insn (gen_rtx_SET (VOIDmode, operands[0], x));
+})
+
+(define_insn "*altivec_vmrglw_internal"
[(set (match_operand:V4SI 0 "register_operand" "=v")
(vec_select:V4SI
(vec_concat:V8SI
@@ -936,6 +1180,20 @@
(parallel [(const_int 2) (const_int 6)
(const_int 3) (const_int 7)])))]
"VECTOR_MEM_ALTIVEC_P (V4SImode)"
+{
+ if (BYTES_BIG_ENDIAN)
+ return "vmrglw %0,%1,%2";
+ else
+ return "vmrghw %0,%2,%1";
+}
+ [(set_attr "type" "vecperm")])
+
+(define_insn "altivec_vmrglw_direct"
+ [(set (match_operand:V4SI 0 "register_operand" "=v")
+ (unspec:V4SI [(match_operand:V4SI 1 "register_operand" "v")
+ (match_operand:V4SI 2 "register_operand" "v")]
+ UNSPEC_VMRGL_DIRECT))]
+ "TARGET_ALTIVEC"
"vmrglw %0,%1,%2"
[(set_attr "type" "vecperm")])
@@ -948,7 +1206,12 @@
(parallel [(const_int 2) (const_int 6)
(const_int 3) (const_int 7)])))]
"VECTOR_MEM_ALTIVEC_P (V4SFmode)"
- "vmrglw %0,%1,%2"
+{
+ if (BYTES_BIG_ENDIAN)
+ return "vmrglw %0,%1,%2";
+ else
+ return "vmrghw %0,%2,%1";
+}
[(set_attr "type" "vecperm")])
;; Power8 vector merge even/odd
@@ -2225,13 +2488,13 @@
{
emit_insn (gen_altivec_vmuleub (ve, operands[1], operands[2]));
emit_insn (gen_altivec_vmuloub (vo, operands[1], operands[2]));
- emit_insn (gen_altivec_vmrghh (operands[0], ve, vo));
+ emit_insn (gen_altivec_vmrghh_direct (operands[0], ve, vo));
}
else
{
emit_insn (gen_altivec_vmuloub (ve, operands[1], operands[2]));
emit_insn (gen_altivec_vmuleub (vo, operands[1], operands[2]));
- emit_insn (gen_altivec_vmrghh (operands[0], vo, ve));
+ emit_insn (gen_altivec_vmrghh_direct (operands[0], vo, ve));
}
DONE;
}")
@@ -2251,13 +2514,13 @@
{
emit_insn (gen_altivec_vmuleub (ve, operands[1], operands[2]));
emit_insn (gen_altivec_vmuloub (vo, operands[1], operands[2]));
- emit_insn (gen_altivec_vmrglh (operands[0], ve, vo));
+ emit_insn (gen_altivec_vmrglh_direct (operands[0], ve, vo));
}
else
{
emit_insn (gen_altivec_vmuloub (ve, operands[1], operands[2]));
emit_insn (gen_altivec_vmuleub (vo, operands[1], operands[2]));
- emit_insn (gen_altivec_vmrglh (operands[0], vo, ve));
+ emit_insn (gen_altivec_vmrglh_direct (operands[0], vo, ve));
}
DONE;
}")
@@ -2277,13 +2540,13 @@
{
emit_insn (gen_altivec_vmulesb (ve, operands[1], operands[2]));
emit_insn (gen_altivec_vmulosb (vo, operands[1], operands[2]));
- emit_insn (gen_altivec_vmrghh (operands[0], ve, vo));
+ emit_insn (gen_altivec_vmrghh_direct (operands[0], ve, vo));
}
else
{
emit_insn (gen_altivec_vmulosb (ve, operands[1], operands[2]));
emit_insn (gen_altivec_vmulesb (vo, operands[1], operands[2]));
- emit_insn (gen_altivec_vmrghh (operands[0], vo, ve));
+ emit_insn (gen_altivec_vmrghh_direct (operands[0], vo, ve));
}
DONE;
}")
@@ -2303,13 +2566,13 @@
{
emit_insn (gen_altivec_vmulesb (ve, operands[1], operands[2]));
emit_insn (gen_altivec_vmulosb (vo, operands[1], operands[2]));
- emit_insn (gen_altivec_vmrglh (operands[0], ve, vo));
+ emit_insn (gen_altivec_vmrglh_direct (operands[0], ve, vo));
}
else
{
emit_insn (gen_altivec_vmulosb (ve, operands[1], operands[2]));
emit_insn (gen_altivec_vmulesb (vo, operands[1], operands[2]));
- emit_insn (gen_altivec_vmrglh (operands[0], vo, ve));
+ emit_insn (gen_altivec_vmrglh_direct (operands[0], vo, ve));
}
DONE;
}")
@@ -2329,13 +2592,13 @@
{
emit_insn (gen_altivec_vmuleuh (ve, operands[1], operands[2]));
emit_insn (gen_altivec_vmulouh (vo, operands[1], operands[2]));
- emit_insn (gen_altivec_vmrghw (operands[0], ve, vo));
+ emit_insn (gen_altivec_vmrghw_direct (operands[0], ve, vo));
}
else
{
emit_insn (gen_altivec_vmulouh (ve, operands[1], operands[2]));
emit_insn (gen_altivec_vmuleuh (vo, operands[1], operands[2]));
- emit_insn (gen_altivec_vmrghw (operands[0], vo, ve));
+ emit_insn (gen_altivec_vmrghw_direct (operands[0], vo, ve));
}
DONE;
}")
@@ -2355,13 +2618,13 @@
{
emit_insn (gen_altivec_vmuleuh (ve, operands[1], operands[2]));
emit_insn (gen_altivec_vmulouh (vo, operands[1], operands[2]));
- emit_insn (gen_altivec_vmrglw (operands[0], ve, vo));
+ emit_insn (gen_altivec_vmrglw_direct (operands[0], ve, vo));
}
else
{
emit_insn (gen_altivec_vmulouh (ve, operands[1], operands[2]));
emit_insn (gen_altivec_vmuleuh (vo, operands[1], operands[2]));
- emit_insn (gen_altivec_vmrglw (operands[0], vo, ve));
+ emit_insn (gen_altivec_vmrglw_direct (operands[0], vo, ve));
}
DONE;
}")
@@ -2381,13 +2644,13 @@
{
emit_insn (gen_altivec_vmulesh (ve, operands[1], operands[2]));
emit_insn (gen_altivec_vmulosh (vo, operands[1], operands[2]));
- emit_insn (gen_altivec_vmrghw (operands[0], ve, vo));
+ emit_insn (gen_altivec_vmrghw_direct (operands[0], ve, vo));
}
else
{
emit_insn (gen_altivec_vmulosh (ve, operands[1], operands[2]));
emit_insn (gen_altivec_vmulesh (vo, operands[1], operands[2]));
- emit_insn (gen_altivec_vmrghw (operands[0], vo, ve));
+ emit_insn (gen_altivec_vmrghw_direct (operands[0], vo, ve));
}
DONE;
}")
@@ -2407,13 +2670,13 @@
{
emit_insn (gen_altivec_vmulesh (ve, operands[1], operands[2]));
emit_insn (gen_altivec_vmulosh (vo, operands[1], operands[2]));
- emit_insn (gen_altivec_vmrglw (operands[0], ve, vo));
+ emit_insn (gen_altivec_vmrglw_direct (operands[0], ve, vo));
}
else
{
emit_insn (gen_altivec_vmulosh (ve, operands[1], operands[2]));
emit_insn (gen_altivec_vmulesh (vo, operands[1], operands[2]));
- emit_insn (gen_altivec_vmrglw (operands[0], vo, ve));
+ emit_insn (gen_altivec_vmrglw_direct (operands[0], vo, ve));
}
DONE;
}")
Hi, David suggested privately that I rework some of the pattern names to fit in with existing practice. I've done this, and the result is below. Bootstrapped and tested on powerpc64{,le}-unknown-linux-gnu with no regressions. Is this ok for trunk? Thanks, Bill On Thu, 2014-01-23 at 18:08 -0600, Bill Schmidt wrote: > Hi, > > This patch continues the series of changes to the Altivec builtins to > accommodate big-endian element order when targeting a little endian > machine. Here the focus is on the vector merge-high and merge-low > operations. > > The primary change is in altivec.md. As an example, look at the pattern > altivec_vmrghw. Previously, this was represented with a single > define_insn. Now it's been split into a define_expand to create the > RTL, and a define_insn (*altivec_vmrghw_endian) to generate the hardware > instruction. This is because we need a different selection vector when > using -maltivec=be and targeting LE. (Normally LE and BE can use the > same selection vector, and GCC takes care of interpreting the indices as > left-to-right or right-to-left.) The new define_insn also substitutes > vmrglw with swapped operands for vmrghw for little-endian mode, since > the hardware instruction has "big-endian bias" in the interpretation of > "high" and "low." > > Because -maltivec=be applies only to programmer-specified builtins, we > need to adjust internal uses of altivec_vmrghw and friends. Thus we > have a new define_insn altivec_vmrghw_internal that generates the > hardware instruction directly with none of the above transformations. > New unspecs are needed for these internal forms. > > The VSX flavors of merge-high and merge-low are a little simpler (see > vsx.md). Here we already had a define_expand where the instructions are > generated by separate xxpermdi patterns, and there are no internal uses > to worry about. So we only need to change the selection vector in the > generated RTL. > > There are four new test cases that cover all of the supported data > types. Tests are divided between those that require only VMX > instructions and those that require VSX instructions. There are also > variants for -maltivec and -maltivec=be. > > Bootstrapped and tested on powerpc64{,le}-unknown-linux-gnu with no > regressions. Ok for trunk? > > Thanks, > Bill > gcc: 2014-01-28 Bill Schmidt <wschmidt@linux.vnet.ibm.com> * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*. * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for -maltivec=be with LE targets. (vsx_mergeh_<mode>): Likewise. * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs. (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct. (altivec_vmrghb): Replace with define_expand and new *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets. (altivec_vmrghb_direct): New define_insn. (altivec_vmrghh): Replace with define_expand and new *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets. (altivec_vmrghh_direct): New define_insn. (altivec_vmrghw): Replace with define_expand and new *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets. (altivec_vmrghw_direct): New define_insn. (*altivec_vmrghsf): Adjust for endianness. (altivec_vmrglb): Replace with define_expand and new *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets. (altivec_vmrglb_direct): New define_insn. (altivec_vmrglh): Replace with define_expand and new *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets. (altivec_vmrglh_direct): New define_insn. (altivec_vmrglw): Replace with define_expand and new *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets. (altivec_vmrglw_direct): New define_insn. (*altivec_vmrglsf): Adjust for endianness. (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct. (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct. (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct. (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct. (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct. (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct. (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct. (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct. gcc/testsuite: 2014-01-28 Bill Schmidt <wschmidt@linux.vnet.ibm.com> * gcc.dg/vmx/merge-be-order.c: New. * gcc.dg/vmx/merge.c: New. * gcc.dg/vmx/merge-vsx-be-order.c: New. * gcc.dg/vmx/merge-vsx.c: New.