[llvm] 666ee84 - [PowerPC] Fix shift amount of xxsldwi when performing vector int_to_double

[Kai Luo via llvm-commits]

Author: Kai Luo
Date: 2021-08-06T06:01:29Z
New Revision: 666ee849f0778160b4660acccb796ac5bd238b2d

URL: https://github.com/llvm/llvm-project/commit/666ee849f0778160b4660acccb796ac5bd238b2d
DIFF: https://github.com/llvm/llvm-project/commit/666ee849f0778160b4660acccb796ac5bd238b2d.diff

LOG: [PowerPC] Fix shift amount of xxsldwi when performing vector int_to_double

POC
```
// main.c
#include <stdio.h>
#include <altivec.h>
extern vector double foo(vector int s);
int main() {
  vector int s = {0, 1, 0, 4};
  vector double vd;
  vd = foo(s);
  printf("%lf %lf\n", vd[0], vd[1]);
  return 0;
}
// poc.c
vector double foo(vector int s) {
  int x1 = s[1];
  int x3 = s[3];
  double d1 = x1;
  double d3 = x3;
  vector double x = { d1, d3 };
  return x;
}
```
Compiled with `poc.c main.c -mcpu=pwr8 -O3` on BE machine.
Current clang gives
```
4.000000 1.000000
```
while xlc gives
```
1.000000 4.000000
```
Xlc's output should be correct.

Reviewed By: shchenz, #powerpc

Differential Revision: https://reviews.llvm.org/D107428

Added: 
    

Modified: 
    llvm/lib/Target/PowerPC/PPCInstrVSX.td
    llvm/test/CodeGen/PowerPC/build-vector-tests.ll
    llvm/test/CodeGen/PowerPC/vec_int_to_double_shuffle.ll

Removed: 
    


################################################################################
diff  --git a/llvm/lib/Target/PowerPC/PPCInstrVSX.td b/llvm/lib/Target/PowerPC/PPCInstrVSX.td
index a4101a8f656d..0855b0186dab 100644
--- a/llvm/lib/Target/PowerPC/PPCInstrVSX.td
+++ b/llvm/lib/Target/PowerPC/PPCInstrVSX.td
@@ -2962,11 +2962,11 @@ def : Pat<(v2i64 (fp_to_uint
 def : Pat<WToDPExtractConv.BV02S,
           (v2f64 (XVCVSXWDP $A))>;
 def : Pat<WToDPExtractConv.BV13S,
-          (v2f64 (XVCVSXWDP (XXSLDWI $A, $A, 3)))>;
+          (v2f64 (XVCVSXWDP (XXSLDWI $A, $A, 1)))>;
 def : Pat<WToDPExtractConv.BV02U,
           (v2f64 (XVCVUXWDP $A))>;
 def : Pat<WToDPExtractConv.BV13U,
-          (v2f64 (XVCVUXWDP (XXSLDWI $A, $A, 3)))>;
+          (v2f64 (XVCVUXWDP (XXSLDWI $A, $A, 1)))>;
 def : Pat<(v2f64 (insertelt v2f64:$A, f64:$B, 0)),
           (v2f64 (XXPERMDI (SUBREG_TO_REG (i64 1), $B, sub_64), $A, 1))>;
 def : Pat<(v2f64 (insertelt v2f64:$A, f64:$B, 1)),

diff  --git a/llvm/test/CodeGen/PowerPC/build-vector-tests.ll b/llvm/test/CodeGen/PowerPC/build-vector-tests.ll
index 4bc1a9461d25..edf42bf4a9b0 100644
--- a/llvm/test/CodeGen/PowerPC/build-vector-tests.ll
+++ b/llvm/test/CodeGen/PowerPC/build-vector-tests.ll
@@ -6160,7 +6160,7 @@ entry:
 define dso_local <2 x double> @sint_to_fp_widen13(<4 x i32> %a) {
 ; P9BE-LABEL: sint_to_fp_widen13:
 ; P9BE:       # %bb.0: # %entry
-; P9BE-NEXT:    xxsldwi vs0, v2, v2, 3
+; P9BE-NEXT:    xxsldwi vs0, v2, v2, 1
 ; P9BE-NEXT:    xvcvsxwdp v2, vs0
 ; P9BE-NEXT:    blr
 ;
@@ -6171,7 +6171,7 @@ define dso_local <2 x double> @sint_to_fp_widen13(<4 x i32> %a) {
 ;
 ; P8BE-LABEL: sint_to_fp_widen13:
 ; P8BE:       # %bb.0: # %entry
-; P8BE-NEXT:    xxsldwi vs0, v2, v2, 3
+; P8BE-NEXT:    xxsldwi vs0, v2, v2, 1
 ; P8BE-NEXT:    xvcvsxwdp v2, vs0
 ; P8BE-NEXT:    blr
 ;
@@ -6224,7 +6224,7 @@ entry:
 define dso_local <2 x double> @uint_to_fp_widen13(<4 x i32> %a) {
 ; P9BE-LABEL: uint_to_fp_widen13:
 ; P9BE:       # %bb.0: # %entry
-; P9BE-NEXT:    xxsldwi vs0, v2, v2, 3
+; P9BE-NEXT:    xxsldwi vs0, v2, v2, 1
 ; P9BE-NEXT:    xvcvuxwdp v2, vs0
 ; P9BE-NEXT:    blr
 ;
@@ -6235,7 +6235,7 @@ define dso_local <2 x double> @uint_to_fp_widen13(<4 x i32> %a) {
 ;
 ; P8BE-LABEL: uint_to_fp_widen13:
 ; P8BE:       # %bb.0: # %entry
-; P8BE-NEXT:    xxsldwi vs0, v2, v2, 3
+; P8BE-NEXT:    xxsldwi vs0, v2, v2, 1
 ; P8BE-NEXT:    xvcvuxwdp v2, vs0
 ; P8BE-NEXT:    blr
 ;

diff  --git a/llvm/test/CodeGen/PowerPC/vec_int_to_double_shuffle.ll b/llvm/test/CodeGen/PowerPC/vec_int_to_double_shuffle.ll
index c92a26e8ad95..9bba60435c62 100644
--- a/llvm/test/CodeGen/PowerPC/vec_int_to_double_shuffle.ll
+++ b/llvm/test/CodeGen/PowerPC/vec_int_to_double_shuffle.ll
@@ -12,7 +12,7 @@ define <2 x double> @foo(<4 x i32> %s) {
 ;
 ; CHECK-BE-LABEL: foo:
 ; CHECK-BE:       # %bb.0: # %entry
-; CHECK-BE-NEXT:    xxsldwi 0, 34, 34, 3
+; CHECK-BE-NEXT:    xxsldwi 0, 34, 34, 1
 ; CHECK-BE-NEXT:    xvcvsxwdp 34, 0
 ; CHECK-BE-NEXT:    blr
 entry:
@@ -29,7 +29,7 @@ define <2 x double> @bar(<4 x i32> %s) {
 ;
 ; CHECK-BE-LABEL: bar:
 ; CHECK-BE:       # %bb.0: # %entry
-; CHECK-BE-NEXT:    xxsldwi 0, 34, 34, 3
+; CHECK-BE-NEXT:    xxsldwi 0, 34, 34, 1
 ; CHECK-BE-NEXT:    xvcvuxwdp 34, 0
 ; CHECK-BE-NEXT:    blr
 entry: