[PATCH] D49424: [PowerPC] Handle __builtin_xxpermdi the same way as GCC does

Nemanja Ivanovic via Phabricator via cfe-commits Tue, 17 Jul 2018 06:30:18 -0700

nemanjai created this revision.
nemanjai added reviewers: lu-zero, hfinkel.
Herald added a subscriber: kbarton.


The codegen for this builtin was initially implemented to match GCC. However, 
due to interest from users GCC changed behaviour to account for the big endian 
bias of the instruction and correct it. This patch brings the handling inline 
with GCC.


Repository:
  rC Clang

https://reviews.llvm.org/D49424

Files:
  lib/CodeGen/CGBuiltin.cpp
  test/CodeGen/builtins-ppc-vsx.c


Index: test/CodeGen/builtins-ppc-vsx.c
===================================================================
--- test/CodeGen/builtins-ppc-vsx.c
+++ test/CodeGen/builtins-ppc-vsx.c
@@ -1694,43 +1694,43 @@
 
 res_vd = vec_xxpermdi(vd, vd, 0);
 // CHECK: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x 
i32> <i32 0, i32 2>
-// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x 
i32> <i32 3, i32 1>
+// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x 
i32> <i32 0, i32 2>
 
 res_vf = vec_xxpermdi(vf, vf, 1);
 // CHECK: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x 
i32> <i32 0, i32 3>
-// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x 
i32> <i32 2, i32 1>
+// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x 
i32> <i32 0, i32 3>
 
 res_vsll = vec_xxpermdi(vsll, vsll, 2);
 // CHECK: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x 
i32> <i32 1, i32 2>
-// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x 
i32> <i32 3, i32 0>
+// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x 
i32> <i32 1, i32 2>
 
 res_vull = vec_xxpermdi(vull, vull, 3);
 // CHECK: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x 
i32> <i32 1, i32 3>
-// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x 
i32> <i32 2, i32 0>
+// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x 
i32> <i32 1, i32 3>
 
 res_vsi = vec_xxpermdi(vsi, vsi, 0);
 // CHECK: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x 
i32> <i32 0, i32 2>
-// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x 
i32> <i32 3, i32 1>
+// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x 
i32> <i32 0, i32 2>
 
 res_vui = vec_xxpermdi(vui, vui, 1);
 // CHECK: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x 
i32> <i32 0, i32 3>
-// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x 
i32> <i32 2, i32 1>
+// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x 
i32> <i32 0, i32 3>
 
 res_vss = vec_xxpermdi(vss, vss, 2);
 // CHECK: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x 
i32> <i32 1, i32 2>
-// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x 
i32> <i32 3, i32 0>
+// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x 
i32> <i32 1, i32 2>
 
 res_vus = vec_xxpermdi(vus, vus, 3);
 // CHECK: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x 
i32> <i32 1, i32 3>
-// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x 
i32> <i32 2, i32 0>
+// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x 
i32> <i32 1, i32 3>
 
 res_vsc = vec_xxpermdi(vsc, vsc, 0);
 // CHECK: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x 
i32> <i32 0, i32 2>
-// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x 
i32> <i32 3, i32 1>
+// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x 
i32> <i32 0, i32 2>
 
 res_vuc = vec_xxpermdi(vuc, vuc, 1);
 // CHECK: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x 
i32> <i32 0, i32 3>
-// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x 
i32> <i32 2, i32 1>
+// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x 
i32> <i32 0, i32 3>
 
 res_vd = vec_xxsldwi(vd, vd, 0);
 // CHECK: shufflevector <4 x i32> %{{[0-9]+}}, <4 x i32> %{{[0-9]+}}, <4 x 
i32> <i32 0, i32 1, i32 2, i32 3>
@@ -1786,7 +1786,7 @@
 
 // CHECK-LE:  bitcast <4 x i32> %{{[0-9]+}} to <2 x i64>
 // CHECK-LE-NEXT:  bitcast <4 x i32> %{{[0-9]+}} to <2 x i64>
-// CHECK-LE-NEXT:  shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, 
<2 x i32> <i32 3, i32 1>
+// CHECK-LE-NEXT:  shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, 
<2 x i32> <i32 0, i32 2>
 // CHECK-LE-NEXT:  bitcast <2 x i64> %{{[0-9]+}} to <4 x i32>
 }
 
Index: lib/CodeGen/CGBuiltin.cpp
===================================================================
--- lib/CodeGen/CGBuiltin.cpp
+++ lib/CodeGen/CGBuiltin.cpp
@@ -10778,15 +10778,8 @@
     // the second. The element indices within each vector are numbered in big
     // endian order so the shuffle mask must be adjusted for this on little
     // endian platforms (i.e. index is complemented and source vector 
reversed).
-    unsigned ElemIdx0;
-    unsigned ElemIdx1;
-    if (getTarget().isLittleEndian()) {
-      ElemIdx0 = (~Index & 1) + 2;
-      ElemIdx1 = (~Index & 2) >> 1;
-    } else { // BigEndian
-      ElemIdx0 = (Index & 2) >> 1;
-      ElemIdx1 = 2 + (Index & 1);
-    }
+    unsigned ElemIdx0 = (Index & 2) >> 1;;
+    unsigned ElemIdx1 = 2 + (Index & 1);;
 
     Constant *ShuffleElts[2] = {ConstantInt::get(Int32Ty, ElemIdx0),
                                 ConstantInt::get(Int32Ty, ElemIdx1)};

Index: test/CodeGen/builtins-ppc-vsx.c
===================================================================
--- test/CodeGen/builtins-ppc-vsx.c
+++ test/CodeGen/builtins-ppc-vsx.c
@@ -1694,43 +1694,43 @@
 
 res_vd = vec_xxpermdi(vd, vd, 0);
 // CHECK: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 0, i32 2>
-// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 3, i32 1>
+// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 0, i32 2>
 
 res_vf = vec_xxpermdi(vf, vf, 1);
 // CHECK: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 0, i32 3>
-// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 2, i32 1>
+// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 0, i32 3>
 
 res_vsll = vec_xxpermdi(vsll, vsll, 2);
 // CHECK: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 1, i32 2>
-// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 3, i32 0>
+// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 1, i32 2>
 
 res_vull = vec_xxpermdi(vull, vull, 3);
 // CHECK: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 1, i32 3>
-// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 2, i32 0>
+// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 1, i32 3>
 
 res_vsi = vec_xxpermdi(vsi, vsi, 0);
 // CHECK: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 0, i32 2>
-// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 3, i32 1>
+// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 0, i32 2>
 
 res_vui = vec_xxpermdi(vui, vui, 1);
 // CHECK: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 0, i32 3>
-// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 2, i32 1>
+// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 0, i32 3>
 
 res_vss = vec_xxpermdi(vss, vss, 2);
 // CHECK: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 1, i32 2>
-// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 3, i32 0>
+// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 1, i32 2>
 
 res_vus = vec_xxpermdi(vus, vus, 3);
 // CHECK: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 1, i32 3>
-// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 2, i32 0>
+// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 1, i32 3>
 
 res_vsc = vec_xxpermdi(vsc, vsc, 0);
 // CHECK: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 0, i32 2>
-// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 3, i32 1>
+// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 0, i32 2>
 
 res_vuc = vec_xxpermdi(vuc, vuc, 1);
 // CHECK: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 0, i32 3>
-// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 2, i32 1>
+// CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 0, i32 3>
 
 res_vd = vec_xxsldwi(vd, vd, 0);
 // CHECK: shufflevector <4 x i32> %{{[0-9]+}}, <4 x i32> %{{[0-9]+}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
@@ -1786,7 +1786,7 @@
 
 // CHECK-LE:  bitcast <4 x i32> %{{[0-9]+}} to <2 x i64>
 // CHECK-LE-NEXT:  bitcast <4 x i32> %{{[0-9]+}} to <2 x i64>
-// CHECK-LE-NEXT:  shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 3, i32 1>
+// CHECK-LE-NEXT:  shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 0, i32 2>
 // CHECK-LE-NEXT:  bitcast <2 x i64> %{{[0-9]+}} to <4 x i32>
 }
 
Index: lib/CodeGen/CGBuiltin.cpp
===================================================================
--- lib/CodeGen/CGBuiltin.cpp
+++ lib/CodeGen/CGBuiltin.cpp
@@ -10778,15 +10778,8 @@
     // the second. The element indices within each vector are numbered in big
     // endian order so the shuffle mask must be adjusted for this on little
     // endian platforms (i.e. index is complemented and source vector reversed).
-    unsigned ElemIdx0;
-    unsigned ElemIdx1;
-    if (getTarget().isLittleEndian()) {
-      ElemIdx0 = (~Index & 1) + 2;
-      ElemIdx1 = (~Index & 2) >> 1;
-    } else { // BigEndian
-      ElemIdx0 = (Index & 2) >> 1;
-      ElemIdx1 = 2 + (Index & 1);
-    }
+    unsigned ElemIdx0 = (Index & 2) >> 1;;
+    unsigned ElemIdx1 = 2 + (Index & 1);;
 
     Constant *ShuffleElts[2] = {ConstantInt::get(Int32Ty, ElemIdx0),
                                 ConstantInt::get(Int32Ty, ElemIdx1)};

_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D49424: [PowerPC] Handle __builtin_xxpermdi the same way as GCC does

Reply via email to