From ba33a0928e789ce6ebd21bd53ba2c31dd42b0ce8 Mon Sep 17 00:00:00 2001
From: Nemanja Ivanovic <nemanja.i.ibm@gmail.com>
Date: Wed, 26 Oct 2016 18:25:45 +0000
Subject: [PATCH] [PPC] Implement vector reverse elements builtins (vec_reve)

This patch corresponds to review https://reviews.llvm.org/D25906.
Committing on behalf of Tony Jiang.



git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@285218 91177308-0d34-0410-b5e6-96231b3b80d8
---
 lib/Headers/altivec.h               | 71 +++++++++++++++++++++++++++++
 test/CodeGen/builtins-ppc-altivec.c | 45 ++++++++++++++++++
 test/CodeGen/builtins-ppc-vsx.c     | 16 +++++++
 3 files changed, 132 insertions(+)

diff --git a/lib/Headers/altivec.h b/lib/Headers/altivec.h
index 87a3d0688d..5dd093a9ef 100644
--- a/lib/Headers/altivec.h
+++ b/lib/Headers/altivec.h
@@ -15034,6 +15034,77 @@ vec_bperm(vector unsigned __int128 __a, vector unsigned char __b) {
 #endif
 #endif
 
+
+/* vec_reve */
+
+static inline __ATTRS_o_ai vector bool char vec_reve(vector bool char __a) {
+  return __builtin_shufflevector(__a, __a, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6,
+                                 5, 4, 3, 2, 1, 0);
+}
+
+static inline __ATTRS_o_ai vector signed char vec_reve(vector signed char __a) {
+  return __builtin_shufflevector(__a, __a, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6,
+                                 5, 4, 3, 2, 1, 0);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_reve(vector unsigned char __a) {
+  return __builtin_shufflevector(__a, __a, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6,
+                                 5, 4, 3, 2, 1, 0);
+}
+
+static inline __ATTRS_o_ai vector bool int vec_reve(vector bool int __a) {
+  return __builtin_shufflevector(__a, __a, 3, 2, 1, 0);
+}
+
+static inline __ATTRS_o_ai vector signed int vec_reve(vector signed int __a) {
+  return __builtin_shufflevector(__a, __a, 3, 2, 1, 0);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_reve(vector unsigned int __a) {
+  return __builtin_shufflevector(__a, __a, 3, 2, 1, 0);
+}
+
+static inline __ATTRS_o_ai vector bool short vec_reve(vector bool short __a) {
+  return __builtin_shufflevector(__a, __a, 7, 6, 5, 4, 3, 2, 1, 0);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_reve(vector signed short __a) {
+  return __builtin_shufflevector(__a, __a, 7, 6, 5, 4, 3, 2, 1, 0);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_reve(vector unsigned short __a) {
+  return __builtin_shufflevector(__a, __a, 7, 6, 5, 4, 3, 2, 1, 0);
+}
+
+static inline __ATTRS_o_ai vector float vec_reve(vector float __a) {
+  return __builtin_shufflevector(__a, __a, 3, 2, 1, 0);
+}
+
+#ifdef __VSX__
+static inline __ATTRS_o_ai vector bool long long
+vec_reve(vector bool long long __a) {
+  return __builtin_shufflevector(__a, __a, 1, 0);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_reve(vector signed long long __a) {
+  return __builtin_shufflevector(__a, __a, 1, 0);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_reve(vector unsigned long long __a) {
+  return __builtin_shufflevector(__a, __a, 1, 0);
+}
+
+static inline __ATTRS_o_ai vector double vec_reve(vector double __a) {
+  return __builtin_shufflevector(__a, __a, 1, 0);
+}
+#endif
+
 #undef __ATTRS_o_ai
 
 #endif /* __ALTIVEC_H */
diff --git a/test/CodeGen/builtins-ppc-altivec.c b/test/CodeGen/builtins-ppc-altivec.c
index 1edf99f868..895938cb44 100644
--- a/test/CodeGen/builtins-ppc-altivec.c
+++ b/test/CodeGen/builtins-ppc-altivec.c
@@ -8996,3 +8996,48 @@ void test7() {
 // CHECK: @llvm.ppc.altivec.vcmpgefp.p(i32 2
 // CHECK-LE: @llvm.ppc.altivec.vcmpgefp.p(i32 2
 }
+
+void test8() {
+  // CHECK-LABEL: define void @test8
+  // CHECK-LE-LABEL: define void @test8
+  res_vbc = vec_reve(vbc);
+  // CHECK: shufflevector <16 x i8> %{{[0-9]+}}, <16 x i8> %{{[0-9]+}}, <16 x i32> <i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+  // CHECK-LE: shufflevector <16 x i8> %{{[0-9]+}}, <16 x i8> %{{[0-9]+}}, <16 x i32> <i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+
+  res_vsc = vec_reve(vsc);
+  // CHECK: shufflevector <16 x i8> %{{[0-9]+}}, <16 x i8> %{{[0-9]+}}, <16 x i32> <i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+  // CHECK-LE: shufflevector <16 x i8> %{{[0-9]+}}, <16 x i8> %{{[0-9]+}}, <16 x i32> <i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+
+  res_vuc = vec_reve(vuc);
+  // CHECK: shufflevector <16 x i8> %{{[0-9]+}}, <16 x i8> %{{[0-9]+}}, <16 x i32> <i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+  // CHECK-LE: shufflevector <16 x i8> %{{[0-9]+}}, <16 x i8> %{{[0-9]+}}, <16 x i32> <i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+
+  res_vbi = vec_reve(vbi);
+  // CHECK: shufflevector <4 x i32> %{{[0-9]+}}, <4 x i32> %{{[0-9]+}}, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
+  // CHECK-LE: shufflevector <4 x i32> %{{[0-9]+}}, <4 x i32> %{{[0-9]+}}, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
+
+  res_vi = vec_reve(vi);
+  // CHECK: shufflevector <4 x i32> %{{[0-9]+}}, <4 x i32> %{{[0-9]+}}, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
+  // CHECK-LE: shufflevector <4 x i32> %{{[0-9]+}}, <4 x i32> %{{[0-9]+}}, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
+
+  res_vui = vec_reve(vui);
+  // CHECK: shufflevector <4 x i32> %{{[0-9]+}}, <4 x i32> %{{[0-9]+}}, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
+  // CHECK-LE: shufflevector <4 x i32> %{{[0-9]+}}, <4 x i32> %{{[0-9]+}}, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
+
+  res_vbs = vec_reve(vbs);
+  // CHECK: shufflevector <8 x i16> %{{[0-9]+}}, <8 x i16> %{{[0-9]+}}, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+  // CHECK-LE: shufflevector <8 x i16> %{{[0-9]+}}, <8 x i16> %{{[0-9]+}}, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+
+  res_vbs = vec_reve(vs);
+  // CHECK: shufflevector <8 x i16> %{{[0-9]+}}, <8 x i16> %{{[0-9]+}}, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+  // CHECK-LE: shufflevector <8 x i16> %{{[0-9]+}}, <8 x i16> %{{[0-9]+}}, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+
+  res_vbs = vec_reve(vus);
+  // CHECK: shufflevector <8 x i16> %{{[0-9]+}}, <8 x i16> %{{[0-9]+}}, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+  // CHECK-LE: shufflevector <8 x i16> %{{[0-9]+}}, <8 x i16> %{{[0-9]+}}, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+
+  res_vf = vec_reve(vf);
+  // CHECK: shufflevector <4 x float> %{{[0-9]+}}, <4 x float> %{{[0-9]+}}, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
+  // CHECK-LE: shufflevector <4 x float> %{{[0-9]+}}, <4 x float> %{{[0-9]+}}, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
+
+}
diff --git a/test/CodeGen/builtins-ppc-vsx.c b/test/CodeGen/builtins-ppc-vsx.c
index e58afdd94d..481d7dc44e 100644
--- a/test/CodeGen/builtins-ppc-vsx.c
+++ b/test/CodeGen/builtins-ppc-vsx.c
@@ -1080,4 +1080,20 @@ void test1() {
 // CHECK: fmul <2 x double>
 // CHECK-LE: uitofp <2 x i64> %{{.*}} to <2 x double>
 // CHECK-LE: fmul <2 x double>
+
+  res_vbll = vec_reve(vbll);
+  // CHECK: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 1, i32 0>
+  // CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 1, i32 0>
+
+  res_vsll = vec_reve(vsll);
+  // CHECK: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 1, i32 0>
+  // CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 1, i32 0>
+
+  res_vull = vec_reve(vull);
+  // CHECK: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 1, i32 0>
+  // CHECK-LE: shufflevector <2 x i64> %{{[0-9]+}}, <2 x i64> %{{[0-9]+}}, <2 x i32> <i32 1, i32 0>
+
+  res_vd = vec_reve(vd);
+  // CHECK: shufflevector <2 x double> %{{[0-9]+}}, <2 x double> %{{[0-9]+}}, <2 x i32> <i32 1, i32 0>
+  // CHECK-LE: shufflevector <2 x double> %{{[0-9]+}}, <2 x double> %{{[0-9]+}}, <2 x i32> <i32 1, i32 0>
 }
-- 
2.40.0