[llvm] [clang] [PowerPC][X86] Make cpu id builtins target independent and lower for PPC (PR #68919)
Nemanja Ivanovic via cfe-commits
cfe-commits at lists.llvm.org
Thu Jan 25 22:46:01 PST 2024
https://github.com/nemanjai updated https://github.com/llvm/llvm-project/pull/68919
>From 65c84f2ba78efcbf92ce9c8232fc40f493414930 Mon Sep 17 00:00:00 2001
From: Nemanja Ivanovic <nemanjaivanovic at nemanjas-air.kpn>
Date: Thu, 12 Oct 2023 14:08:42 -0400
Subject: [PATCH 1/5] [PowerPC][X86] Make cpu id builtins target independent
and lower for PPC
Make __builtin_cpu_{init|supports|is} target independent and provide
an opt-in query for targets that want to support it. Each target is
still responsible for their specific lowering/code-gen.
Also provide code-gen for PowerPC.
---
clang/include/clang/Basic/Builtins.td | 20 +++
clang/include/clang/Basic/BuiltinsX86.def | 7 -
clang/include/clang/Basic/TargetInfo.h | 6 +
clang/lib/Basic/Targets/PPC.cpp | 14 ++
clang/lib/Basic/Targets/PPC.h | 7 +
clang/lib/Basic/Targets/X86.h | 4 +
clang/lib/CodeGen/CGBuiltin.cpp | 42 +++++-
clang/lib/Sema/SemaChecking.cpp | 124 +++++++++++-------
clang/test/CodeGen/builtin-cpu-supports.c | 68 ++++++----
clang/test/Sema/builtin-cpu-supports.c | 8 +-
llvm/include/llvm/IR/IntrinsicsPowerPC.td | 6 +
.../llvm/TargetParser/PPCTargetParser.def | 80 +++++++++++
llvm/lib/Target/PowerPC/PPCAsmPrinter.cpp | 4 +
llvm/lib/Target/PowerPC/PPCInstrInfo.cpp | 33 +++++
llvm/lib/Target/PowerPC/PPCInstrInfo.td | 3 +
llvm/lib/Target/PowerPC/PPCTargetMachine.h | 3 +
llvm/test/CodeGen/PowerPC/cpu-supports.ll | 111 ++++++++++++++++
17 files changed, 458 insertions(+), 82 deletions(-)
create mode 100644 llvm/include/llvm/TargetParser/PPCTargetParser.def
create mode 100644 llvm/test/CodeGen/PowerPC/cpu-supports.ll
diff --git a/clang/include/clang/Basic/Builtins.td b/clang/include/clang/Basic/Builtins.td
index 22e616e6cde599..1af01fe0d700c9 100644
--- a/clang/include/clang/Basic/Builtins.td
+++ b/clang/include/clang/Basic/Builtins.td
@@ -727,6 +727,26 @@ def RotateRight : BitInt8_16_32_64BuiltinsTemplate, Builtin {
// FIXME: The builtins marked FunctionWithBuiltinPrefix below should be
// merged with the library definitions. They are currently not because
// the attributes are different.
+
+// Builtins for checking CPU features based on the GCC builtins.
+def BuiltinCPUIs : Builtin {
+ let Spellings = ["__builtin_cpu_is"];
+ let Attributes = [NoThrow, Const];
+ let Prototype = "bool(char const*)";
+}
+
+def BuiltinCPUSupports : Builtin {
+ let Spellings = ["__builtin_cpu_supports"];
+ let Attributes = [NoThrow, Const];
+ let Prototype = "bool(char const*)";
+}
+
+def BuiltinCPUInit : Builtin {
+ let Spellings = ["__builtin_cpu_init"];
+ let Attributes = [NoThrow];
+ let Prototype = "void()";
+}
+
def BuiltinCalloc : Builtin {
let Spellings = ["__builtin_calloc"];
let Attributes = [FunctionWithBuiltinPrefix, NoThrow];
diff --git a/clang/include/clang/Basic/BuiltinsX86.def b/clang/include/clang/Basic/BuiltinsX86.def
index 60b752ad48548f..207cde0414b54e 100644
--- a/clang/include/clang/Basic/BuiltinsX86.def
+++ b/clang/include/clang/Basic/BuiltinsX86.def
@@ -26,13 +26,6 @@
# define TARGET_HEADER_BUILTIN(ID, TYPE, ATTRS, HEADER, LANG, FEATURE) BUILTIN(ID, TYPE, ATTRS)
#endif
-// Miscellaneous builtin for checking x86 cpu features.
-// TODO: Make this somewhat generic so that other backends
-// can use it?
-BUILTIN(__builtin_cpu_init, "v", "n")
-BUILTIN(__builtin_cpu_supports, "bcC*", "nc")
-BUILTIN(__builtin_cpu_is, "bcC*", "nc")
-
// Undefined Values
//
TARGET_BUILTIN(__builtin_ia32_undef128, "V2d", "ncV:128:", "")
diff --git a/clang/include/clang/Basic/TargetInfo.h b/clang/include/clang/Basic/TargetInfo.h
index 3eb23ebdacf0ed..9432154d5063ce 100644
--- a/clang/include/clang/Basic/TargetInfo.h
+++ b/clang/include/clang/Basic/TargetInfo.h
@@ -1432,6 +1432,12 @@ class TargetInfo : public TransferrableTargetInfo,
getTriple().isOSFreeBSD());
}
+ // Identify whether this target supports __builtin_cpu_supports and
+ // __builtin_cpu_is.
+ virtual bool supportsCpuSupports() const { return false; }
+ virtual bool supportsCpuIs() const { return false; }
+ virtual bool supportsCpuInit() const { return false; }
+
// Validate the contents of the __builtin_cpu_supports(const char*)
// argument.
virtual bool validateCpuSupports(StringRef Name) const { return false; }
diff --git a/clang/lib/Basic/Targets/PPC.cpp b/clang/lib/Basic/Targets/PPC.cpp
index 41935abfb65d3b..2cf1bacd95fd95 100644
--- a/clang/lib/Basic/Targets/PPC.cpp
+++ b/clang/lib/Basic/Targets/PPC.cpp
@@ -878,3 +878,17 @@ ArrayRef<Builtin::Info> PPCTargetInfo::getTargetBuiltins() const {
return llvm::ArrayRef(BuiltinInfo,
clang::PPC::LastTSBuiltin - Builtin::FirstTSBuiltin);
}
+
+bool PPCTargetInfo::validateCpuSupports(StringRef FeatureStr) const {
+#define PPC_FEATURE(NAME, DESC, ENUMNAME, ENUMVAL, HWCAPN) .Case(NAME, true)
+ return llvm::StringSwitch<bool>(FeatureStr)
+#include "llvm/TargetParser/PPCTargetParser.def"
+ .Default(false);
+}
+
+bool PPCTargetInfo::validateCpuIs(StringRef CPUName) const {
+#define PPC_CPU(NAME, NUM) .Case(NAME, true)
+ return llvm::StringSwitch<bool>(CPUName)
+#include "llvm/TargetParser/PPCTargetParser.def"
+ .Default(false);
+}
diff --git a/clang/lib/Basic/Targets/PPC.h b/clang/lib/Basic/Targets/PPC.h
index 4d62673ba7fb8c..f700b625b79030 100644
--- a/clang/lib/Basic/Targets/PPC.h
+++ b/clang/lib/Basic/Targets/PPC.h
@@ -359,6 +359,13 @@ class LLVM_LIBRARY_VISIBILITY PPCTargetInfo : public TargetInfo {
bool isSPRegName(StringRef RegName) const override {
return RegName.equals("r1") || RegName.equals("x1");
}
+
+ // We support __builtin_cpu_supports/__builtin_cpu_is on targets that
+ // have GLIBC since it is GLIBC that provides the HWCAP[2] in the auxv.
+ bool supportsCpuSupports() const override { return getTriple().isOSGlibc(); }
+ bool supportsCpuIs() const override { return getTriple().isOSGlibc(); }
+ bool validateCpuSupports(StringRef Feature) const override;
+ bool validateCpuIs(StringRef Name) const override;
};
class LLVM_LIBRARY_VISIBILITY PPC32TargetInfo : public PPCTargetInfo {
diff --git a/clang/lib/Basic/Targets/X86.h b/clang/lib/Basic/Targets/X86.h
index 0ab1c10833db26..1845f5a747af40 100644
--- a/clang/lib/Basic/Targets/X86.h
+++ b/clang/lib/Basic/Targets/X86.h
@@ -220,6 +220,10 @@ class LLVM_LIBRARY_VISIBILITY X86TargetInfo : public TargetInfo {
return RegName.equals("esp") || RegName.equals("rsp");
}
+ bool supportsCpuSupports() const override { return true; }
+ bool supportsCpuIs() const override { return true; }
+ bool supportsCpuInit() const override { return true; }
+
bool validateCpuSupports(StringRef FeatureStr) const override;
bool validateCpuIs(StringRef FeatureStr) const override;
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 27183c69b7ea37..3c9d8155569fac 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -14053,11 +14053,11 @@ CodeGenFunction::EmitAArch64CpuSupports(ArrayRef<StringRef> FeaturesStrs) {
Value *CodeGenFunction::EmitX86BuiltinExpr(unsigned BuiltinID,
const CallExpr *E) {
- if (BuiltinID == X86::BI__builtin_cpu_is)
+ if (BuiltinID == Builtin::BI__builtin_cpu_is)
return EmitX86CpuIs(E);
- if (BuiltinID == X86::BI__builtin_cpu_supports)
+ if (BuiltinID == Builtin::BI__builtin_cpu_supports)
return EmitX86CpuSupports(E);
- if (BuiltinID == X86::BI__builtin_cpu_init)
+ if (BuiltinID == Builtin::BI__builtin_cpu_init)
return EmitX86CpuInit();
// Handle MSVC intrinsics before argument evaluation to prevent double
@@ -16545,6 +16545,42 @@ Value *CodeGenFunction::EmitPPCBuiltinExpr(unsigned BuiltinID,
switch (BuiltinID) {
default: return nullptr;
+ case Builtin::BI__builtin_cpu_is: {
+ const Expr *CPUExpr = E->getArg(0)->IgnoreParenCasts();
+ StringRef CPUStr = cast<clang::StringLiteral>(CPUExpr)->getString();
+ unsigned NumCPUID = StringSwitch<unsigned>(CPUStr)
+#define PPC_CPU(Name, NumericID) .Case(Name, NumericID)
+#include "llvm/TargetParser/PPCTargetParser.def"
+ .Default(-1U);
+ Value *Op0 =
+ llvm::ConstantInt::get(Int32Ty, PPC_FAWORD_CPUID);
+ llvm::Function *F = CGM.getIntrinsic(Intrinsic::ppc_fixed_addr_ld);
+ Value *TheCall = Builder.CreateCall(F, {Op0}, "cpu_is");
+ return Builder.CreateICmpEQ(TheCall,
+ llvm::ConstantInt::get(Int32Ty, NumCPUID));
+ }
+ case Builtin::BI__builtin_cpu_supports: {
+ unsigned FeatureWord;
+ unsigned BitMask;
+ const Expr *CPUExpr = E->getArg(0)->IgnoreParenCasts();
+ StringRef CPUStr = cast<clang::StringLiteral>(CPUExpr)->getString();
+ std::tie(FeatureWord, BitMask) =
+ StringSwitch<std::pair<unsigned, unsigned>>(CPUStr)
+#define PPC_FEATURE(Name, Description, EnumName, Bitmask, FA_WORD) \
+ .Case(Name, {FA_WORD, Bitmask})
+#include "llvm/TargetParser/PPCTargetParser.def"
+ .Default({0, 0});
+ Value *Op0 = llvm::ConstantInt::get(Int32Ty, FeatureWord);
+ llvm::Function *F = CGM.getIntrinsic(Intrinsic::ppc_fixed_addr_ld);
+ Value *TheCall = Builder.CreateCall(F, {Op0}, "cpu_supports");
+ Value *Mask =
+ Builder.CreateAnd(TheCall, llvm::ConstantInt::get(Int32Ty, BitMask));
+ return Builder.CreateICmpNE(Mask, llvm::Constant::getNullValue(Int32Ty));
+#undef PPC_FAWORD_HWCAP
+#undef PPC_FAWORD_HWCAP2
+#undef PPC_FAWORD_CPUID
+ }
+
// __builtin_ppc_get_timebase is GCC 4.8+'s PowerPC-specific name for what we
// call __builtin_readcyclecounter.
case PPC::BI__builtin_ppc_get_timebase:
diff --git a/clang/lib/Sema/SemaChecking.cpp b/clang/lib/Sema/SemaChecking.cpp
index 4d280f25cc04c2..36305eeac8c9de 100644
--- a/clang/lib/Sema/SemaChecking.cpp
+++ b/clang/lib/Sema/SemaChecking.cpp
@@ -2143,6 +2143,66 @@ static bool checkFPMathBuiltinElementType(Sema &S, SourceLocation Loc,
return false;
}
+/// SemaBuiltinCpuSupports - Handle __builtin_cpu_supports(char *).
+/// This checks that the target supports __builtin_cpu_supports and
+/// that the string argument is constant and valid.
+static bool SemaBuiltinCpuSupports(Sema &S, const TargetInfo &TI,
+ const TargetInfo *AuxTI, CallExpr *TheCall) {
+ Expr *Arg = TheCall->getArg(0);
+
+ const TargetInfo *TheTI = nullptr;
+ if (TI.supportsCpuSupports())
+ TheTI = &TI;
+ else if (AuxTI && AuxTI->supportsCpuSupports())
+ TheTI = AuxTI;
+ else
+ return S.Diag(TheCall->getBeginLoc(), diag::err_builtin_target_unsupported)
+ << SourceRange(TheCall->getBeginLoc(), TheCall->getEndLoc());
+
+ // Check if the argument is a string literal.
+ if (!isa<StringLiteral>(Arg->IgnoreParenImpCasts()))
+ return S.Diag(TheCall->getBeginLoc(), diag::err_expr_not_string_literal)
+ << Arg->getSourceRange();
+
+ // Check the contents of the string.
+ StringRef Feature =
+ cast<StringLiteral>(Arg->IgnoreParenImpCasts())->getString();
+ if (!TheTI->validateCpuSupports(Feature))
+ return S.Diag(TheCall->getBeginLoc(), diag::err_invalid_cpu_supports)
+ << Arg->getSourceRange();
+ return false;
+}
+
+/// SemaBuiltinCpuIs - Handle __builtin_cpu_is(char *).
+/// This checks that the target supports __builtin_cpu_is and
+/// that the string argument is constant and valid.
+static bool SemaBuiltinCpuIs(Sema &S, const TargetInfo &TI,
+ const TargetInfo *AuxTI, CallExpr *TheCall) {
+ Expr *Arg = TheCall->getArg(0);
+
+ const TargetInfo *TheTI = nullptr;
+ if (TI.supportsCpuIs())
+ TheTI = &TI;
+ else if (AuxTI && AuxTI->supportsCpuIs())
+ TheTI = AuxTI;
+ else
+ return S.Diag(TheCall->getBeginLoc(), diag::err_builtin_target_unsupported)
+ << SourceRange(TheCall->getBeginLoc(), TheCall->getEndLoc());
+
+ // Check if the argument is a string literal.
+ if (!isa<StringLiteral>(Arg->IgnoreParenImpCasts()))
+ return S.Diag(TheCall->getBeginLoc(), diag::err_expr_not_string_literal)
+ << Arg->getSourceRange();
+
+ // Check the contents of the string.
+ StringRef Feature =
+ cast<StringLiteral>(Arg->IgnoreParenImpCasts())->getString();
+ if (!TheTI->validateCpuIs(Feature))
+ return S.Diag(TheCall->getBeginLoc(), diag::err_invalid_cpu_is)
+ << Arg->getSourceRange();
+ return false;
+}
+
ExprResult
Sema::CheckBuiltinFunctionCall(FunctionDecl *FDecl, unsigned BuiltinID,
CallExpr *TheCall) {
@@ -2171,6 +2231,23 @@ Sema::CheckBuiltinFunctionCall(FunctionDecl *FDecl, unsigned BuiltinID,
FPOptions FPO;
switch (BuiltinID) {
+ case Builtin::BI__builtin_cpu_supports:
+ if (SemaBuiltinCpuSupports(*this, Context.getTargetInfo(),
+ Context.getAuxTargetInfo(), TheCall))
+ return ExprError();
+ break;
+ case Builtin::BI__builtin_cpu_is:
+ if (SemaBuiltinCpuIs(*this, Context.getTargetInfo(),
+ Context.getAuxTargetInfo(), TheCall))
+ return ExprError();
+ break;
+ case Builtin::BI__builtin_cpu_init:
+ if (!Context.getTargetInfo().supportsCpuInit()) {
+ Diag(TheCall->getBeginLoc(), diag::err_builtin_target_unsupported)
+ << SourceRange(TheCall->getBeginLoc(), TheCall->getEndLoc());
+ return ExprError();
+ }
+ break;
case Builtin::BI__builtin___CFStringMakeConstantString:
// CFStringMakeConstantString is currently not implemented for GOFF (i.e.,
// on z/OS) and for XCOFF (i.e., on AIX). Emit unsupported
@@ -6256,47 +6333,6 @@ bool Sema::CheckNVPTXBuiltinFunctionCall(const TargetInfo &TI,
return false;
}
-/// SemaBuiltinCpuSupports - Handle __builtin_cpu_supports(char *).
-/// This checks that the target supports __builtin_cpu_supports and
-/// that the string argument is constant and valid.
-static bool SemaBuiltinCpuSupports(Sema &S, const TargetInfo &TI,
- CallExpr *TheCall) {
- Expr *Arg = TheCall->getArg(0);
-
- // Check if the argument is a string literal.
- if (!isa<StringLiteral>(Arg->IgnoreParenImpCasts()))
- return S.Diag(TheCall->getBeginLoc(), diag::err_expr_not_string_literal)
- << Arg->getSourceRange();
-
- // Check the contents of the string.
- StringRef Feature =
- cast<StringLiteral>(Arg->IgnoreParenImpCasts())->getString();
- if (!TI.validateCpuSupports(Feature))
- return S.Diag(TheCall->getBeginLoc(), diag::err_invalid_cpu_supports)
- << Arg->getSourceRange();
- return false;
-}
-
-/// SemaBuiltinCpuIs - Handle __builtin_cpu_is(char *).
-/// This checks that the target supports __builtin_cpu_is and
-/// that the string argument is constant and valid.
-static bool SemaBuiltinCpuIs(Sema &S, const TargetInfo &TI, CallExpr *TheCall) {
- Expr *Arg = TheCall->getArg(0);
-
- // Check if the argument is a string literal.
- if (!isa<StringLiteral>(Arg->IgnoreParenImpCasts()))
- return S.Diag(TheCall->getBeginLoc(), diag::err_expr_not_string_literal)
- << Arg->getSourceRange();
-
- // Check the contents of the string.
- StringRef Feature =
- cast<StringLiteral>(Arg->IgnoreParenImpCasts())->getString();
- if (!TI.validateCpuIs(Feature))
- return S.Diag(TheCall->getBeginLoc(), diag::err_invalid_cpu_is)
- << Arg->getSourceRange();
- return false;
-}
-
// Check if the rounding mode is legal.
bool Sema::CheckX86BuiltinRoundingOrSAE(unsigned BuiltinID, CallExpr *TheCall) {
// Indicates if this instruction has rounding control or just SAE.
@@ -6771,12 +6807,6 @@ static bool isX86_32Builtin(unsigned BuiltinID) {
bool Sema::CheckX86BuiltinFunctionCall(const TargetInfo &TI, unsigned BuiltinID,
CallExpr *TheCall) {
- if (BuiltinID == X86::BI__builtin_cpu_supports)
- return SemaBuiltinCpuSupports(*this, TI, TheCall);
-
- if (BuiltinID == X86::BI__builtin_cpu_is)
- return SemaBuiltinCpuIs(*this, TI, TheCall);
-
// Check for 32-bit only builtins on a 64-bit target.
const llvm::Triple &TT = TI.getTriple();
if (TT.getArch() != llvm::Triple::x86 && isX86_32Builtin(BuiltinID))
diff --git a/clang/test/CodeGen/builtin-cpu-supports.c b/clang/test/CodeGen/builtin-cpu-supports.c
index 796611c1fcd9af..42497d8845437a 100644
--- a/clang/test/CodeGen/builtin-cpu-supports.c
+++ b/clang/test/CodeGen/builtin-cpu-supports.c
@@ -1,11 +1,16 @@
-// RUN: %clang_cc1 -triple x86_64-pc-linux-gnu -emit-llvm < %s| FileCheck %s
+// RUN: %clang_cc1 -triple x86_64-pc-linux-gnu -emit-llvm < %s | FileCheck %s \
+// RUN: --check-prefix=CHECK-X86
+// RUN: %clang_cc1 -triple ppc64le-linux-gnu -emit-llvm < %s | FileCheck %s \
+// RUN: --check-prefix=CHECK-PPC
+
+#ifndef __PPC__
// Test that we have the structure definition, the gep offsets, the name of the
// global, the bit grab, and the icmp correct.
extern void a(const char *);
-// CHECK: @__cpu_model = external dso_local global { i32, i32, i32, [1 x i32] }
-// CHECK: @__cpu_features2 = external dso_local global [3 x i32]
+// CHECK-X86: @__cpu_model = external dso_local global { i32, i32, i32, [1 x i32] }
+// CHECK-X86: @__cpu_features2 = external dso_local global [3 x i32]
int main(void) {
__builtin_cpu_init();
@@ -15,38 +20,57 @@ int main(void) {
if (__builtin_cpu_supports("sse4.2"))
a("sse4.2");
- // CHECK: [[LOAD:%[^ ]+]] = load i32, ptr getelementptr inbounds ({ i32, i32, i32, [1 x i32] }, ptr @__cpu_model, i32 0, i32 3, i32 0)
- // CHECK: [[AND:%[^ ]+]] = and i32 [[LOAD]], 256
- // CHECK: = icmp eq i32 [[AND]], 256
+ // CHECK-X86: [[LOAD:%[^ ]+]] = load i32, ptr getelementptr inbounds ({ i32, i32, i32, [1 x i32] }, ptr @__cpu_model, i32 0, i32 3, i32 0)
+ // CHECK-X86: [[AND:%[^ ]+]] = and i32 [[LOAD]], 256
+ // CHECK-X86: = icmp eq i32 [[AND]], 256
if (__builtin_cpu_supports("gfni"))
a("gfni");
- // CHECK: [[LOAD:%[^ ]+]] = load i32, ptr @__cpu_features2
- // CHECK: [[AND:%[^ ]+]] = and i32 [[LOAD]], 1
- // CHECK: = icmp eq i32 [[AND]], 1
+ // CHECK-X86: [[LOAD:%[^ ]+]] = load i32, ptr @__cpu_features2
+ // CHECK-X86: [[AND:%[^ ]+]] = and i32 [[LOAD]], 1
+ // CHECK-X86: = icmp eq i32 [[AND]], 1
return 0;
}
-// CHECK: declare dso_local void @__cpu_indicator_init()
+// CHECK-X86: declare dso_local void @__cpu_indicator_init()
-// CHECK-LABEL: define{{.*}} @baseline(
-// CHECK: [[LOAD:%.*]] = load i32, ptr getelementptr inbounds ([[[#]] x i32], ptr @__cpu_features2, i32 0, i32 1)
-// CHECK-NEXT: and i32 [[LOAD]], -2147483648
+// CHECK-X86-LABEL: define{{.*}} @baseline(
+// CHECK-X86: [[LOAD:%.*]] = load i32, ptr getelementptr inbounds ([[[#]] x i32], ptr @__cpu_features2, i32 0, i32 1)
+// CHECK-X86-NEXT: and i32 [[LOAD]], -2147483648
int baseline() { return __builtin_cpu_supports("x86-64"); }
-// CHECK-LABEL: define{{.*}} @v2(
-// CHECK: [[LOAD:%.*]] = load i32, ptr getelementptr inbounds ([[[#]] x i32], ptr @__cpu_features2, i32 0, i32 2)
-// CHECK-NEXT: and i32 [[LOAD]], 1
+// CHECK-X86-LABEL: define{{.*}} @v2(
+// CHECK-X86: [[LOAD:%.*]] = load i32, ptr getelementptr inbounds ([[[#]] x i32], ptr @__cpu_features2, i32 0, i32 2)
+// CHECK-X86-NEXT: and i32 [[LOAD]], 1
int v2() { return __builtin_cpu_supports("x86-64-v2"); }
-// CHECK-LABEL: define{{.*}} @v3(
-// CHECK: [[LOAD:%.*]] = load i32, ptr getelementptr inbounds ([[[#]] x i32], ptr @__cpu_features2, i32 0, i32 2)
-// CHECK-NEXT: and i32 [[LOAD]], 2
+// CHECK-X86-LABEL: define{{.*}} @v3(
+// CHECK-X86: [[LOAD:%.*]] = load i32, ptr getelementptr inbounds ([[[#]] x i32], ptr @__cpu_features2, i32 0, i32 2)
+// CHECK-X86-NEXT: and i32 [[LOAD]], 2
int v3() { return __builtin_cpu_supports("x86-64-v3"); }
-// CHECK-LABEL: define{{.*}} @v4(
-// CHECK: [[LOAD:%.*]] = load i32, ptr getelementptr inbounds ([[[#]] x i32], ptr @__cpu_features2, i32 0, i32 2)
-// CHECK-NEXT: and i32 [[LOAD]], 4
+// CHECK-X86-LABEL: define{{.*}} @v4(
+// CHECK-X86: [[LOAD:%.*]] = load i32, ptr getelementptr inbounds ([[[#]] x i32], ptr @__cpu_features2, i32 0, i32 2)
+// CHECK-X86-NEXT: and i32 [[LOAD]], 4
int v4() { return __builtin_cpu_supports("x86-64-v4"); }
+#else
+int test(int a) {
+// CHECK-PPC: [[CPUSUP:%[^ ]+]] = call i32 @llvm.ppc.fixed.addr.ld(i32 2)
+// CHECK-PPC: [[AND:%[^ ]+]] = and i32 [[CPUSUP]], 8388608
+// CHECK-PPC: icmp ne i32 [[AND]], 0
+// CHECK-PPC: [[CPUSUP2:%[^ ]+]] = call i32 @llvm.ppc.fixed.addr.ld(i32 1)
+// CHECK-PPC: [[AND2:%[^ ]+]] = and i32 [[CPUSUP2]], 67108864
+// CHECK-PPC: icmp ne i32 [[AND2]], 0
+// CHECK-PPC: [[CPUID:%[^ ]+]] = call i32 @llvm.ppc.fixed.addr.ld(i32 3)
+// CHECK-PPC: icmp eq i32 [[CPUID]], 39
+ if (__builtin_cpu_supports("arch_3_00")) // HWCAP2
+ return a;
+ else if (__builtin_cpu_supports("mmu")) // HWCAP
+ return a - 5;
+ else if (__builtin_cpu_is("power7")) // CPUID
+ return a + a;
+ return a + 5;
+}
+#endif
diff --git a/clang/test/Sema/builtin-cpu-supports.c b/clang/test/Sema/builtin-cpu-supports.c
index ad310128fecebf..cc6f1beb5d8a7c 100644
--- a/clang/test/Sema/builtin-cpu-supports.c
+++ b/clang/test/Sema/builtin-cpu-supports.c
@@ -1,5 +1,5 @@
// RUN: %clang_cc1 -fsyntax-only -triple x86_64-pc-linux-gnu -verify %s
-// RUN: %clang_cc1 -fsyntax-only -triple powerpc64le-linux-gnu -verify %s
+// RUN: %clang_cc1 -fsyntax-only -triple aarch64-linux-gnu -verify %s
extern void a(const char *);
@@ -27,11 +27,13 @@ int main(void) {
(void)__builtin_cpu_supports("x86-64-v4");
(void)__builtin_cpu_supports("x86-64-v5"); // expected-error {{invalid cpu feature string for builtin}}
#else
- if (__builtin_cpu_supports("vsx")) // expected-error {{use of unknown builtin}}
+ if (__builtin_cpu_supports("aes")) // expected-error {{builtin is not supported on this target}}
a("vsx");
- if (__builtin_cpu_is("pwr9")) // expected-error {{use of unknown builtin}}
+ if (__builtin_cpu_is("cortex-x3")) // expected-error {{builtin is not supported on this target}}
a("pwr9");
+
+ __builtin_cpu_init(); // expected-error {{builtin is not supported on this target}}
#endif
return 0;
diff --git a/llvm/include/llvm/IR/IntrinsicsPowerPC.td b/llvm/include/llvm/IR/IntrinsicsPowerPC.td
index 6d1e8eb47405dd..d1abc37e072f70 100644
--- a/llvm/include/llvm/IR/IntrinsicsPowerPC.td
+++ b/llvm/include/llvm/IR/IntrinsicsPowerPC.td
@@ -215,6 +215,12 @@ let TargetPrefix = "ppc" in { // All intrinsics start with "llvm.ppc.".
[llvm_float_ty],
[llvm_float_ty, llvm_float_ty, llvm_float_ty, llvm_vararg_ty],
[IntrNoMem]>;
+ // Load of a value provided by the system library at a fixed address. Used for
+ // accessing things like HWCAP word provided by GLIBC.
+ def int_ppc_fixed_addr_ld
+ : DefaultAttrsIntrinsic<[llvm_i32_ty], [llvm_i32_ty],
+ [IntrInaccessibleMemOnly, ImmArg<ArgIndex<0>>]>;
+
}
let TargetPrefix = "ppc" in { // All PPC intrinsics start with "llvm.ppc.".
diff --git a/llvm/include/llvm/TargetParser/PPCTargetParser.def b/llvm/include/llvm/TargetParser/PPCTargetParser.def
new file mode 100644
index 00000000000000..e935e0112956fe
--- /dev/null
+++ b/llvm/include/llvm/TargetParser/PPCTargetParser.def
@@ -0,0 +1,80 @@
+#ifndef PPC_FEATURE
+#define PPC_FEATURE(NAME, DESC, ENUMNAME, ENUMVAL, HWCAPN)
+#endif
+#ifndef PPC_CPU
+#define PPC_CPU(NAME, NUM)
+#endif
+#ifndef PPC_FAWORD_HWCAP
+#define PPC_FAWORD_HWCAP 1
+#endif
+#ifndef PPC_FAWORD_HWCAP2
+#define PPC_FAWORD_HWCAP2 2
+#endif
+#ifndef PPC_FAWORD_CPUID
+#define PPC_FAWORD_CPUID 3
+#endif
+
+// PPC_FEATURE(Name, Description, EnumName, BitMask, PPC_FAWORD_WORD)
+PPC_FEATURE("4xxmac","4xx CPU has a Multiply Accumulator",PPCF_4XXMAC,0x02000000,PPC_FAWORD_HWCAP)
+PPC_FEATURE("altivec","CPU has a SIMD/Vector Unit",PPCF_ALTIVEC,0x10000000,PPC_FAWORD_HWCAP)
+PPC_FEATURE("arch_2_05","CPU supports ISA 205 (eg, POWER6)",PPCF_ARCH205,0x00001000,PPC_FAWORD_HWCAP)
+PPC_FEATURE("arch_2_06","CPU supports ISA 206 (eg, POWER7)",PPCF_ARCH206,0x00000100,PPC_FAWORD_HWCAP)
+PPC_FEATURE("arch_2_07","CPU supports ISA 207 (eg, POWER8)",PPCF_ARCH207,0x80000000,PPC_FAWORD_HWCAP2)
+PPC_FEATURE("arch_3_00","CPU supports ISA 30 (eg, POWER9)",PPCF_ARCH30,0x00800000,PPC_FAWORD_HWCAP2)
+PPC_FEATURE("arch_3_1","CPU supports ISA 31 (eg, POWER10)",PPCF_ARCH31,0x00040000,PPC_FAWORD_HWCAP2)
+PPC_FEATURE("archpmu","CPU supports the set of compatible performance monitoring events",PPCF_ARCHPMU,0x00000040,PPC_FAWORD_HWCAP)
+PPC_FEATURE("booke","CPU supports the Embedded ISA category",PPCF_BOOKE,0x00008000,PPC_FAWORD_HWCAP)
+PPC_FEATURE("cellbe","CPU has a CELL broadband engine",PPCF_CELLBE,0x00010000,PPC_FAWORD_HWCAP)
+PPC_FEATURE("darn","CPU supports the darn (deliver a random number) instruction",PPCF_DARN,0x00200000,PPC_FAWORD_HWCAP2)
+PPC_FEATURE("dfp","CPU has a decimal floating point unit",PPCF_DFP,0x00000400,PPC_FAWORD_HWCAP)
+PPC_FEATURE("dscr","CPU supports the data stream control register",PPCF_DSCR,0x20000000,PPC_FAWORD_HWCAP2)
+PPC_FEATURE("ebb","CPU supports event base branching",PPCF_EBB,0x10000000,PPC_FAWORD_HWCAP2)
+PPC_FEATURE("efpdouble","CPU has a SPE double precision floating point unit",PPCF_EFPDOUBLE,0x00200000,PPC_FAWORD_HWCAP)
+PPC_FEATURE("efpsingle","CPU has a SPE single precision floating point unit",PPCF_EFPSINGLE,0x00400000,PPC_FAWORD_HWCAP)
+PPC_FEATURE("fpu","CPU has a floating point unit",PPCF_FPU,0x08000000,PPC_FAWORD_HWCAP)
+PPC_FEATURE("htm","CPU has hardware transaction memory instructions",PPCF_HTM,0x40000000,PPC_FAWORD_HWCAP2)
+PPC_FEATURE("htm-nosc","Kernel aborts hardware transactions when a syscall is made",PPCF_HTM_NOSC,0x01000000,PPC_FAWORD_HWCAP2)
+PPC_FEATURE("htm-no-suspend","CPU supports hardware transaction memory but does not support the tsuspend instruction.",PPCF_HTM_NO_SUSPEND,0x00080000,PPC_FAWORD_HWCAP2)
+PPC_FEATURE("ic_snoop","CPU supports icache snooping capabilities",PPCF_IC_SNOOP,0x00002000,PPC_FAWORD_HWCAP)
+PPC_FEATURE("ieee128","CPU supports 128-bit IEEE binary floating point instructions",PPCF_IEEE128,0x00400000,PPC_FAWORD_HWCAP2)
+PPC_FEATURE("isel","CPU supports the integer select instruction",PPCF_ISEL,0x08000000,PPC_FAWORD_HWCAP2)
+PPC_FEATURE("mma","CPU supports the matrix-multiply assist instructions",PPCF_MMA,0x00020000,PPC_FAWORD_HWCAP2)
+PPC_FEATURE("mmu","CPU has a memory management unit",PPCF_MMU,0x04000000,PPC_FAWORD_HWCAP)
+PPC_FEATURE("notb","CPU does not have a timebase (eg, 601 and 403gx)",PPCF_NOTB,0x00100000,PPC_FAWORD_HWCAP)
+PPC_FEATURE("pa6t","CPU supports the PA Semi 6T CORE ISA",PPCF_PA6T,0x00000800,PPC_FAWORD_HWCAP)
+PPC_FEATURE("power4","CPU supports ISA 200 (eg, POWER4)",PPCF_POWER4,0x00080000,PPC_FAWORD_HWCAP)
+PPC_FEATURE("power5","CPU supports ISA 202 (eg, POWER5)",PPCF_POWER5,0x00040000,PPC_FAWORD_HWCAP)
+PPC_FEATURE("power5+","CPU supports ISA 203 (eg, POWER5+)",PPCF_POWER5P,0x00020000,PPC_FAWORD_HWCAP)
+PPC_FEATURE("power6x","CPU supports ISA 205 (eg, POWER6) extended opcodes mffgpr and mftgpr.",PPCF_POWER6X,0x00000200,PPC_FAWORD_HWCAP)
+PPC_FEATURE("ppc32","CPU supports 32-bit mode execution",PPCF_PPC32,0x80000000,PPC_FAWORD_HWCAP)
+PPC_FEATURE("ppc601","CPU supports the old POWER ISA (eg, 601)",PPCF_PPC601,0x20000000,PPC_FAWORD_HWCAP)
+PPC_FEATURE("ppc64","CPU supports 64-bit mode execution",PPCF_PPC64,0x40000000,PPC_FAWORD_HWCAP)
+PPC_FEATURE("ppcle","CPU supports a little-endian mode that uses address swizzling",PPCF_PPCLE,0x00000001,PPC_FAWORD_HWCAP)
+PPC_FEATURE("scv","Kernel supports system call vectored",PPCF_SCV,0x00100000,PPC_FAWORD_HWCAP2)
+PPC_FEATURE("smt","CPU support simultaneous multi-threading",PPCF_SMT,0x00004000,PPC_FAWORD_HWCAP)
+PPC_FEATURE("spe","CPU has a signal processing extension unit",PPCF_SPE,0x00800000,PPC_FAWORD_HWCAP)
+PPC_FEATURE("tar","CPU supports the target address register",PPCF_TAR,0x04000000,PPC_FAWORD_HWCAP2)
+PPC_FEATURE("true_le","CPU supports true little-endian mode",PPCF_TRUE_LE,0x00000002,PPC_FAWORD_HWCAP)
+PPC_FEATURE("ucache","CPU has unified I/D cache",PPCF_UCACHE,0x01000000,PPC_FAWORD_HWCAP)
+PPC_FEATURE("vcrypto","CPU supports the vector cryptography instructions",PPCF_VCRYPTO,0x02000000,PPC_FAWORD_HWCAP2)
+PPC_FEATURE("vsx","CPU supports the vector-scalar extension",PPCF_VSX,0x00000080,PPC_FAWORD_HWCAP)
+
+// PPC_CPU(Name, NumericID)
+PPC_CPU("power4",32)
+PPC_CPU("ppc970",33)
+PPC_CPU("power5",34)
+PPC_CPU("power5+",35)
+PPC_CPU("power6",36)
+PPC_CPU("ppc-cell-be",37)
+PPC_CPU("power6x",38)
+PPC_CPU("power7",39)
+PPC_CPU("ppca2",40)
+PPC_CPU("ppc405",41)
+PPC_CPU("ppc440",42)
+PPC_CPU("ppc464",43)
+PPC_CPU("ppc476",44)
+PPC_CPU("power8",45)
+PPC_CPU("power9",46)
+PPC_CPU("power10",47)
+#undef PPC_FEATURE
+#undef PPC_CPU
diff --git a/llvm/lib/Target/PowerPC/PPCAsmPrinter.cpp b/llvm/lib/Target/PowerPC/PPCAsmPrinter.cpp
index 780b22b4fbe65e..46afb882d80c5d 100644
--- a/llvm/lib/Target/PowerPC/PPCAsmPrinter.cpp
+++ b/llvm/lib/Target/PowerPC/PPCAsmPrinter.cpp
@@ -1821,6 +1821,10 @@ void PPCLinuxAsmPrinter::emitEndOfAsmFile(Module &M) {
PPCTargetStreamer *TS =
static_cast<PPCTargetStreamer *>(OutStreamer->getTargetStreamer());
+ if (static_cast<const PPCTargetMachine &>(TM).hasGlibcHWCAPAccess())
+ OutStreamer->emitSymbolValue(
+ GetExternalSymbolSymbol("__parse_hwcap_and_convert_at_platform"),
+ MAI->getCodePointerSize());
emitGNUAttributes(M);
if (!TOC.empty()) {
diff --git a/llvm/lib/Target/PowerPC/PPCInstrInfo.cpp b/llvm/lib/Target/PowerPC/PPCInstrInfo.cpp
index 15f6b65dea83c4..77fe34a6cbbcb3 100644
--- a/llvm/lib/Target/PowerPC/PPCInstrInfo.cpp
+++ b/llvm/lib/Target/PowerPC/PPCInstrInfo.cpp
@@ -1079,6 +1079,7 @@ bool PPCInstrInfo::isReallyTriviallyReMaterializable(
case PPC::ADDIStocHA8:
case PPC::ADDItocL:
case PPC::LOAD_STACK_GUARD:
+ case PPC::PPCLdFixedAddr:
case PPC::XXLXORz:
case PPC::XXLXORspz:
case PPC::XXLXORdpz:
@@ -3099,6 +3100,38 @@ bool PPCInstrInfo::expandPostRAPseudo(MachineInstr &MI) const {
.addReg(Reg);
return true;
}
+ case PPC::PPCLdFixedAddr: {
+ assert(Subtarget.isTargetLinux() &&
+ "Only Linux target is expected to contain PPCLdFixedAddr");
+ int64_t Offset = 0;
+ const unsigned Reg = Subtarget.isPPC64() ? PPC::X13 : PPC::R2;
+ MI.setDesc(get(PPC::LWZ));
+ uint64_t FAType = MI.getOperand(1).getImm();
+#undef PPC_FEATURE
+#undef PPC_CPU
+#include "llvm/TargetParser/PPCTargetParser.def"
+ // The HWCAP and HWCAP2 word offsets are reversed on big endian Linux.
+ if ((FAType == PPC_FAWORD_HWCAP && Subtarget.isLittleEndian()) ||
+ (FAType == PPC_FAWORD_HWCAP2 && !Subtarget.isLittleEndian()))
+ Offset = Subtarget.isPPC64() ? -0x7064 : -0x703C;
+ else if ((FAType == PPC_FAWORD_HWCAP2 &&
+ Subtarget.isLittleEndian()) ||
+ (FAType == PPC_FAWORD_HWCAP &&
+ !Subtarget.isLittleEndian()))
+ Offset = Subtarget.isPPC64() ? -0x7068 : -0x7040;
+ else if (FAType == PPC_FAWORD_CPUID)
+ Offset = Subtarget.isPPC64() ? -0x705C : -0x7034;
+ assert(Offset && "Do not know the offset for this fixed addr load");
+ MI.removeOperand(1);
+ Subtarget.getTargetMachine().setGlibcHWCAPAccess();
+ MachineInstrBuilder(*MI.getParent()->getParent(), MI)
+ .addImm(Offset)
+ .addReg(Reg);
+ return true;
+#undef PPC_FAWORD_HWCAP
+#undef PPC_FAWORD_HWCAP2
+#undef PPC_FAWORD_CPUID
+ }
case PPC::DFLOADf32:
case PPC::DFLOADf64:
case PPC::DFSTOREf32:
diff --git a/llvm/lib/Target/PowerPC/PPCInstrInfo.td b/llvm/lib/Target/PowerPC/PPCInstrInfo.td
index 5550ba4207392b..3abd97f2c38c09 100644
--- a/llvm/lib/Target/PowerPC/PPCInstrInfo.td
+++ b/llvm/lib/Target/PowerPC/PPCInstrInfo.td
@@ -4830,6 +4830,9 @@ def RLWNMbm : PPCAsmPseudo<"rlwnm $rA, $rS, $n, $b",
(ins g8rc:$rA, g8rc:$rS, u5imm:$n, i32imm:$b)>;
def RLWNMbm_rec : PPCAsmPseudo<"rlwnm. $rA, $rS, $n, $b",
(ins g8rc:$rA, g8rc:$rS, u5imm:$n, i32imm:$b)>;
+def PPCLdFixedAddr :
+ PPCPostRAExpPseudo<(outs gprc:$rT), (ins i32imm:$imm), "#FA_LOAD",
+ [(set i32:$rT, (int_ppc_fixed_addr_ld timm:$imm))]>;
// These generic branch instruction forms are used for the assembler parser only.
// Defs and Uses are conservative, since we don't know the BO value.
diff --git a/llvm/lib/Target/PowerPC/PPCTargetMachine.h b/llvm/lib/Target/PowerPC/PPCTargetMachine.h
index 56145a2eb39ce4..9d0d3e727170a3 100644
--- a/llvm/lib/Target/PowerPC/PPCTargetMachine.h
+++ b/llvm/lib/Target/PowerPC/PPCTargetMachine.h
@@ -32,6 +32,7 @@ class PPCTargetMachine final : public LLVMTargetMachine {
std::unique_ptr<TargetLoweringObjectFile> TLOF;
PPCABI TargetABI;
Endian Endianness = Endian::NOT_DETECTED;
+ mutable bool HasGlibcHWCAPAccess = false;
mutable StringMap<std::unique_ptr<PPCSubtarget>> SubtargetMap;
@@ -64,6 +65,8 @@ class PPCTargetMachine final : public LLVMTargetMachine {
const TargetSubtargetInfo *STI) const override;
bool isELFv2ABI() const { return TargetABI == PPC_ABI_ELFv2; }
+ bool hasGlibcHWCAPAccess() const { return HasGlibcHWCAPAccess; }
+ void setGlibcHWCAPAccess(bool Val = true) const { HasGlibcHWCAPAccess = Val; }
bool isPPC64() const {
const Triple &TT = getTargetTriple();
return (TT.getArch() == Triple::ppc64 || TT.getArch() == Triple::ppc64le);
diff --git a/llvm/test/CodeGen/PowerPC/cpu-supports.ll b/llvm/test/CodeGen/PowerPC/cpu-supports.ll
new file mode 100644
index 00000000000000..912d67e5c555e6
--- /dev/null
+++ b/llvm/test/CodeGen/PowerPC/cpu-supports.ll
@@ -0,0 +1,111 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 2
+; RUN: llc -mcpu=pwr9 -ppc-asm-full-reg-names \
+; RUN: -mtriple=powerpc64-linux-gnu < %s | FileCheck %s \
+; RUN: -check-prefix=BE64
+; RUN: llc -mcpu=pwr9 -ppc-asm-full-reg-names \
+; RUN: -mtriple=powerpc-linux-gnu < %s | FileCheck %s \
+; RUN: -check-prefix=BE32
+; RUN: llc -mcpu=pwr9 -ppc-asm-full-reg-names \
+; RUN: -mtriple=powerpc64le-linux-gnu < %s | FileCheck %s \
+; RUN: -check-prefix=LE
+define dso_local signext i32 @test(i32 noundef signext %a) local_unnamed_addr #0 {
+; BE64-LABEL: test:
+; BE64: # %bb.0: # %entry
+; BE64-NEXT: lwz r4, -28772(r13)
+; BE64-NEXT: andis. r4, r4, 128
+; BE64-NEXT: bne cr0, .LBB0_3
+; BE64-NEXT: # %bb.1: # %if.else
+; BE64-NEXT: lwz r4, -28776(r13)
+; BE64-NEXT: andis. r4, r4, 1024
+; BE64-NEXT: bne cr0, .LBB0_4
+; BE64-NEXT: # %bb.2: # %if.else3
+; BE64-NEXT: lwz r4, -28764(r13)
+; BE64-NEXT: cmplwi r4, 39
+; BE64-NEXT: addi r4, r3, 5
+; BE64-NEXT: slwi r3, r3, 1
+; BE64-NEXT: iseleq r3, r3, r4
+; BE64-NEXT: .LBB0_3: # %return
+; BE64-NEXT: extsw r3, r3
+; BE64-NEXT: blr
+; BE64-NEXT: .LBB0_4: # %if.then2
+; BE64-NEXT: addi r3, r3, -5
+; BE64-NEXT: extsw r3, r3
+; BE64-NEXT: blr
+;
+; BE32-LABEL: test:
+; BE32: # %bb.0: # %entry
+; BE32-NEXT: lwz r4, -28732(r2)
+; BE32-NEXT: andis. r4, r4, 128
+; BE32-NEXT: bnelr cr0
+; BE32-NEXT: # %bb.1: # %if.else
+; BE32-NEXT: lwz r4, -28736(r2)
+; BE32-NEXT: andis. r4, r4, 1024
+; BE32-NEXT: bne cr0, .LBB0_3
+; BE32-NEXT: # %bb.2: # %if.else3
+; BE32-NEXT: lwz r4, -28724(r2)
+; BE32-NEXT: cmplwi r4, 39
+; BE32-NEXT: addi r4, r3, 5
+; BE32-NEXT: slwi r3, r3, 1
+; BE32-NEXT: iseleq r3, r3, r4
+; BE32-NEXT: blr
+; BE32-NEXT: .LBB0_3: # %if.then2
+; BE32-NEXT: addi r3, r3, -5
+; BE32-NEXT: blr
+;
+; LE-LABEL: test:
+; LE: # %bb.0: # %entry
+; LE-NEXT: lwz r4, -28776(r13)
+; LE-NEXT: andis. r4, r4, 128
+; LE-NEXT: bne cr0, .LBB0_3
+; LE-NEXT: # %bb.1: # %if.else
+; LE-NEXT: lwz r4, -28772(r13)
+; LE-NEXT: andis. r4, r4, 1024
+; LE-NEXT: bne cr0, .LBB0_4
+; LE-NEXT: # %bb.2: # %if.else3
+; LE-NEXT: lwz r4, -28764(r13)
+; LE-NEXT: cmplwi r4, 39
+; LE-NEXT: addi r4, r3, 5
+; LE-NEXT: slwi r3, r3, 1
+; LE-NEXT: iseleq r3, r3, r4
+; LE-NEXT: .LBB0_3: # %return
+; LE-NEXT: extsw r3, r3
+; LE-NEXT: blr
+; LE-NEXT: .LBB0_4: # %if.then2
+; LE-NEXT: addi r3, r3, -5
+; LE-NEXT: extsw r3, r3
+; LE-NEXT: blr
+entry:
+ %cpu_supports = tail call i32 @llvm.ppc.fixed.addr.ld(i32 2)
+ %0 = and i32 %cpu_supports, 8388608
+ %.not = icmp eq i32 %0, 0
+ br i1 %.not, label %if.else, label %return
+
+if.else: ; preds = %entry
+ %cpu_supports1 = tail call i32 @llvm.ppc.fixed.addr.ld(i32 1)
+ %1 = and i32 %cpu_supports1, 67108864
+ %.not12 = icmp eq i32 %1, 0
+ br i1 %.not12, label %if.else3, label %if.then2
+
+if.then2: ; preds = %if.else
+ %sub = add nsw i32 %a, -5
+ br label %return
+
+if.else3: ; preds = %if.else
+ %cpu_is = tail call i32 @llvm.ppc.fixed.addr.ld(i32 3)
+ %2 = icmp eq i32 %cpu_is, 39
+ br i1 %2, label %if.then4, label %if.end6
+
+if.then4: ; preds = %if.else3
+ %add = shl nsw i32 %a, 1
+ br label %return
+
+if.end6: ; preds = %if.else3
+ %add7 = add nsw i32 %a, 5
+ br label %return
+
+return: ; preds = %entry, %if.end6, %if.then4, %if.then2
+ %retval.0 = phi i32 [ %sub, %if.then2 ], [ %add, %if.then4 ], [ %add7, %if.end6 ], [ %a, %entry ]
+ ret i32 %retval.0
+}
+
+declare i32 @llvm.ppc.fixed.addr.ld(i32 immarg) #1
>From 60e1aba57ad08d69349bb9fe768e6fd3abf832bc Mon Sep 17 00:00:00 2001
From: Nemanja Ivanovic <nemanjaivanovic at nemanjas-air.kpn>
Date: Thu, 12 Oct 2023 14:32:07 -0400
Subject: [PATCH 2/5] Run clang-format which I appear to have forgotten to run
---
clang/lib/CodeGen/CGBuiltin.cpp | 3 +--
llvm/lib/Target/PowerPC/PPCInstrInfo.cpp | 6 ++----
2 files changed, 3 insertions(+), 6 deletions(-)
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 3c9d8155569fac..fa6859d6921585 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -16552,8 +16552,7 @@ Value *CodeGenFunction::EmitPPCBuiltinExpr(unsigned BuiltinID,
#define PPC_CPU(Name, NumericID) .Case(Name, NumericID)
#include "llvm/TargetParser/PPCTargetParser.def"
.Default(-1U);
- Value *Op0 =
- llvm::ConstantInt::get(Int32Ty, PPC_FAWORD_CPUID);
+ Value *Op0 = llvm::ConstantInt::get(Int32Ty, PPC_FAWORD_CPUID);
llvm::Function *F = CGM.getIntrinsic(Intrinsic::ppc_fixed_addr_ld);
Value *TheCall = Builder.CreateCall(F, {Op0}, "cpu_is");
return Builder.CreateICmpEQ(TheCall,
diff --git a/llvm/lib/Target/PowerPC/PPCInstrInfo.cpp b/llvm/lib/Target/PowerPC/PPCInstrInfo.cpp
index 77fe34a6cbbcb3..34d41d05089586 100644
--- a/llvm/lib/Target/PowerPC/PPCInstrInfo.cpp
+++ b/llvm/lib/Target/PowerPC/PPCInstrInfo.cpp
@@ -3114,10 +3114,8 @@ bool PPCInstrInfo::expandPostRAPseudo(MachineInstr &MI) const {
if ((FAType == PPC_FAWORD_HWCAP && Subtarget.isLittleEndian()) ||
(FAType == PPC_FAWORD_HWCAP2 && !Subtarget.isLittleEndian()))
Offset = Subtarget.isPPC64() ? -0x7064 : -0x703C;
- else if ((FAType == PPC_FAWORD_HWCAP2 &&
- Subtarget.isLittleEndian()) ||
- (FAType == PPC_FAWORD_HWCAP &&
- !Subtarget.isLittleEndian()))
+ else if ((FAType == PPC_FAWORD_HWCAP2 && Subtarget.isLittleEndian()) ||
+ (FAType == PPC_FAWORD_HWCAP && !Subtarget.isLittleEndian()))
Offset = Subtarget.isPPC64() ? -0x7068 : -0x7040;
else if (FAType == PPC_FAWORD_CPUID)
Offset = Subtarget.isPPC64() ? -0x705C : -0x7034;
>From dd421e7e6aba4886b6d6d7231d5ec3e456640da8 Mon Sep 17 00:00:00 2001
From: Nemanja Ivanovic <nemanja at synopsys.com>
Date: Thu, 11 Jan 2024 13:23:28 +0100
Subject: [PATCH 3/5] Address comments
- Use script for generating test checks
- Clean up code and rename macros for clarity
- Remove magic number offsets
- Get rid of auxiliary targets in Sema checking
- Clarify some comments and add missing ones
- Add check for undefined symbol added for safety
---
clang/lib/Basic/Targets/PPC.cpp | 4 +-
clang/lib/CodeGen/CGBuiltin.cpp | 4 +-
clang/lib/Sema/SemaChecking.cpp | 71 ++-----
clang/test/CodeGen/builtin-cpu-supports.c | 149 +++++++++++---
llvm/include/llvm/IR/IntrinsicsPowerPC.td | 5 +-
.../llvm/TargetParser/PPCTargetParser.def | 183 +++++++++++-------
llvm/lib/Target/PowerPC/PPCAsmPrinter.cpp | 4 +
llvm/lib/Target/PowerPC/PPCInstrInfo.cpp | 42 ++--
llvm/test/CodeGen/PowerPC/cpu-supports.ll | 3 +
9 files changed, 294 insertions(+), 171 deletions(-)
diff --git a/clang/lib/Basic/Targets/PPC.cpp b/clang/lib/Basic/Targets/PPC.cpp
index 2cf1bacd95fd95..c03350ceb6d965 100644
--- a/clang/lib/Basic/Targets/PPC.cpp
+++ b/clang/lib/Basic/Targets/PPC.cpp
@@ -880,14 +880,14 @@ ArrayRef<Builtin::Info> PPCTargetInfo::getTargetBuiltins() const {
}
bool PPCTargetInfo::validateCpuSupports(StringRef FeatureStr) const {
-#define PPC_FEATURE(NAME, DESC, ENUMNAME, ENUMVAL, HWCAPN) .Case(NAME, true)
+#define PPC_LNX_FEATURE(NAME, DESC, ENUMNAME, ENUMVAL, HWCAPN) .Case(NAME, true)
return llvm::StringSwitch<bool>(FeatureStr)
#include "llvm/TargetParser/PPCTargetParser.def"
.Default(false);
}
bool PPCTargetInfo::validateCpuIs(StringRef CPUName) const {
-#define PPC_CPU(NAME, NUM) .Case(NAME, true)
+#define PPC_LNX_CPU(NAME, NUM) .Case(NAME, true)
return llvm::StringSwitch<bool>(CPUName)
#include "llvm/TargetParser/PPCTargetParser.def"
.Default(false);
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index fa6859d6921585..e5b15a74da392b 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -16549,7 +16549,7 @@ Value *CodeGenFunction::EmitPPCBuiltinExpr(unsigned BuiltinID,
const Expr *CPUExpr = E->getArg(0)->IgnoreParenCasts();
StringRef CPUStr = cast<clang::StringLiteral>(CPUExpr)->getString();
unsigned NumCPUID = StringSwitch<unsigned>(CPUStr)
-#define PPC_CPU(Name, NumericID) .Case(Name, NumericID)
+#define PPC_LNX_CPU(Name, NumericID) .Case(Name, NumericID)
#include "llvm/TargetParser/PPCTargetParser.def"
.Default(-1U);
Value *Op0 = llvm::ConstantInt::get(Int32Ty, PPC_FAWORD_CPUID);
@@ -16565,7 +16565,7 @@ Value *CodeGenFunction::EmitPPCBuiltinExpr(unsigned BuiltinID,
StringRef CPUStr = cast<clang::StringLiteral>(CPUExpr)->getString();
std::tie(FeatureWord, BitMask) =
StringSwitch<std::pair<unsigned, unsigned>>(CPUStr)
-#define PPC_FEATURE(Name, Description, EnumName, Bitmask, FA_WORD) \
+#define PPC_LNX_FEATURE(Name, Description, EnumName, Bitmask, FA_WORD) \
.Case(Name, {FA_WORD, Bitmask})
#include "llvm/TargetParser/PPCTargetParser.def"
.Default({0, 0});
diff --git a/clang/lib/Sema/SemaChecking.cpp b/clang/lib/Sema/SemaChecking.cpp
index 36305eeac8c9de..9edf568ec95f05 100644
--- a/clang/lib/Sema/SemaChecking.cpp
+++ b/clang/lib/Sema/SemaChecking.cpp
@@ -2143,61 +2143,35 @@ static bool checkFPMathBuiltinElementType(Sema &S, SourceLocation Loc,
return false;
}
-/// SemaBuiltinCpuSupports - Handle __builtin_cpu_supports(char *).
-/// This checks that the target supports __builtin_cpu_supports and
-/// that the string argument is constant and valid.
-static bool SemaBuiltinCpuSupports(Sema &S, const TargetInfo &TI,
- const TargetInfo *AuxTI, CallExpr *TheCall) {
- Expr *Arg = TheCall->getArg(0);
-
- const TargetInfo *TheTI = nullptr;
- if (TI.supportsCpuSupports())
- TheTI = &TI;
- else if (AuxTI && AuxTI->supportsCpuSupports())
- TheTI = AuxTI;
- else
+/// SemaBuiltinCpu{Supports|Is} - Handle __builtin_cpu_{supports|is}(char *).
+/// This checks that the target supports the builtin and that the string
+/// argument is constant and valid.
+static bool SemaBuiltinCpu(Sema &S, const TargetInfo &TI, CallExpr *TheCall,
+ unsigned BuiltinID) {
+ assert((BuiltinID == Builtin::BI__builtin_cpu_supports ||
+ BuiltinID == Builtin::BI__builtin_cpu_is) &&
+ "Expecting __builtin_cpu_...");
+
+ bool IsCPUSupports = BuiltinID == Builtin::BI__builtin_cpu_supports;
+ if (IsCPUSupports && !TI.supportsCpuSupports())
return S.Diag(TheCall->getBeginLoc(), diag::err_builtin_target_unsupported)
<< SourceRange(TheCall->getBeginLoc(), TheCall->getEndLoc());
-
- // Check if the argument is a string literal.
- if (!isa<StringLiteral>(Arg->IgnoreParenImpCasts()))
- return S.Diag(TheCall->getBeginLoc(), diag::err_expr_not_string_literal)
- << Arg->getSourceRange();
-
- // Check the contents of the string.
- StringRef Feature =
- cast<StringLiteral>(Arg->IgnoreParenImpCasts())->getString();
- if (!TheTI->validateCpuSupports(Feature))
- return S.Diag(TheCall->getBeginLoc(), diag::err_invalid_cpu_supports)
- << Arg->getSourceRange();
- return false;
-}
-
-/// SemaBuiltinCpuIs - Handle __builtin_cpu_is(char *).
-/// This checks that the target supports __builtin_cpu_is and
-/// that the string argument is constant and valid.
-static bool SemaBuiltinCpuIs(Sema &S, const TargetInfo &TI,
- const TargetInfo *AuxTI, CallExpr *TheCall) {
- Expr *Arg = TheCall->getArg(0);
-
- const TargetInfo *TheTI = nullptr;
- if (TI.supportsCpuIs())
- TheTI = &TI;
- else if (AuxTI && AuxTI->supportsCpuIs())
- TheTI = AuxTI;
- else
+ if (!IsCPUSupports && !TI.supportsCpuIs())
return S.Diag(TheCall->getBeginLoc(), diag::err_builtin_target_unsupported)
<< SourceRange(TheCall->getBeginLoc(), TheCall->getEndLoc());
+ Expr *Arg = TheCall->getArg(0)->IgnoreParenImpCasts();
// Check if the argument is a string literal.
- if (!isa<StringLiteral>(Arg->IgnoreParenImpCasts()))
+ if (!isa<StringLiteral>(Arg))
return S.Diag(TheCall->getBeginLoc(), diag::err_expr_not_string_literal)
<< Arg->getSourceRange();
// Check the contents of the string.
- StringRef Feature =
- cast<StringLiteral>(Arg->IgnoreParenImpCasts())->getString();
- if (!TheTI->validateCpuIs(Feature))
+ StringRef Feature = cast<StringLiteral>(Arg)->getString();
+ if (IsCPUSupports && !TI.validateCpuSupports(Feature))
+ return S.Diag(TheCall->getBeginLoc(), diag::err_invalid_cpu_supports)
+ << Arg->getSourceRange();
+ if (!IsCPUSupports && !TI.validateCpuIs(Feature))
return S.Diag(TheCall->getBeginLoc(), diag::err_invalid_cpu_is)
<< Arg->getSourceRange();
return false;
@@ -2232,13 +2206,8 @@ Sema::CheckBuiltinFunctionCall(FunctionDecl *FDecl, unsigned BuiltinID,
FPOptions FPO;
switch (BuiltinID) {
case Builtin::BI__builtin_cpu_supports:
- if (SemaBuiltinCpuSupports(*this, Context.getTargetInfo(),
- Context.getAuxTargetInfo(), TheCall))
- return ExprError();
- break;
case Builtin::BI__builtin_cpu_is:
- if (SemaBuiltinCpuIs(*this, Context.getTargetInfo(),
- Context.getAuxTargetInfo(), TheCall))
+ if (SemaBuiltinCpu(*this, Context.getTargetInfo(), TheCall, BuiltinID))
return ExprError();
break;
case Builtin::BI__builtin_cpu_init:
diff --git a/clang/test/CodeGen/builtin-cpu-supports.c b/clang/test/CodeGen/builtin-cpu-supports.c
index 42497d8845437a..88eb7b0fa786e5 100644
--- a/clang/test/CodeGen/builtin-cpu-supports.c
+++ b/clang/test/CodeGen/builtin-cpu-supports.c
@@ -1,6 +1,7 @@
-// RUN: %clang_cc1 -triple x86_64-pc-linux-gnu -emit-llvm < %s | FileCheck %s \
-// RUN: --check-prefix=CHECK-X86
-// RUN: %clang_cc1 -triple ppc64le-linux-gnu -emit-llvm < %s | FileCheck %s \
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 3
+// RUN: %clang_cc1 -triple x86_64-pc-linux-gnu -emit-llvm -o - %s | \
+// RUN: FileCheck %s --check-prefix=CHECK-X86
+// RUN: %clang_cc1 -triple ppc64le-linux-gnu -emit-llvm -o - %s | FileCheck %s \
// RUN: --check-prefix=CHECK-PPC
#ifndef __PPC__
@@ -9,9 +10,33 @@
// global, the bit grab, and the icmp correct.
extern void a(const char *);
-// CHECK-X86: @__cpu_model = external dso_local global { i32, i32, i32, [1 x i32] }
-// CHECK-X86: @__cpu_features2 = external dso_local global [3 x i32]
+// CHECK-X86-LABEL: define dso_local i32 @main(
+// CHECK-X86-SAME: ) #[[ATTR0:[0-9]+]] {
+// CHECK-X86-NEXT: entry:
+// CHECK-X86-NEXT: [[RETVAL:%.*]] = alloca i32, align 4
+// CHECK-X86-NEXT: store i32 0, ptr [[RETVAL]], align 4
+// CHECK-X86-NEXT: call void @__cpu_indicator_init()
+// CHECK-X86-NEXT: [[TMP0:%.*]] = load i32, ptr getelementptr inbounds ({ i32, i32, i32, [1 x i32] }, ptr @__cpu_model, i32 0, i32 3, i32 0), align 4
+// CHECK-X86-NEXT: [[TMP1:%.*]] = and i32 [[TMP0]], 256
+// CHECK-X86-NEXT: [[TMP2:%.*]] = icmp eq i32 [[TMP1]], 256
+// CHECK-X86-NEXT: [[TMP3:%.*]] = and i1 true, [[TMP2]]
+// CHECK-X86-NEXT: br i1 [[TMP3]], label [[IF_THEN:%.*]], label [[IF_END:%.*]]
+// CHECK-X86: if.then:
+// CHECK-X86-NEXT: call void @a(ptr noundef @.str)
+// CHECK-X86-NEXT: br label [[IF_END]]
+// CHECK-X86: if.end:
+// CHECK-X86-NEXT: [[TMP4:%.*]] = load i32, ptr @__cpu_features2, align 4
+// CHECK-X86-NEXT: [[TMP5:%.*]] = and i32 [[TMP4]], 1
+// CHECK-X86-NEXT: [[TMP6:%.*]] = icmp eq i32 [[TMP5]], 1
+// CHECK-X86-NEXT: [[TMP7:%.*]] = and i1 true, [[TMP6]]
+// CHECK-X86-NEXT: br i1 [[TMP7]], label [[IF_THEN1:%.*]], label [[IF_END2:%.*]]
+// CHECK-X86: if.then1:
+// CHECK-X86-NEXT: call void @a(ptr noundef @.str.1)
+// CHECK-X86-NEXT: br label [[IF_END2]]
+// CHECK-X86: if.end2:
+// CHECK-X86-NEXT: ret i32 0
+//
int main(void) {
__builtin_cpu_init();
@@ -20,51 +45,111 @@ int main(void) {
if (__builtin_cpu_supports("sse4.2"))
a("sse4.2");
- // CHECK-X86: [[LOAD:%[^ ]+]] = load i32, ptr getelementptr inbounds ({ i32, i32, i32, [1 x i32] }, ptr @__cpu_model, i32 0, i32 3, i32 0)
- // CHECK-X86: [[AND:%[^ ]+]] = and i32 [[LOAD]], 256
- // CHECK-X86: = icmp eq i32 [[AND]], 256
if (__builtin_cpu_supports("gfni"))
a("gfni");
- // CHECK-X86: [[LOAD:%[^ ]+]] = load i32, ptr @__cpu_features2
- // CHECK-X86: [[AND:%[^ ]+]] = and i32 [[LOAD]], 1
- // CHECK-X86: = icmp eq i32 [[AND]], 1
return 0;
}
-// CHECK-X86: declare dso_local void @__cpu_indicator_init()
-// CHECK-X86-LABEL: define{{.*}} @baseline(
-// CHECK-X86: [[LOAD:%.*]] = load i32, ptr getelementptr inbounds ([[[#]] x i32], ptr @__cpu_features2, i32 0, i32 1)
-// CHECK-X86-NEXT: and i32 [[LOAD]], -2147483648
+// CHECK-X86-LABEL: define dso_local i32 @baseline(
+// CHECK-X86-SAME: ) #[[ATTR0]] {
+// CHECK-X86-NEXT: entry:
+// CHECK-X86-NEXT: [[TMP0:%.*]] = load i32, ptr getelementptr inbounds ([3 x i32], ptr @__cpu_features2, i32 0, i32 1), align 4
+// CHECK-X86-NEXT: [[TMP1:%.*]] = and i32 [[TMP0]], -2147483648
+// CHECK-X86-NEXT: [[TMP2:%.*]] = icmp eq i32 [[TMP1]], -2147483648
+// CHECK-X86-NEXT: [[TMP3:%.*]] = and i1 true, [[TMP2]]
+// CHECK-X86-NEXT: [[CONV:%.*]] = zext i1 [[TMP3]] to i32
+// CHECK-X86-NEXT: ret i32 [[CONV]]
+//
int baseline() { return __builtin_cpu_supports("x86-64"); }
-// CHECK-X86-LABEL: define{{.*}} @v2(
-// CHECK-X86: [[LOAD:%.*]] = load i32, ptr getelementptr inbounds ([[[#]] x i32], ptr @__cpu_features2, i32 0, i32 2)
-// CHECK-X86-NEXT: and i32 [[LOAD]], 1
+// CHECK-X86-LABEL: define dso_local i32 @v2(
+// CHECK-X86-SAME: ) #[[ATTR0]] {
+// CHECK-X86-NEXT: entry:
+// CHECK-X86-NEXT: [[TMP0:%.*]] = load i32, ptr getelementptr inbounds ([3 x i32], ptr @__cpu_features2, i32 0, i32 2), align 4
+// CHECK-X86-NEXT: [[TMP1:%.*]] = and i32 [[TMP0]], 1
+// CHECK-X86-NEXT: [[TMP2:%.*]] = icmp eq i32 [[TMP1]], 1
+// CHECK-X86-NEXT: [[TMP3:%.*]] = and i1 true, [[TMP2]]
+// CHECK-X86-NEXT: [[CONV:%.*]] = zext i1 [[TMP3]] to i32
+// CHECK-X86-NEXT: ret i32 [[CONV]]
+//
int v2() { return __builtin_cpu_supports("x86-64-v2"); }
-// CHECK-X86-LABEL: define{{.*}} @v3(
-// CHECK-X86: [[LOAD:%.*]] = load i32, ptr getelementptr inbounds ([[[#]] x i32], ptr @__cpu_features2, i32 0, i32 2)
-// CHECK-X86-NEXT: and i32 [[LOAD]], 2
+// CHECK-X86-LABEL: define dso_local i32 @v3(
+// CHECK-X86-SAME: ) #[[ATTR0]] {
+// CHECK-X86-NEXT: entry:
+// CHECK-X86-NEXT: [[TMP0:%.*]] = load i32, ptr getelementptr inbounds ([3 x i32], ptr @__cpu_features2, i32 0, i32 2), align 4
+// CHECK-X86-NEXT: [[TMP1:%.*]] = and i32 [[TMP0]], 2
+// CHECK-X86-NEXT: [[TMP2:%.*]] = icmp eq i32 [[TMP1]], 2
+// CHECK-X86-NEXT: [[TMP3:%.*]] = and i1 true, [[TMP2]]
+// CHECK-X86-NEXT: [[CONV:%.*]] = zext i1 [[TMP3]] to i32
+// CHECK-X86-NEXT: ret i32 [[CONV]]
+//
int v3() { return __builtin_cpu_supports("x86-64-v3"); }
-// CHECK-X86-LABEL: define{{.*}} @v4(
-// CHECK-X86: [[LOAD:%.*]] = load i32, ptr getelementptr inbounds ([[[#]] x i32], ptr @__cpu_features2, i32 0, i32 2)
-// CHECK-X86-NEXT: and i32 [[LOAD]], 4
+// CHECK-X86-LABEL: define dso_local i32 @v4(
+// CHECK-X86-SAME: ) #[[ATTR0]] {
+// CHECK-X86-NEXT: entry:
+// CHECK-X86-NEXT: [[TMP0:%.*]] = load i32, ptr getelementptr inbounds ([3 x i32], ptr @__cpu_features2, i32 0, i32 2), align 4
+// CHECK-X86-NEXT: [[TMP1:%.*]] = and i32 [[TMP0]], 4
+// CHECK-X86-NEXT: [[TMP2:%.*]] = icmp eq i32 [[TMP1]], 4
+// CHECK-X86-NEXT: [[TMP3:%.*]] = and i1 true, [[TMP2]]
+// CHECK-X86-NEXT: [[CONV:%.*]] = zext i1 [[TMP3]] to i32
+// CHECK-X86-NEXT: ret i32 [[CONV]]
+//
int v4() { return __builtin_cpu_supports("x86-64-v4"); }
#else
+// CHECK-PPC-LABEL: define dso_local signext i32 @test(
+// CHECK-PPC-SAME: i32 noundef signext [[A:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-PPC-NEXT: entry:
+// CHECK-PPC-NEXT: [[RETVAL:%.*]] = alloca i32, align 4
+// CHECK-PPC-NEXT: [[A_ADDR:%.*]] = alloca i32, align 4
+// CHECK-PPC-NEXT: store i32 [[A]], ptr [[A_ADDR]], align 4
+// CHECK-PPC-NEXT: [[CPU_SUPPORTS:%.*]] = call i32 @llvm.ppc.fixed.addr.ld(i32 2)
+// CHECK-PPC-NEXT: [[TMP0:%.*]] = and i32 [[CPU_SUPPORTS]], 8388608
+// CHECK-PPC-NEXT: [[TMP1:%.*]] = icmp ne i32 [[TMP0]], 0
+// CHECK-PPC-NEXT: br i1 [[TMP1]], label [[IF_THEN:%.*]], label [[IF_ELSE:%.*]]
+// CHECK-PPC: if.then:
+// CHECK-PPC-NEXT: [[TMP2:%.*]] = load i32, ptr [[A_ADDR]], align 4
+// CHECK-PPC-NEXT: store i32 [[TMP2]], ptr [[RETVAL]], align 4
+// CHECK-PPC-NEXT: br label [[RETURN:%.*]]
+// CHECK-PPC: if.else:
+// CHECK-PPC-NEXT: [[CPU_SUPPORTS1:%.*]] = call i32 @llvm.ppc.fixed.addr.ld(i32 1)
+// CHECK-PPC-NEXT: [[TMP3:%.*]] = and i32 [[CPU_SUPPORTS1]], 67108864
+// CHECK-PPC-NEXT: [[TMP4:%.*]] = icmp ne i32 [[TMP3]], 0
+// CHECK-PPC-NEXT: br i1 [[TMP4]], label [[IF_THEN2:%.*]], label [[IF_ELSE3:%.*]]
+// CHECK-PPC: if.then2:
+// CHECK-PPC-NEXT: [[TMP5:%.*]] = load i32, ptr [[A_ADDR]], align 4
+// CHECK-PPC-NEXT: [[SUB:%.*]] = sub nsw i32 [[TMP5]], 5
+// CHECK-PPC-NEXT: store i32 [[SUB]], ptr [[RETVAL]], align 4
+// CHECK-PPC-NEXT: br label [[RETURN]]
+// CHECK-PPC: if.else3:
+// CHECK-PPC-NEXT: [[CPU_IS:%.*]] = call i32 @llvm.ppc.fixed.addr.ld(i32 3)
+// CHECK-PPC-NEXT: [[TMP6:%.*]] = icmp eq i32 [[CPU_IS]], 39
+// CHECK-PPC-NEXT: br i1 [[TMP6]], label [[IF_THEN4:%.*]], label [[IF_END:%.*]]
+// CHECK-PPC: if.then4:
+// CHECK-PPC-NEXT: [[TMP7:%.*]] = load i32, ptr [[A_ADDR]], align 4
+// CHECK-PPC-NEXT: [[TMP8:%.*]] = load i32, ptr [[A_ADDR]], align 4
+// CHECK-PPC-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP7]], [[TMP8]]
+// CHECK-PPC-NEXT: store i32 [[ADD]], ptr [[RETVAL]], align 4
+// CHECK-PPC-NEXT: br label [[RETURN]]
+// CHECK-PPC: if.end:
+// CHECK-PPC-NEXT: br label [[IF_END5:%.*]]
+// CHECK-PPC: if.end5:
+// CHECK-PPC-NEXT: br label [[IF_END6:%.*]]
+// CHECK-PPC: if.end6:
+// CHECK-PPC-NEXT: [[TMP9:%.*]] = load i32, ptr [[A_ADDR]], align 4
+// CHECK-PPC-NEXT: [[ADD7:%.*]] = add nsw i32 [[TMP9]], 5
+// CHECK-PPC-NEXT: store i32 [[ADD7]], ptr [[RETVAL]], align 4
+// CHECK-PPC-NEXT: br label [[RETURN]]
+// CHECK-PPC: return:
+// CHECK-PPC-NEXT: [[TMP10:%.*]] = load i32, ptr [[RETVAL]], align 4
+// CHECK-PPC-NEXT: ret i32 [[TMP10]]
+//
int test(int a) {
-// CHECK-PPC: [[CPUSUP:%[^ ]+]] = call i32 @llvm.ppc.fixed.addr.ld(i32 2)
-// CHECK-PPC: [[AND:%[^ ]+]] = and i32 [[CPUSUP]], 8388608
-// CHECK-PPC: icmp ne i32 [[AND]], 0
-// CHECK-PPC: [[CPUSUP2:%[^ ]+]] = call i32 @llvm.ppc.fixed.addr.ld(i32 1)
-// CHECK-PPC: [[AND2:%[^ ]+]] = and i32 [[CPUSUP2]], 67108864
-// CHECK-PPC: icmp ne i32 [[AND2]], 0
-// CHECK-PPC: [[CPUID:%[^ ]+]] = call i32 @llvm.ppc.fixed.addr.ld(i32 3)
-// CHECK-PPC: icmp eq i32 [[CPUID]], 39
if (__builtin_cpu_supports("arch_3_00")) // HWCAP2
return a;
else if (__builtin_cpu_supports("mmu")) // HWCAP
diff --git a/llvm/include/llvm/IR/IntrinsicsPowerPC.td b/llvm/include/llvm/IR/IntrinsicsPowerPC.td
index d1abc37e072f70..77dc4126c697ae 100644
--- a/llvm/include/llvm/IR/IntrinsicsPowerPC.td
+++ b/llvm/include/llvm/IR/IntrinsicsPowerPC.td
@@ -216,7 +216,10 @@ let TargetPrefix = "ppc" in { // All intrinsics start with "llvm.ppc.".
[llvm_float_ty, llvm_float_ty, llvm_float_ty, llvm_vararg_ty],
[IntrNoMem]>;
// Load of a value provided by the system library at a fixed address. Used for
- // accessing things like HWCAP word provided by GLIBC.
+ // accessing things like HWCAP word provided by GLIBC. The immediate argument
+ // is not an address but a value defined in
+ // include/llvm/TargetParser/PPCTargetParser.def. Each of the values provided
+ // by Glibc is a 32-bit word.
def int_ppc_fixed_addr_ld
: DefaultAttrsIntrinsic<[llvm_i32_ty], [llvm_i32_ty],
[IntrInaccessibleMemOnly, ImmArg<ArgIndex<0>>]>;
diff --git a/llvm/include/llvm/TargetParser/PPCTargetParser.def b/llvm/include/llvm/TargetParser/PPCTargetParser.def
index e935e0112956fe..f2c44b46fa6730 100644
--- a/llvm/include/llvm/TargetParser/PPCTargetParser.def
+++ b/llvm/include/llvm/TargetParser/PPCTargetParser.def
@@ -1,8 +1,41 @@
-#ifndef PPC_FEATURE
-#define PPC_FEATURE(NAME, DESC, ENUMNAME, ENUMVAL, HWCAPN)
+//===- PPCTargetParser.def - PPC target parsing defines ---------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// This file provides defines to build up the PPC target parser's logic.
+//
+//===----------------------------------------------------------------------===//
+
+// NOTE: NO INCLUDE GUARD DESIRED!
+
+#ifdef PPC_TGT_PARSER_UNDEF_MACROS
+#undef PPC_LNX_FEATURE
+#undef PPC_LNX_CPU
+#undef PPC_FAWORD_HWCAP
+#undef PPC_FAWORD_HWCAP2
+#undef PPC_FAWORD_CPUID
+#undef PPC_HWCAP_OFFSET_LE32
+#undef PPC_HWCAP_OFFSET_LE64
+#undef PPC_HWCAP_OFFSET_BE32
+#undef PPC_HWCAP_OFFSET_BE64
+#undef PPC_HWCAP2_OFFSET_LE32
+#undef PPC_HWCAP2_OFFSET_LE64
+#undef PPC_HWCAP2_OFFSET_BE32
+#undef PPC_HWCAP2_OFFSET_BE64
+#undef PPC_CPUID_OFFSET_LE32
+#undef PPC_CPUID_OFFSET_LE64
+#undef PPC_CPUID_OFFSET_BE32
+#undef PPC_CPUID_OFFSET_BE64
+#else
+#ifndef PPC_LNX_FEATURE
+#define PPC_LNX_FEATURE(NAME, DESC, ENUMNAME, ENUMVAL, HWCAPN)
#endif
-#ifndef PPC_CPU
-#define PPC_CPU(NAME, NUM)
+#ifndef PPC_LNX_CPU
+#define PPC_LNX_CPU(NAME, NUM)
#endif
#ifndef PPC_FAWORD_HWCAP
#define PPC_FAWORD_HWCAP 1
@@ -14,67 +47,83 @@
#define PPC_FAWORD_CPUID 3
#endif
-// PPC_FEATURE(Name, Description, EnumName, BitMask, PPC_FAWORD_WORD)
-PPC_FEATURE("4xxmac","4xx CPU has a Multiply Accumulator",PPCF_4XXMAC,0x02000000,PPC_FAWORD_HWCAP)
-PPC_FEATURE("altivec","CPU has a SIMD/Vector Unit",PPCF_ALTIVEC,0x10000000,PPC_FAWORD_HWCAP)
-PPC_FEATURE("arch_2_05","CPU supports ISA 205 (eg, POWER6)",PPCF_ARCH205,0x00001000,PPC_FAWORD_HWCAP)
-PPC_FEATURE("arch_2_06","CPU supports ISA 206 (eg, POWER7)",PPCF_ARCH206,0x00000100,PPC_FAWORD_HWCAP)
-PPC_FEATURE("arch_2_07","CPU supports ISA 207 (eg, POWER8)",PPCF_ARCH207,0x80000000,PPC_FAWORD_HWCAP2)
-PPC_FEATURE("arch_3_00","CPU supports ISA 30 (eg, POWER9)",PPCF_ARCH30,0x00800000,PPC_FAWORD_HWCAP2)
-PPC_FEATURE("arch_3_1","CPU supports ISA 31 (eg, POWER10)",PPCF_ARCH31,0x00040000,PPC_FAWORD_HWCAP2)
-PPC_FEATURE("archpmu","CPU supports the set of compatible performance monitoring events",PPCF_ARCHPMU,0x00000040,PPC_FAWORD_HWCAP)
-PPC_FEATURE("booke","CPU supports the Embedded ISA category",PPCF_BOOKE,0x00008000,PPC_FAWORD_HWCAP)
-PPC_FEATURE("cellbe","CPU has a CELL broadband engine",PPCF_CELLBE,0x00010000,PPC_FAWORD_HWCAP)
-PPC_FEATURE("darn","CPU supports the darn (deliver a random number) instruction",PPCF_DARN,0x00200000,PPC_FAWORD_HWCAP2)
-PPC_FEATURE("dfp","CPU has a decimal floating point unit",PPCF_DFP,0x00000400,PPC_FAWORD_HWCAP)
-PPC_FEATURE("dscr","CPU supports the data stream control register",PPCF_DSCR,0x20000000,PPC_FAWORD_HWCAP2)
-PPC_FEATURE("ebb","CPU supports event base branching",PPCF_EBB,0x10000000,PPC_FAWORD_HWCAP2)
-PPC_FEATURE("efpdouble","CPU has a SPE double precision floating point unit",PPCF_EFPDOUBLE,0x00200000,PPC_FAWORD_HWCAP)
-PPC_FEATURE("efpsingle","CPU has a SPE single precision floating point unit",PPCF_EFPSINGLE,0x00400000,PPC_FAWORD_HWCAP)
-PPC_FEATURE("fpu","CPU has a floating point unit",PPCF_FPU,0x08000000,PPC_FAWORD_HWCAP)
-PPC_FEATURE("htm","CPU has hardware transaction memory instructions",PPCF_HTM,0x40000000,PPC_FAWORD_HWCAP2)
-PPC_FEATURE("htm-nosc","Kernel aborts hardware transactions when a syscall is made",PPCF_HTM_NOSC,0x01000000,PPC_FAWORD_HWCAP2)
-PPC_FEATURE("htm-no-suspend","CPU supports hardware transaction memory but does not support the tsuspend instruction.",PPCF_HTM_NO_SUSPEND,0x00080000,PPC_FAWORD_HWCAP2)
-PPC_FEATURE("ic_snoop","CPU supports icache snooping capabilities",PPCF_IC_SNOOP,0x00002000,PPC_FAWORD_HWCAP)
-PPC_FEATURE("ieee128","CPU supports 128-bit IEEE binary floating point instructions",PPCF_IEEE128,0x00400000,PPC_FAWORD_HWCAP2)
-PPC_FEATURE("isel","CPU supports the integer select instruction",PPCF_ISEL,0x08000000,PPC_FAWORD_HWCAP2)
-PPC_FEATURE("mma","CPU supports the matrix-multiply assist instructions",PPCF_MMA,0x00020000,PPC_FAWORD_HWCAP2)
-PPC_FEATURE("mmu","CPU has a memory management unit",PPCF_MMU,0x04000000,PPC_FAWORD_HWCAP)
-PPC_FEATURE("notb","CPU does not have a timebase (eg, 601 and 403gx)",PPCF_NOTB,0x00100000,PPC_FAWORD_HWCAP)
-PPC_FEATURE("pa6t","CPU supports the PA Semi 6T CORE ISA",PPCF_PA6T,0x00000800,PPC_FAWORD_HWCAP)
-PPC_FEATURE("power4","CPU supports ISA 200 (eg, POWER4)",PPCF_POWER4,0x00080000,PPC_FAWORD_HWCAP)
-PPC_FEATURE("power5","CPU supports ISA 202 (eg, POWER5)",PPCF_POWER5,0x00040000,PPC_FAWORD_HWCAP)
-PPC_FEATURE("power5+","CPU supports ISA 203 (eg, POWER5+)",PPCF_POWER5P,0x00020000,PPC_FAWORD_HWCAP)
-PPC_FEATURE("power6x","CPU supports ISA 205 (eg, POWER6) extended opcodes mffgpr and mftgpr.",PPCF_POWER6X,0x00000200,PPC_FAWORD_HWCAP)
-PPC_FEATURE("ppc32","CPU supports 32-bit mode execution",PPCF_PPC32,0x80000000,PPC_FAWORD_HWCAP)
-PPC_FEATURE("ppc601","CPU supports the old POWER ISA (eg, 601)",PPCF_PPC601,0x20000000,PPC_FAWORD_HWCAP)
-PPC_FEATURE("ppc64","CPU supports 64-bit mode execution",PPCF_PPC64,0x40000000,PPC_FAWORD_HWCAP)
-PPC_FEATURE("ppcle","CPU supports a little-endian mode that uses address swizzling",PPCF_PPCLE,0x00000001,PPC_FAWORD_HWCAP)
-PPC_FEATURE("scv","Kernel supports system call vectored",PPCF_SCV,0x00100000,PPC_FAWORD_HWCAP2)
-PPC_FEATURE("smt","CPU support simultaneous multi-threading",PPCF_SMT,0x00004000,PPC_FAWORD_HWCAP)
-PPC_FEATURE("spe","CPU has a signal processing extension unit",PPCF_SPE,0x00800000,PPC_FAWORD_HWCAP)
-PPC_FEATURE("tar","CPU supports the target address register",PPCF_TAR,0x04000000,PPC_FAWORD_HWCAP2)
-PPC_FEATURE("true_le","CPU supports true little-endian mode",PPCF_TRUE_LE,0x00000002,PPC_FAWORD_HWCAP)
-PPC_FEATURE("ucache","CPU has unified I/D cache",PPCF_UCACHE,0x01000000,PPC_FAWORD_HWCAP)
-PPC_FEATURE("vcrypto","CPU supports the vector cryptography instructions",PPCF_VCRYPTO,0x02000000,PPC_FAWORD_HWCAP2)
-PPC_FEATURE("vsx","CPU supports the vector-scalar extension",PPCF_VSX,0x00000080,PPC_FAWORD_HWCAP)
+// PPC_LNX_FEATURE(Name, Description, EnumName, BitMask, PPC_FAWORD_WORD)
+PPC_LNX_FEATURE("4xxmac","4xx CPU has a Multiply Accumulator",PPCF_4XXMAC,0x02000000,PPC_FAWORD_HWCAP)
+PPC_LNX_FEATURE("altivec","CPU has a SIMD/Vector Unit",PPCF_ALTIVEC,0x10000000,PPC_FAWORD_HWCAP)
+PPC_LNX_FEATURE("arch_2_05","CPU supports ISA 205 (eg, POWER6)",PPCF_ARCH205,0x00001000,PPC_FAWORD_HWCAP)
+PPC_LNX_FEATURE("arch_2_06","CPU supports ISA 206 (eg, POWER7)",PPCF_ARCH206,0x00000100,PPC_FAWORD_HWCAP)
+PPC_LNX_FEATURE("arch_2_07","CPU supports ISA 207 (eg, POWER8)",PPCF_ARCH207,0x80000000,PPC_FAWORD_HWCAP2)
+PPC_LNX_FEATURE("arch_3_00","CPU supports ISA 30 (eg, POWER9)",PPCF_ARCH30,0x00800000,PPC_FAWORD_HWCAP2)
+PPC_LNX_FEATURE("arch_3_1","CPU supports ISA 31 (eg, POWER10)",PPCF_ARCH31,0x00040000,PPC_FAWORD_HWCAP2)
+PPC_LNX_FEATURE("archpmu","CPU supports the set of compatible performance monitoring events",PPCF_ARCHPMU,0x00000040,PPC_FAWORD_HWCAP)
+PPC_LNX_FEATURE("booke","CPU supports the Embedded ISA category",PPCF_BOOKE,0x00008000,PPC_FAWORD_HWCAP)
+PPC_LNX_FEATURE("cellbe","CPU has a CELL broadband engine",PPCF_CELLBE,0x00010000,PPC_FAWORD_HWCAP)
+PPC_LNX_FEATURE("darn","CPU supports the darn (deliver a random number) instruction",PPCF_DARN,0x00200000,PPC_FAWORD_HWCAP2)
+PPC_LNX_FEATURE("dfp","CPU has a decimal floating point unit",PPCF_DFP,0x00000400,PPC_FAWORD_HWCAP)
+PPC_LNX_FEATURE("dscr","CPU supports the data stream control register",PPCF_DSCR,0x20000000,PPC_FAWORD_HWCAP2)
+PPC_LNX_FEATURE("ebb","CPU supports event base branching",PPCF_EBB,0x10000000,PPC_FAWORD_HWCAP2)
+PPC_LNX_FEATURE("efpdouble","CPU has a SPE double precision floating point unit",PPCF_EFPDOUBLE,0x00200000,PPC_FAWORD_HWCAP)
+PPC_LNX_FEATURE("efpsingle","CPU has a SPE single precision floating point unit",PPCF_EFPSINGLE,0x00400000,PPC_FAWORD_HWCAP)
+PPC_LNX_FEATURE("fpu","CPU has a floating point unit",PPCF_FPU,0x08000000,PPC_FAWORD_HWCAP)
+PPC_LNX_FEATURE("htm","CPU has hardware transaction memory instructions",PPCF_HTM,0x40000000,PPC_FAWORD_HWCAP2)
+PPC_LNX_FEATURE("htm-nosc","Kernel aborts hardware transactions when a syscall is made",PPCF_HTM_NOSC,0x01000000,PPC_FAWORD_HWCAP2)
+PPC_LNX_FEATURE("htm-no-suspend","CPU supports hardware transaction memory but does not support the tsuspend instruction.",PPCF_HTM_NO_SUSPEND,0x00080000,PPC_FAWORD_HWCAP2)
+PPC_LNX_FEATURE("ic_snoop","CPU supports icache snooping capabilities",PPCF_IC_SNOOP,0x00002000,PPC_FAWORD_HWCAP)
+PPC_LNX_FEATURE("ieee128","CPU supports 128-bit IEEE binary floating point instructions",PPCF_IEEE128,0x00400000,PPC_FAWORD_HWCAP2)
+PPC_LNX_FEATURE("isel","CPU supports the integer select instruction",PPCF_ISEL,0x08000000,PPC_FAWORD_HWCAP2)
+PPC_LNX_FEATURE("mma","CPU supports the matrix-multiply assist instructions",PPCF_MMA,0x00020000,PPC_FAWORD_HWCAP2)
+PPC_LNX_FEATURE("mmu","CPU has a memory management unit",PPCF_MMU,0x04000000,PPC_FAWORD_HWCAP)
+PPC_LNX_FEATURE("notb","CPU does not have a timebase (eg, 601 and 403gx)",PPCF_NOTB,0x00100000,PPC_FAWORD_HWCAP)
+PPC_LNX_FEATURE("pa6t","CPU supports the PA Semi 6T CORE ISA",PPCF_PA6T,0x00000800,PPC_FAWORD_HWCAP)
+PPC_LNX_FEATURE("power4","CPU supports ISA 200 (eg, POWER4)",PPCF_POWER4,0x00080000,PPC_FAWORD_HWCAP)
+PPC_LNX_FEATURE("power5","CPU supports ISA 202 (eg, POWER5)",PPCF_POWER5,0x00040000,PPC_FAWORD_HWCAP)
+PPC_LNX_FEATURE("power5+","CPU supports ISA 203 (eg, POWER5+)",PPCF_POWER5P,0x00020000,PPC_FAWORD_HWCAP)
+PPC_LNX_FEATURE("power6x","CPU supports ISA 205 (eg, POWER6) extended opcodes mffgpr and mftgpr.",PPCF_POWER6X,0x00000200,PPC_FAWORD_HWCAP)
+PPC_LNX_FEATURE("ppc32","CPU supports 32-bit mode execution",PPCF_PPC32,0x80000000,PPC_FAWORD_HWCAP)
+PPC_LNX_FEATURE("ppc601","CPU supports the old POWER ISA (eg, 601)",PPCF_PPC601,0x20000000,PPC_FAWORD_HWCAP)
+PPC_LNX_FEATURE("ppc64","CPU supports 64-bit mode execution",PPCF_PPC64,0x40000000,PPC_FAWORD_HWCAP)
+PPC_LNX_FEATURE("ppcle","CPU supports a little-endian mode that uses address swizzling",PPCF_PPCLE,0x00000001,PPC_FAWORD_HWCAP)
+PPC_LNX_FEATURE("scv","Kernel supports system call vectored",PPCF_SCV,0x00100000,PPC_FAWORD_HWCAP2)
+PPC_LNX_FEATURE("smt","CPU support simultaneous multi-threading",PPCF_SMT,0x00004000,PPC_FAWORD_HWCAP)
+PPC_LNX_FEATURE("spe","CPU has a signal processing extension unit",PPCF_SPE,0x00800000,PPC_FAWORD_HWCAP)
+PPC_LNX_FEATURE("tar","CPU supports the target address register",PPCF_TAR,0x04000000,PPC_FAWORD_HWCAP2)
+PPC_LNX_FEATURE("true_le","CPU supports true little-endian mode",PPCF_TRUE_LE,0x00000002,PPC_FAWORD_HWCAP)
+PPC_LNX_FEATURE("ucache","CPU has unified I/D cache",PPCF_UCACHE,0x01000000,PPC_FAWORD_HWCAP)
+PPC_LNX_FEATURE("vcrypto","CPU supports the vector cryptography instructions",PPCF_VCRYPTO,0x02000000,PPC_FAWORD_HWCAP2)
+PPC_LNX_FEATURE("vsx","CPU supports the vector-scalar extension",PPCF_VSX,0x00000080,PPC_FAWORD_HWCAP)
-// PPC_CPU(Name, NumericID)
-PPC_CPU("power4",32)
-PPC_CPU("ppc970",33)
-PPC_CPU("power5",34)
-PPC_CPU("power5+",35)
-PPC_CPU("power6",36)
-PPC_CPU("ppc-cell-be",37)
-PPC_CPU("power6x",38)
-PPC_CPU("power7",39)
-PPC_CPU("ppca2",40)
-PPC_CPU("ppc405",41)
-PPC_CPU("ppc440",42)
-PPC_CPU("ppc464",43)
-PPC_CPU("ppc476",44)
-PPC_CPU("power8",45)
-PPC_CPU("power9",46)
-PPC_CPU("power10",47)
-#undef PPC_FEATURE
-#undef PPC_CPU
+// PPC_LNX_CPU(Name, NumericID)
+PPC_LNX_CPU("power4",32)
+PPC_LNX_CPU("ppc970",33)
+PPC_LNX_CPU("power5",34)
+PPC_LNX_CPU("power5+",35)
+PPC_LNX_CPU("power6",36)
+PPC_LNX_CPU("ppc-cell-be",37)
+PPC_LNX_CPU("power6x",38)
+PPC_LNX_CPU("power7",39)
+PPC_LNX_CPU("ppca2",40)
+PPC_LNX_CPU("ppc405",41)
+PPC_LNX_CPU("ppc440",42)
+PPC_LNX_CPU("ppc464",43)
+PPC_LNX_CPU("ppc476",44)
+PPC_LNX_CPU("power8",45)
+PPC_LNX_CPU("power9",46)
+PPC_LNX_CPU("power10",47)
+#ifdef PPC_LNX_DEFINE_OFFSETS
+# define PPC_HWCAP_OFFSET_LE32 -0x703C
+# define PPC_HWCAP_OFFSET_LE64 -0x7064
+# define PPC_HWCAP_OFFSET_BE32 -0x7040
+# define PPC_HWCAP_OFFSET_BE64 -0x7068
+# define PPC_HWCAP2_OFFSET_LE32 -0x7040
+# define PPC_HWCAP2_OFFSET_LE64 -0x7068
+# define PPC_HWCAP2_OFFSET_BE32 -0x703C
+# define PPC_HWCAP2_OFFSET_BE64 -0x7064
+# define PPC_CPUID_OFFSET_LE32 -0x7034
+# define PPC_CPUID_OFFSET_LE64 -0x705C
+# define PPC_CPUID_OFFSET_BE32 -0x7034
+# define PPC_CPUID_OFFSET_BE64 -0x705C
+#endif
+#undef PPC_LNX_DEFINE_OFFSETS
+#undef PPC_LNX_FEATURE
+#undef PPC_LNX_CPU
+#endif // !PPC_TGT_PARSER_UNDEF_MACROS
diff --git a/llvm/lib/Target/PowerPC/PPCAsmPrinter.cpp b/llvm/lib/Target/PowerPC/PPCAsmPrinter.cpp
index 46afb882d80c5d..528267cb013291 100644
--- a/llvm/lib/Target/PowerPC/PPCAsmPrinter.cpp
+++ b/llvm/lib/Target/PowerPC/PPCAsmPrinter.cpp
@@ -1821,6 +1821,10 @@ void PPCLinuxAsmPrinter::emitEndOfAsmFile(Module &M) {
PPCTargetStreamer *TS =
static_cast<PPCTargetStreamer *>(OutStreamer->getTargetStreamer());
+ // If we are using any values provided by Glibc at fixed addresses,
+ // we need to ensure that the Glibc used at link time actually provides
+ // those values. All versions of Glibc that do will define the symbol
+ // named "__parse_hwcap_and_convert_at_platform".
if (static_cast<const PPCTargetMachine &>(TM).hasGlibcHWCAPAccess())
OutStreamer->emitSymbolValue(
GetExternalSymbolSymbol("__parse_hwcap_and_convert_at_platform"),
diff --git a/llvm/lib/Target/PowerPC/PPCInstrInfo.cpp b/llvm/lib/Target/PowerPC/PPCInstrInfo.cpp
index 34d41d05089586..38eeca0fbf821a 100644
--- a/llvm/lib/Target/PowerPC/PPCInstrInfo.cpp
+++ b/llvm/lib/Target/PowerPC/PPCInstrInfo.cpp
@@ -3101,24 +3101,34 @@ bool PPCInstrInfo::expandPostRAPseudo(MachineInstr &MI) const {
return true;
}
case PPC::PPCLdFixedAddr: {
- assert(Subtarget.isTargetLinux() &&
- "Only Linux target is expected to contain PPCLdFixedAddr");
+ assert(Subtarget.getTargetTriple().isOSGlibc() &&
+ "Only targets with Glibc expected to contain PPCLdFixedAddr");
int64_t Offset = 0;
const unsigned Reg = Subtarget.isPPC64() ? PPC::X13 : PPC::R2;
MI.setDesc(get(PPC::LWZ));
uint64_t FAType = MI.getOperand(1).getImm();
-#undef PPC_FEATURE
-#undef PPC_CPU
+#undef PPC_LNX_FEATURE
+#undef PPC_LNX_CPU
+#define PPC_LNX_DEFINE_OFFSETS
#include "llvm/TargetParser/PPCTargetParser.def"
- // The HWCAP and HWCAP2 word offsets are reversed on big endian Linux.
- if ((FAType == PPC_FAWORD_HWCAP && Subtarget.isLittleEndian()) ||
- (FAType == PPC_FAWORD_HWCAP2 && !Subtarget.isLittleEndian()))
- Offset = Subtarget.isPPC64() ? -0x7064 : -0x703C;
- else if ((FAType == PPC_FAWORD_HWCAP2 && Subtarget.isLittleEndian()) ||
- (FAType == PPC_FAWORD_HWCAP && !Subtarget.isLittleEndian()))
- Offset = Subtarget.isPPC64() ? -0x7068 : -0x7040;
- else if (FAType == PPC_FAWORD_CPUID)
- Offset = Subtarget.isPPC64() ? -0x705C : -0x7034;
+ bool IsLE = Subtarget.isLittleEndian();
+ bool Is64 = Subtarget.isPPC64();
+ if (FAType == PPC_FAWORD_HWCAP) {
+ if (IsLE)
+ Offset = Is64 ? PPC_HWCAP_OFFSET_LE64 : PPC_HWCAP_OFFSET_LE32;
+ else
+ Offset = Is64 ? PPC_HWCAP_OFFSET_BE64 : PPC_HWCAP_OFFSET_BE32;
+ } else if (FAType == PPC_FAWORD_HWCAP2) {
+ if (IsLE)
+ Offset = Is64 ? PPC_HWCAP2_OFFSET_LE64 : PPC_HWCAP2_OFFSET_LE32;
+ else
+ Offset = Is64 ? PPC_HWCAP2_OFFSET_BE64 : PPC_HWCAP2_OFFSET_BE32;
+ } else if (FAType == PPC_FAWORD_CPUID) {
+ if (IsLE)
+ Offset = Is64 ? PPC_CPUID_OFFSET_LE64 : PPC_CPUID_OFFSET_LE32;
+ else
+ Offset = Is64 ? PPC_CPUID_OFFSET_BE64 : PPC_CPUID_OFFSET_BE32;
+ }
assert(Offset && "Do not know the offset for this fixed addr load");
MI.removeOperand(1);
Subtarget.getTargetMachine().setGlibcHWCAPAccess();
@@ -3126,9 +3136,9 @@ bool PPCInstrInfo::expandPostRAPseudo(MachineInstr &MI) const {
.addImm(Offset)
.addReg(Reg);
return true;
-#undef PPC_FAWORD_HWCAP
-#undef PPC_FAWORD_HWCAP2
-#undef PPC_FAWORD_CPUID
+#define PPC_TGT_PARSER_UNDEF_MACROS
+#include "llvm/TargetParser/PPCTargetParser.def"
+#undef PPC_TGT_PARSER_UNDEF_MACROS
}
case PPC::DFLOADf32:
case PPC::DFLOADf64:
diff --git a/llvm/test/CodeGen/PowerPC/cpu-supports.ll b/llvm/test/CodeGen/PowerPC/cpu-supports.ll
index 912d67e5c555e6..82ceb15084f07b 100644
--- a/llvm/test/CodeGen/PowerPC/cpu-supports.ll
+++ b/llvm/test/CodeGen/PowerPC/cpu-supports.ll
@@ -31,6 +31,7 @@ define dso_local signext i32 @test(i32 noundef signext %a) local_unnamed_addr #0
; BE64-NEXT: addi r3, r3, -5
; BE64-NEXT: extsw r3, r3
; BE64-NEXT: blr
+; BE64: .quad __parse_hwcap_and_convert_at_platform
;
; BE32-LABEL: test:
; BE32: # %bb.0: # %entry
@@ -51,6 +52,7 @@ define dso_local signext i32 @test(i32 noundef signext %a) local_unnamed_addr #0
; BE32-NEXT: .LBB0_3: # %if.then2
; BE32-NEXT: addi r3, r3, -5
; BE32-NEXT: blr
+; BE32: .long __parse_hwcap_and_convert_at_platform
;
; LE-LABEL: test:
; LE: # %bb.0: # %entry
@@ -74,6 +76,7 @@ define dso_local signext i32 @test(i32 noundef signext %a) local_unnamed_addr #0
; LE-NEXT: addi r3, r3, -5
; LE-NEXT: extsw r3, r3
; LE-NEXT: blr
+; LE: .quad __parse_hwcap_and_convert_at_platform
entry:
%cpu_supports = tail call i32 @llvm.ppc.fixed.addr.ld(i32 2)
%0 = and i32 %cpu_supports, 8388608
>From eef7dd67ca7f0e2f87b7fc4483b8c6198364a29c Mon Sep 17 00:00:00 2001
From: Nemanja Ivanovic <nemanja at synopsys.com>
Date: Thu, 11 Jan 2024 21:56:12 +0100
Subject: [PATCH 4/5] Add the Auxiliary target checks back in for Sema.
---
clang/lib/Sema/SemaChecking.cpp | 21 +++++++++++++++------
1 file changed, 15 insertions(+), 6 deletions(-)
diff --git a/clang/lib/Sema/SemaChecking.cpp b/clang/lib/Sema/SemaChecking.cpp
index 9edf568ec95f05..299dc2ed7afc37 100644
--- a/clang/lib/Sema/SemaChecking.cpp
+++ b/clang/lib/Sema/SemaChecking.cpp
@@ -2147,16 +2147,24 @@ static bool checkFPMathBuiltinElementType(Sema &S, SourceLocation Loc,
/// This checks that the target supports the builtin and that the string
/// argument is constant and valid.
static bool SemaBuiltinCpu(Sema &S, const TargetInfo &TI, CallExpr *TheCall,
- unsigned BuiltinID) {
+ const TargetInfo *AuxTI, unsigned BuiltinID) {
assert((BuiltinID == Builtin::BI__builtin_cpu_supports ||
BuiltinID == Builtin::BI__builtin_cpu_is) &&
"Expecting __builtin_cpu_...");
bool IsCPUSupports = BuiltinID == Builtin::BI__builtin_cpu_supports;
- if (IsCPUSupports && !TI.supportsCpuSupports())
+ const TargetInfo *TheTI = &TI;
+ auto SupportsBI = [=](const TargetInfo *TInfo) {
+ return TInfo && ((IsCPUSupports && TInfo->supportsCpuSupports()) ||
+ (!IsCPUSupports && TInfo->supportsCpuIs()));
+ };
+ if (!SupportsBI(&TI) && SupportsBI(AuxTI))
+ TheTI = AuxTI;
+
+ if (IsCPUSupports && !TheTI->supportsCpuSupports())
return S.Diag(TheCall->getBeginLoc(), diag::err_builtin_target_unsupported)
<< SourceRange(TheCall->getBeginLoc(), TheCall->getEndLoc());
- if (!IsCPUSupports && !TI.supportsCpuIs())
+ if (!IsCPUSupports && !TheTI->supportsCpuIs())
return S.Diag(TheCall->getBeginLoc(), diag::err_builtin_target_unsupported)
<< SourceRange(TheCall->getBeginLoc(), TheCall->getEndLoc());
@@ -2168,10 +2176,10 @@ static bool SemaBuiltinCpu(Sema &S, const TargetInfo &TI, CallExpr *TheCall,
// Check the contents of the string.
StringRef Feature = cast<StringLiteral>(Arg)->getString();
- if (IsCPUSupports && !TI.validateCpuSupports(Feature))
+ if (IsCPUSupports && !TheTI->validateCpuSupports(Feature))
return S.Diag(TheCall->getBeginLoc(), diag::err_invalid_cpu_supports)
<< Arg->getSourceRange();
- if (!IsCPUSupports && !TI.validateCpuIs(Feature))
+ if (!IsCPUSupports && !TheTI->validateCpuIs(Feature))
return S.Diag(TheCall->getBeginLoc(), diag::err_invalid_cpu_is)
<< Arg->getSourceRange();
return false;
@@ -2207,7 +2215,8 @@ Sema::CheckBuiltinFunctionCall(FunctionDecl *FDecl, unsigned BuiltinID,
switch (BuiltinID) {
case Builtin::BI__builtin_cpu_supports:
case Builtin::BI__builtin_cpu_is:
- if (SemaBuiltinCpu(*this, Context.getTargetInfo(), TheCall, BuiltinID))
+ if (SemaBuiltinCpu(*this, Context.getTargetInfo(), TheCall,
+ Context.getAuxTargetInfo(), BuiltinID))
return ExprError();
break;
case Builtin::BI__builtin_cpu_init:
>From b1216646d375d840bbc81bff42558e9935ee3bb1 Mon Sep 17 00:00:00 2001
From: Nemanja Ivanovic <nemanja.i.ibm at gmail.com>
Date: Fri, 26 Jan 2024 07:42:02 +0100
Subject: [PATCH 5/5] Address last round of reviews and rebase
The way we define target-independent builtins has changed which
required a rebase. Also cleaned up some comments and improved
asserts as suggested in the review.
---
clang/lib/Basic/Targets/PPC.h | 2 +-
clang/lib/CodeGen/CGBuiltin.cpp | 2 ++
llvm/include/llvm/IR/IntrinsicsPowerPC.td | 4 ++--
3 files changed, 5 insertions(+), 3 deletions(-)
diff --git a/clang/lib/Basic/Targets/PPC.h b/clang/lib/Basic/Targets/PPC.h
index f700b625b79030..a91bdede53e40d 100644
--- a/clang/lib/Basic/Targets/PPC.h
+++ b/clang/lib/Basic/Targets/PPC.h
@@ -361,7 +361,7 @@ class LLVM_LIBRARY_VISIBILITY PPCTargetInfo : public TargetInfo {
}
// We support __builtin_cpu_supports/__builtin_cpu_is on targets that
- // have GLIBC since it is GLIBC that provides the HWCAP[2] in the auxv.
+ // have Glibc since it is Glibc that provides the HWCAP[2] in the auxv.
bool supportsCpuSupports() const override { return getTriple().isOSGlibc(); }
bool supportsCpuIs() const override { return getTriple().isOSGlibc(); }
bool validateCpuSupports(StringRef Feature) const override;
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index e5b15a74da392b..f3ab5ad7b08ec8 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -16552,6 +16552,7 @@ Value *CodeGenFunction::EmitPPCBuiltinExpr(unsigned BuiltinID,
#define PPC_LNX_CPU(Name, NumericID) .Case(Name, NumericID)
#include "llvm/TargetParser/PPCTargetParser.def"
.Default(-1U);
+ assert(NumCPUID < -1U && "Invalid CPU name. Missed by SemaChecking?");
Value *Op0 = llvm::ConstantInt::get(Int32Ty, PPC_FAWORD_CPUID);
llvm::Function *F = CGM.getIntrinsic(Intrinsic::ppc_fixed_addr_ld);
Value *TheCall = Builder.CreateCall(F, {Op0}, "cpu_is");
@@ -16569,6 +16570,7 @@ Value *CodeGenFunction::EmitPPCBuiltinExpr(unsigned BuiltinID,
.Case(Name, {FA_WORD, Bitmask})
#include "llvm/TargetParser/PPCTargetParser.def"
.Default({0, 0});
+ assert(BitMask && "Invalid target feature string. Missed by SemaChecking?");
Value *Op0 = llvm::ConstantInt::get(Int32Ty, FeatureWord);
llvm::Function *F = CGM.getIntrinsic(Intrinsic::ppc_fixed_addr_ld);
Value *TheCall = Builder.CreateCall(F, {Op0}, "cpu_supports");
diff --git a/llvm/include/llvm/IR/IntrinsicsPowerPC.td b/llvm/include/llvm/IR/IntrinsicsPowerPC.td
index 77dc4126c697ae..bfc2b17043bc79 100644
--- a/llvm/include/llvm/IR/IntrinsicsPowerPC.td
+++ b/llvm/include/llvm/IR/IntrinsicsPowerPC.td
@@ -216,8 +216,8 @@ let TargetPrefix = "ppc" in { // All intrinsics start with "llvm.ppc.".
[llvm_float_ty, llvm_float_ty, llvm_float_ty, llvm_vararg_ty],
[IntrNoMem]>;
// Load of a value provided by the system library at a fixed address. Used for
- // accessing things like HWCAP word provided by GLIBC. The immediate argument
- // is not an address but a value defined in
+ // accessing things like the HWCAP word provided by Glibc. The immediate
+ // argument is not an address but a value defined in
// include/llvm/TargetParser/PPCTargetParser.def. Each of the values provided
// by Glibc is a 32-bit word.
def int_ppc_fixed_addr_ld
More information about the cfe-commits
mailing list