aboutsummaryrefslogtreecommitdiff
path: root/clang/lib/CodeGen
diff options
context:
space:
mode:
Diffstat (limited to 'clang/lib/CodeGen')
-rw-r--r--clang/lib/CodeGen/CGBuiltin.cpp12
-rw-r--r--clang/lib/CodeGen/CGStmt.cpp69
-rw-r--r--clang/lib/CodeGen/CodeGenModule.cpp5
-rw-r--r--clang/lib/CodeGen/CodeGenPGO.cpp9
-rw-r--r--clang/lib/CodeGen/TargetBuiltins/ARM.cpp32
5 files changed, 78 insertions, 49 deletions
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index bbcee34..0a2ea41 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -1211,14 +1211,10 @@ llvm::Value *CodeGenFunction::emitCountedByPointerSize(
getContext().getTypeSizeInChars(ElementTy->getPointeeType());
if (ElementSize.isZero()) {
- // This might be a __sized_by on a 'void *', which counts bytes, not
- // elements.
- auto *CAT = ElementTy->getAs<CountAttributedType>();
- if (!CAT || (CAT->getKind() != CountAttributedType::SizedBy &&
- CAT->getKind() != CountAttributedType::SizedByOrNull))
- // Okay, not sure what it is now.
- // FIXME: Should this be an assert?
- return std::optional<CharUnits>();
+ // This might be a __sized_by (or __counted_by) on a
+ // 'void *', which counts bytes, not elements.
+ [[maybe_unused]] auto *CAT = ElementTy->getAs<CountAttributedType>();
+ assert(CAT && "must have an CountAttributedType");
ElementSize = CharUnits::One();
}
diff --git a/clang/lib/CodeGen/CGStmt.cpp b/clang/lib/CodeGen/CGStmt.cpp
index fdc1a11..36be329 100644
--- a/clang/lib/CodeGen/CGStmt.cpp
+++ b/clang/lib/CodeGen/CGStmt.cpp
@@ -582,48 +582,45 @@ CodeGenFunction::EmitCompoundStmtWithoutScope(const CompoundStmt &S,
bool GetLast,
AggValueSlot AggSlot) {
- const Stmt *ExprResult = S.getStmtExprResult();
- assert((!GetLast || (GetLast && ExprResult)) &&
- "If GetLast is true then the CompoundStmt must have a StmtExprResult");
+ for (CompoundStmt::const_body_iterator I = S.body_begin(),
+ E = S.body_end() - GetLast;
+ I != E; ++I)
+ EmitStmt(*I);
Address RetAlloca = Address::invalid();
-
- for (auto *CurStmt : S.body()) {
- if (GetLast && ExprResult == CurStmt) {
- // We have to special case labels here. They are statements, but when put
- // at the end of a statement expression, they yield the value of their
- // subexpression. Handle this by walking through all labels we encounter,
- // emitting them before we evaluate the subexpr.
- // Similar issues arise for attributed statements.
- while (!isa<Expr>(ExprResult)) {
- if (const auto *LS = dyn_cast<LabelStmt>(ExprResult)) {
- EmitLabel(LS->getDecl());
- ExprResult = LS->getSubStmt();
- } else if (const auto *AS = dyn_cast<AttributedStmt>(ExprResult)) {
- // FIXME: Update this if we ever have attributes that affect the
- // semantics of an expression.
- ExprResult = AS->getSubStmt();
- } else {
- llvm_unreachable("unknown value statement");
- }
+ if (GetLast) {
+ // We have to special case labels here. They are statements, but when put
+ // at the end of a statement expression, they yield the value of their
+ // subexpression. Handle this by walking through all labels we encounter,
+ // emitting them before we evaluate the subexpr.
+ // Similar issues arise for attributed statements.
+ const Stmt *LastStmt = S.body_back();
+ while (!isa<Expr>(LastStmt)) {
+ if (const auto *LS = dyn_cast<LabelStmt>(LastStmt)) {
+ EmitLabel(LS->getDecl());
+ LastStmt = LS->getSubStmt();
+ } else if (const auto *AS = dyn_cast<AttributedStmt>(LastStmt)) {
+ // FIXME: Update this if we ever have attributes that affect the
+ // semantics of an expression.
+ LastStmt = AS->getSubStmt();
+ } else {
+ llvm_unreachable("unknown value statement");
}
+ }
- EnsureInsertPoint();
+ EnsureInsertPoint();
- const Expr *E = cast<Expr>(ExprResult);
- QualType ExprTy = E->getType();
- if (hasAggregateEvaluationKind(ExprTy)) {
- EmitAggExpr(E, AggSlot);
- } else {
- // We can't return an RValue here because there might be cleanups at
- // the end of the StmtExpr. Because of that, we have to emit the result
- // here into a temporary alloca.
- RetAlloca = CreateMemTemp(ExprTy);
- EmitAnyExprToMem(E, RetAlloca, Qualifiers(),
- /*IsInit*/ false);
- }
+ const Expr *E = cast<Expr>(LastStmt);
+ QualType ExprTy = E->getType();
+ if (hasAggregateEvaluationKind(ExprTy)) {
+ EmitAggExpr(E, AggSlot);
} else {
- EmitStmt(CurStmt);
+ // We can't return an RValue here because there might be cleanups at
+ // the end of the StmtExpr. Because of that, we have to emit the result
+ // here into a temporary alloca.
+ RetAlloca = CreateMemTemp(ExprTy);
+ EmitAnyExprToMem(E, RetAlloca, Qualifiers(),
+ /*IsInit*/ false);
}
}
diff --git a/clang/lib/CodeGen/CodeGenModule.cpp b/clang/lib/CodeGen/CodeGenModule.cpp
index 0fea57b..98d59b7 100644
--- a/clang/lib/CodeGen/CodeGenModule.cpp
+++ b/clang/lib/CodeGen/CodeGenModule.cpp
@@ -2368,9 +2368,8 @@ static QualType GeneralizeTransparentUnion(QualType Ty) {
const RecordDecl *UD = UT->getDecl()->getDefinitionOrSelf();
if (!UD->hasAttr<TransparentUnionAttr>())
return Ty;
- for (const auto *it : UD->fields()) {
- return it->getType();
- }
+ if (!UD->fields().empty())
+ return UD->fields().begin()->getType();
return Ty;
}
diff --git a/clang/lib/CodeGen/CodeGenPGO.cpp b/clang/lib/CodeGen/CodeGenPGO.cpp
index 8f09564..06d7380 100644
--- a/clang/lib/CodeGen/CodeGenPGO.cpp
+++ b/clang/lib/CodeGen/CodeGenPGO.cpp
@@ -58,9 +58,10 @@ enum PGOHashVersion : unsigned {
PGO_HASH_V1,
PGO_HASH_V2,
PGO_HASH_V3,
+ PGO_HASH_V4,
// Keep this set to the latest hash version.
- PGO_HASH_LATEST = PGO_HASH_V3
+ PGO_HASH_LATEST = PGO_HASH_V4
};
namespace {
@@ -152,7 +153,9 @@ static PGOHashVersion getPGOHashVersion(llvm::IndexedInstrProfReader *PGOReader,
return PGO_HASH_V1;
if (PGOReader->getVersion() <= 5)
return PGO_HASH_V2;
- return PGO_HASH_V3;
+ if (PGOReader->getVersion() <= 12)
+ return PGO_HASH_V3;
+ return PGO_HASH_V4;
}
/// A RecursiveASTVisitor that fills a map of statements to PGO counters.
@@ -1099,6 +1102,8 @@ void CodeGenPGO::mapRegionCounters(const Decl *D) {
assert(Walker.NextCounter > 0 && "no entry counter mapped for decl");
NumRegionCounters = Walker.NextCounter;
FunctionHash = Walker.Hash.finalize();
+ if (HashVersion >= PGO_HASH_V4)
+ FunctionHash &= llvm::NamedInstrProfRecord::FUNC_HASH_MASK;
}
bool CodeGenPGO::skipRegionMappingForDecl(const Decl *D) {
diff --git a/clang/lib/CodeGen/TargetBuiltins/ARM.cpp b/clang/lib/CodeGen/TargetBuiltins/ARM.cpp
index 15fa78d..d4b0b81 100644
--- a/clang/lib/CodeGen/TargetBuiltins/ARM.cpp
+++ b/clang/lib/CodeGen/TargetBuiltins/ARM.cpp
@@ -590,6 +590,7 @@ struct ARMVectorIntrinsicInfo {
Intrinsic::LLVMIntrinsic, Intrinsic::AltLLVMIntrinsic, \
TypeModifier }
+// clang-format off
static const ARMVectorIntrinsicInfo ARMSIMDIntrinsicMap [] = {
NEONMAP1(__a32_vcvt_bf16_f32, arm_neon_vcvtfp2bf, 0),
NEONMAP0(splat_lane_v),
@@ -1217,35 +1218,55 @@ static const ARMVectorIntrinsicInfo AArch64SISDIntrinsicMap[] = {
NEONMAP1(vcales_f32, aarch64_neon_facge, AddRetType | Add1ArgType),
NEONMAP1(vcaltd_f64, aarch64_neon_facgt, AddRetType | Add1ArgType),
NEONMAP1(vcalts_f32, aarch64_neon_facgt, AddRetType | Add1ArgType),
+ NEONMAP1(vcvtad_s32_f64, aarch64_neon_fcvtas, AddRetType | Add1ArgType),
NEONMAP1(vcvtad_s64_f64, aarch64_neon_fcvtas, AddRetType | Add1ArgType),
+ NEONMAP1(vcvtad_u32_f64, aarch64_neon_fcvtau, AddRetType | Add1ArgType),
NEONMAP1(vcvtad_u64_f64, aarch64_neon_fcvtau, AddRetType | Add1ArgType),
NEONMAP1(vcvtas_s32_f32, aarch64_neon_fcvtas, AddRetType | Add1ArgType),
+ NEONMAP1(vcvtas_s64_f32, aarch64_neon_fcvtas, AddRetType | Add1ArgType),
NEONMAP1(vcvtas_u32_f32, aarch64_neon_fcvtau, AddRetType | Add1ArgType),
+ NEONMAP1(vcvtas_u64_f32, aarch64_neon_fcvtau, AddRetType | Add1ArgType),
NEONMAP1(vcvtd_n_f64_s64, aarch64_neon_vcvtfxs2fp, AddRetType | Add1ArgType),
NEONMAP1(vcvtd_n_f64_u64, aarch64_neon_vcvtfxu2fp, AddRetType | Add1ArgType),
NEONMAP1(vcvtd_n_s64_f64, aarch64_neon_vcvtfp2fxs, AddRetType | Add1ArgType),
NEONMAP1(vcvtd_n_u64_f64, aarch64_neon_vcvtfp2fxu, AddRetType | Add1ArgType),
+ NEONMAP1(vcvtd_s32_f64, aarch64_neon_fcvtzs, AddRetType | Add1ArgType),
NEONMAP1(vcvtd_s64_f64, aarch64_neon_fcvtzs, AddRetType | Add1ArgType),
+ NEONMAP1(vcvtd_u32_f64, aarch64_neon_fcvtzu, AddRetType | Add1ArgType),
NEONMAP1(vcvtd_u64_f64, aarch64_neon_fcvtzu, AddRetType | Add1ArgType),
NEONMAP0(vcvth_bf16_f32),
+ NEONMAP1(vcvtmd_s32_f64, aarch64_neon_fcvtms, AddRetType | Add1ArgType),
NEONMAP1(vcvtmd_s64_f64, aarch64_neon_fcvtms, AddRetType | Add1ArgType),
+ NEONMAP1(vcvtmd_u32_f64, aarch64_neon_fcvtmu, AddRetType | Add1ArgType),
NEONMAP1(vcvtmd_u64_f64, aarch64_neon_fcvtmu, AddRetType | Add1ArgType),
NEONMAP1(vcvtms_s32_f32, aarch64_neon_fcvtms, AddRetType | Add1ArgType),
+ NEONMAP1(vcvtms_s64_f32, aarch64_neon_fcvtms, AddRetType | Add1ArgType),
NEONMAP1(vcvtms_u32_f32, aarch64_neon_fcvtmu, AddRetType | Add1ArgType),
+ NEONMAP1(vcvtms_u64_f32, aarch64_neon_fcvtmu, AddRetType | Add1ArgType),
+ NEONMAP1(vcvtnd_s32_f64, aarch64_neon_fcvtns, AddRetType | Add1ArgType),
NEONMAP1(vcvtnd_s64_f64, aarch64_neon_fcvtns, AddRetType | Add1ArgType),
+ NEONMAP1(vcvtnd_u32_f64, aarch64_neon_fcvtnu, AddRetType | Add1ArgType),
NEONMAP1(vcvtnd_u64_f64, aarch64_neon_fcvtnu, AddRetType | Add1ArgType),
NEONMAP1(vcvtns_s32_f32, aarch64_neon_fcvtns, AddRetType | Add1ArgType),
+ NEONMAP1(vcvtns_s64_f32, aarch64_neon_fcvtns, AddRetType | Add1ArgType),
NEONMAP1(vcvtns_u32_f32, aarch64_neon_fcvtnu, AddRetType | Add1ArgType),
+ NEONMAP1(vcvtns_u64_f32, aarch64_neon_fcvtnu, AddRetType | Add1ArgType),
+ NEONMAP1(vcvtpd_s32_f64, aarch64_neon_fcvtps, AddRetType | Add1ArgType),
NEONMAP1(vcvtpd_s64_f64, aarch64_neon_fcvtps, AddRetType | Add1ArgType),
+ NEONMAP1(vcvtpd_u32_f64, aarch64_neon_fcvtpu, AddRetType | Add1ArgType),
NEONMAP1(vcvtpd_u64_f64, aarch64_neon_fcvtpu, AddRetType | Add1ArgType),
NEONMAP1(vcvtps_s32_f32, aarch64_neon_fcvtps, AddRetType | Add1ArgType),
+ NEONMAP1(vcvtps_s64_f32, aarch64_neon_fcvtps, AddRetType | Add1ArgType),
NEONMAP1(vcvtps_u32_f32, aarch64_neon_fcvtpu, AddRetType | Add1ArgType),
+ NEONMAP1(vcvtps_u64_f32, aarch64_neon_fcvtpu, AddRetType | Add1ArgType),
NEONMAP1(vcvts_n_f32_s32, aarch64_neon_vcvtfxs2fp, AddRetType | Add1ArgType),
NEONMAP1(vcvts_n_f32_u32, aarch64_neon_vcvtfxu2fp, AddRetType | Add1ArgType),
NEONMAP1(vcvts_n_s32_f32, aarch64_neon_vcvtfp2fxs, AddRetType | Add1ArgType),
NEONMAP1(vcvts_n_u32_f32, aarch64_neon_vcvtfp2fxu, AddRetType | Add1ArgType),
NEONMAP1(vcvts_s32_f32, aarch64_neon_fcvtzs, AddRetType | Add1ArgType),
+ NEONMAP1(vcvts_s64_f32, aarch64_neon_fcvtzs, AddRetType | Add1ArgType),
NEONMAP1(vcvts_u32_f32, aarch64_neon_fcvtzu, AddRetType | Add1ArgType),
+ NEONMAP1(vcvts_u64_f32, aarch64_neon_fcvtzu, AddRetType | Add1ArgType),
NEONMAP1(vcvtxd_f32_f64, aarch64_sisd_fcvtxn, 0),
NEONMAP1(vmaxnmv_f32, aarch64_neon_fmaxnmv, AddRetType | Add1ArgType),
NEONMAP1(vmaxnmvq_f32, aarch64_neon_fmaxnmv, AddRetType | Add1ArgType),
@@ -1446,6 +1467,7 @@ static const ARMVectorIntrinsicInfo AArch64SISDIntrinsicMap[] = {
NEONMAP1(vrsqrteh_f16, aarch64_neon_frsqrte, Add1ArgType),
NEONMAP1(vrsqrtsh_f16, aarch64_neon_frsqrts, Add1ArgType),
};
+// clang-format on
// Some intrinsics are equivalent for codegen.
static const std::pair<unsigned, unsigned> NEONEquivalentIntrinsicMap[] = {
@@ -7624,6 +7646,16 @@ Value *CodeGenFunction::EmitAArch64BuiltinExpr(unsigned BuiltinID,
Int = Intrinsic::aarch64_neon_vluti4q_laneq_x2;
return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vluti4q_laneq_x2");
}
+ case NEON::BI__builtin_neon_vmmlaq_f16_mf8_fpm:
+ return EmitFP8NeonCall(Intrinsic::aarch64_neon_fmmla,
+ {llvm::FixedVectorType::get(HalfTy, 8),
+ llvm::FixedVectorType::get(Int8Ty, 16)},
+ Ops, E, "fmmla");
+ case NEON::BI__builtin_neon_vmmlaq_f32_mf8_fpm:
+ return EmitFP8NeonCall(Intrinsic::aarch64_neon_fmmla,
+ {llvm::FixedVectorType::get(FloatTy, 4),
+ llvm::FixedVectorType::get(Int8Ty, 16)},
+ Ops, E, "fmmla");
case NEON::BI__builtin_neon_vcvt1_low_bf16_mf8_fpm:
ExtractLow = true;
[[fallthrough]];
generated by cgit v1.1 at 2025-11-08 10:14:47 +0000