[clang] [clang-tools-extra] [lld] [lldb] [llvm] Rework the `Option` library to reduce dynamic relocations (PR #119198)
@@ -53,10 +53,8 @@ class OptTable {
public:
/// Entry for a single option instance in the option data table.
struct Info {
-/// A null terminated array of prefix strings to apply to name while
-/// matching.
-ArrayRef Prefixes;
-StringLiteral PrefixedName;
+unsigned PrefixesOffset;
chandlerc wrote:
FWIW, I put together a simple abstraction here:
https://github.com/llvm/llvm-project/pull/119488
Happy for suggestions on a stronger one, a different API, etc.
My preference would be to land independently and then connect. But if you'd
like to see one or the other land first and roll in, let me know.
https://github.com/llvm/llvm-project/pull/119198
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] cc1a2ea - [AArch64] Implement FP8 SVE intrinsics for widening conversions (#118123)
Author: Momchil Velikov
Date: 2024-12-10T13:32:05Z
New Revision: cc1a2ea61e3f8e790125b10d9ec4e7d179156ddf
URL:
https://github.com/llvm/llvm-project/commit/cc1a2ea61e3f8e790125b10d9ec4e7d179156ddf
DIFF:
https://github.com/llvm/llvm-project/commit/cc1a2ea61e3f8e790125b10d9ec4e7d179156ddf.diff
LOG: [AArch64] Implement FP8 SVE intrinsics for widening conversions (#118123)
This patch adds the following intrinsics:
* 8-bit floating-point convert to half-precision and BFloat16.
// Variants are also available for: _bf16
svfloat16_t svcvt1_f16[_mf8]_fpm(svmfloat8_t zn, fpm_t fpm);
svfloat16_t svcvt2_f16[_mf8]_fpm(svmfloat8_t zn, fpm_t fpm);
* 8-bit floating-point convert to half-precision and BFloat16 (top).
// Variants are also available for: _bf16
svfloat16_t svcvtlt1_f16[_mf8]_fpm(svmfloat8_t zn, fpm_t fpm);
svfloat16_t svcvtlt2_f16[_mf8]_fpm(svmfloat8_t zn, fpm_t fpm);
Added:
clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sve2_fp8_cvt.c
clang/test/Sema/aarch64-sve2-intrinsics/acle_sve2_fp8.c
llvm/test/CodeGen/AArch64/fp8-sve-cvt-cvtlt.ll
Modified:
clang/include/clang/Basic/arm_sve.td
llvm/include/llvm/IR/IntrinsicsAArch64.td
llvm/lib/Target/AArch64/AArch64SVEInstrInfo.td
llvm/lib/Target/AArch64/SVEInstrFormats.td
Removed:
diff --git a/clang/include/clang/Basic/arm_sve.td
b/clang/include/clang/Basic/arm_sve.td
index 9b8a8546b072c0..7b8ecf29a9de6e 100644
--- a/clang/include/clang/Basic/arm_sve.td
+++ b/clang/include/clang/Basic/arm_sve.td
@@ -2430,12 +2430,12 @@ let SVETargetGuard = InvalidMode, SMETargetGuard =
"sme2,fp8" in {
def FSCALE_X4 : Inst<"svscale[_{d}_x4]", "444.x", "fhd", MergeNone,
"aarch64_sme_fp8_scale_x4", [IsStreaming],[]>;
// Convert from FP8 to half-precision/BFloat16 multi-vector
- def SVF1CVT : Inst<"svcvt1_{d}[_mf8]_x2_fpm", "2~>", "bh", MergeNone,
"aarch64_sve_fp8_cvt1_x2", [IsStreaming, SetsFPMR], []>;
- def SVF2CVT : Inst<"svcvt2_{d}[_mf8]_x2_fpm", "2~>", "bh", MergeNone,
"aarch64_sve_fp8_cvt2_x2", [IsStreaming, SetsFPMR], []>;
+ def SVF1CVT_X2 : Inst<"svcvt1_{d}[_mf8]_x2_fpm", "2~>", "bh", MergeNone,
"aarch64_sve_fp8_cvt1_x2", [IsStreaming, SetsFPMR], []>;
+ def SVF2CVT_X2 : Inst<"svcvt2_{d}[_mf8]_x2_fpm", "2~>", "bh", MergeNone,
"aarch64_sve_fp8_cvt2_x2", [IsStreaming, SetsFPMR], []>;
// Convert from FP8 to deinterleaved half-precision/BFloat16 multi-vector
- def SVF1CVTL : Inst<"svcvtl1_{d}[_mf8]_x2_fpm", "2~>", "bh", MergeNone,
"aarch64_sve_fp8_cvtl1_x2", [IsStreaming, SetsFPMR], []>;
- def SVF2CVTL : Inst<"svcvtl2_{d}[_mf8]_x2_fpm", "2~>", "bh", MergeNone,
"aarch64_sve_fp8_cvtl2_x2", [IsStreaming, SetsFPMR], []>;
+ def SVF1CVTL_X2 : Inst<"svcvtl1_{d}[_mf8]_x2_fpm", "2~>", "bh", MergeNone,
"aarch64_sve_fp8_cvtl1_x2", [IsStreaming, SetsFPMR], []>;
+ def SVF2CVTL_X2 : Inst<"svcvtl2_{d}[_mf8]_x2_fpm", "2~>", "bh", MergeNone,
"aarch64_sve_fp8_cvtl2_x2", [IsStreaming, SetsFPMR], []>;
}
let SVETargetGuard = "sve2p1", SMETargetGuard = "sme2" in {
@@ -2451,3 +2451,15 @@ let SVETargetGuard = "sve2,faminmax", SMETargetGuard =
"sme2,faminmax" in {
defm SVAMIN : SInstZPZZ<"svamin", "hfd", "aarch64_sve_famin",
"aarch64_sve_famin_u">;
defm SVAMAX : SInstZPZZ<"svamax", "hfd", "aarch64_sve_famax",
"aarch64_sve_famax_u">;
}
+
+let SVETargetGuard = "sve2,fp8", SMETargetGuard = "sme2,fp8" in {
+ // SVE FP8 widening conversions
+
+ // 8-bit floating-point convert to BFloat16/Float16
+ def SVF1CVT : SInst<"svcvt1_{d}[_mf8]_fpm", "d~>", "bh", MergeNone,
"aarch64_sve_fp8_cvt1", [VerifyRuntimeMode, SetsFPMR]>;
+ def SVF2CVT : SInst<"svcvt2_{d}[_mf8]_fpm", "d~>", "bh", MergeNone,
"aarch64_sve_fp8_cvt2", [VerifyRuntimeMode, SetsFPMR]>;
+
+ // 8-bit floating-point convert to BFloat16/Float16 (top)
+ def SVF1CVTLT : SInst<"svcvtlt1_{d}[_mf8]_fpm", "d~>", "bh", MergeNone,
"aarch64_sve_fp8_cvtlt1", [VerifyRuntimeMode, SetsFPMR]>;
+ def SVF2CVTLT : SInst<"svcvtlt2_{d}[_mf8]_fpm", "d~>", "bh", MergeNone,
"aarch64_sve_fp8_cvtlt2", [VerifyRuntimeMode, SetsFPMR]>;
+}
diff --git a/clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sve2_fp8_cvt.c
b/clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sve2_fp8_cvt.c
new file mode 100644
index 00..c026b8aa216f32
--- /dev/null
+++ b/clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sve2_fp8_cvt.c
@@ -0,0 +1,173 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
UTC_ARGS: --version 5
+// RUN: %clang_cc1-triple aarch64-none-linux-gnu -target-feature +sve
-target-feature +sve2 -target-feature +fp8 -disable-O0-optnone -Werror -Wall
-emit-llvm -o - %s | opt -S -p mem2reg,instcombine,tailcallelim | FileCheck %s
+// RUN: %clang_cc1 -x c++ -triple aarch64-none-linux-gnu -target-feature +sme
-target-feature +sme2 -target-feature +fp8 -disable-O0-optnone -Werror -Wall
-emit-llvm -o - %s | opt -S -p mem2reg,instcombine,tailcallelim |
[clang] Switch builtin strings to use string tables (PR #118734)
zmodem wrote: > I've put it up on Dropbox I looked at Builtins.cpp.obj and compared to the one in my build dir. They're very similar, in particular the string table is there in both files, and looks identical. I don't really have any more ideas for debugging this. https://github.com/llvm/llvm-project/pull/118734 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [clang][CodeGen] `sret` args should always point to the `alloca` AS, so use that (PR #114062)
@@ -225,7 +225,9 @@ ABIArgInfo AMDGPUABIInfo::classifyArgumentType(QualType Ty, bool Variadic, // Records with non-trivial destructors/copy-constructors should not be // passed by value. if (auto RAA = getRecordArgABI(Ty, getCXXABI())) - return getNaturalAlignIndirect(Ty, RAA == CGCXXABI::RAA_DirectInMemory); + return getNaturalAlignIndirect( + Ty, getContext().getTargetAddressSpace(LangAS::Default), + RAA == CGCXXABI::RAA_DirectInMemory); AlexVlx wrote: This is not intended to alter current behaviour, as without this patch non aliased indirect args wouldn't carry an AS, so it's just dealing with the interface extension for `getIndirect` / `getNaturalAlignIndirect`. https://github.com/llvm/llvm-project/pull/114062 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [clang][bytecode] Check vector element types for eligibility (PR #119385)
https://github.com/tbaederr created
https://github.com/llvm/llvm-project/pull/119385
Like we do in ExprConstant.cpp.
>From 003dd9075216766af2b717e5cfd5f38fba15d6ca Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Timm=20B=C3=A4der?=
Date: Tue, 10 Dec 2024 15:12:28 +0100
Subject: [PATCH] [clang][bytecode] Check vector element types for eligibility
Like we do in ExprConstant.cpp.
---
.../lib/AST/ByteCode/InterpBuiltinBitCast.cpp | 29 +++
clang/test/AST/ByteCode/builtin-bit-cast.cpp | 5
2 files changed, 34 insertions(+)
diff --git a/clang/lib/AST/ByteCode/InterpBuiltinBitCast.cpp
b/clang/lib/AST/ByteCode/InterpBuiltinBitCast.cpp
index c9cd113287557b..7ba0faff252530 100644
--- a/clang/lib/AST/ByteCode/InterpBuiltinBitCast.cpp
+++ b/clang/lib/AST/ByteCode/InterpBuiltinBitCast.cpp
@@ -222,6 +222,35 @@ static bool CheckBitcastType(InterpState &S, CodePtr OpPC,
QualType T,
IsToType))
return false;
+ if (const auto *VT = T->getAs()) {
+const ASTContext &ASTCtx = S.getASTContext();
+QualType EltTy = VT->getElementType();
+unsigned NElts = VT->getNumElements();
+unsigned EltSize =
+VT->isExtVectorBoolType() ? 1 : ASTCtx.getTypeSize(EltTy);
+
+if ((NElts * EltSize) % ASTCtx.getCharWidth() != 0) {
+ // The vector's size in bits is not a multiple of the target's byte size,
+ // so its layout is unspecified. For now, we'll simply treat these cases
+ // as unsupported (this should only be possible with OpenCL bool vectors
+ // whose element count isn't a multiple of the byte size).
+ const Expr *E = S.Current->getExpr(OpPC);
+ S.FFDiag(E, diag::note_constexpr_bit_cast_invalid_vector)
+ << QualType(VT, 0) << EltSize << NElts << ASTCtx.getCharWidth();
+ return false;
+}
+
+if (EltTy->isRealFloatingType() &&
+&ASTCtx.getFloatTypeSemantics(EltTy) == &APFloat::x87DoubleExtended())
{
+ // The layout for x86_fp80 vectors seems to be handled very
inconsistently
+ // by both clang and LLVM, so for now we won't allow bit_casts involving
+ // it in a constexpr context.
+ const Expr *E = S.Current->getExpr(OpPC);
+ S.FFDiag(E, diag::note_constexpr_bit_cast_unsupported_type) << EltTy;
+ return false;
+}
+ }
+
return true;
}
diff --git a/clang/test/AST/ByteCode/builtin-bit-cast.cpp
b/clang/test/AST/ByteCode/builtin-bit-cast.cpp
index e99ab3904c339c..8a5bef635b8fde 100644
--- a/clang/test/AST/ByteCode/builtin-bit-cast.cpp
+++ b/clang/test/AST/ByteCode/builtin-bit-cast.cpp
@@ -502,3 +502,8 @@ namespace OversizedBitField {
// ref-note {{constexpr bit_cast involving bit-field is not yet supported}}
#endif
}
+
+typedef bool bool9 __attribute__((ext_vector_type(9)));
+// both-error@+2 {{constexpr variable 'bad_bool9_to_short' must be initialized
by a constant expression}}
+// both-note@+1 {{bit_cast involving type 'bool
__attribute__((ext_vector_type(9)))' (vector of 9 'bool' values) is not allowed
in a constant expression; element size 1 * element count 9 is not a multiple of
the byte size 8}}
+constexpr unsigned short bad_bool9_to_short = __builtin_bit_cast(unsigned
short, bool9{1,1,0,1,0,1,0,1,0});
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [analyzer] Don't assume third iteration in loops (PR #119388)
NagyDonat wrote: The difference between this PR and my earlier PR https://github.com/llvm/llvm-project/pull/109804 "Suppress out of bounds reports after weak loop assumptions" is: | This PR | Earlier PR | | :---: | :---: | | affects all checkers | only affects ArrayBoundV2 | | doesn't traverse unjustified paths | only suppresses reports after traversing the paths | | only discards assuming third or later iteration | also suppresses reports after assuming zero iterations | Slightly offtopic remark: I'm planning to create a separate commit to filter out branches where the analyzer assumes that 0 iterations happen in a loop (with an opaque condition). I'll probably put that logic behind an analyzer option (e.g. `assume-min-iterations` where I suggest 1 as the sane default value), because there is probably user demand for a pedantic analysis mode that aggressively assumes that loops can have 0 iterations. (This PR doesn't introduce an analyzer option, because I don't think that "I want the analyzer to assume that 3+ iterations can happen" is something that the users would want to use.) https://github.com/llvm/llvm-project/pull/119388 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [libcxx] [libcxx] Support for using timespec_get (PR #117362)
ldionne wrote: Gentle ping https://github.com/llvm/llvm-project/pull/117362 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [clang][bytecode][NFC] Avoid calling ASTContext::getTypeSize() (PR #119392)
llvmbot wrote:
@llvm/pr-subscribers-clang
Author: Timm Baeder (tbaederr)
Changes
In the case of primitive arrays, we can do this once and reuse the value for
all elements.
---
Full diff: https://github.com/llvm/llvm-project/pull/119392.diff
1 Files Affected:
- (modified) clang/lib/AST/ByteCode/InterpBuiltinBitCast.cpp (+20-20)
``diff
diff --git a/clang/lib/AST/ByteCode/InterpBuiltinBitCast.cpp
b/clang/lib/AST/ByteCode/InterpBuiltinBitCast.cpp
index c9cd113287557b..3a056acff14a7f 100644
--- a/clang/lib/AST/ByteCode/InterpBuiltinBitCast.cpp
+++ b/clang/lib/AST/ByteCode/InterpBuiltinBitCast.cpp
@@ -33,8 +33,9 @@ using namespace clang::interp;
//bytes to/from the buffer.
/// Used to iterate over pointer fields.
-using DataFunc = llvm::function_ref;
+using DataFunc =
+llvm::function_ref;
#define BITCAST_TYPE_SWITCH(Expr, B)
\
do {
\
@@ -85,21 +86,25 @@ static bool enumerateData(const Pointer &P, const Context
&Ctx, Bits Offset,
assert(FieldDesc);
// Primitives.
- if (FieldDesc->isPrimitive())
-return F(P, FieldDesc->getPrimType(), Offset, /*PackedBools=*/false);
+ if (FieldDesc->isPrimitive()) {
+Bits FullBitWidth =
+Bits(Ctx.getASTContext().getTypeSize(FieldDesc->getType()));
+return F(P, FieldDesc->getPrimType(), Offset, FullBitWidth,
+ /*PackedBools=*/false);
+ }
// Primitive arrays.
if (FieldDesc->isPrimitiveArray()) {
QualType ElemType = FieldDesc->getElemQualType();
-size_t ElemSizeInBits = Ctx.getASTContext().getTypeSize(ElemType);
+Bits ElemSize = Bits(Ctx.getASTContext().getTypeSize(ElemType));
PrimType ElemT = *Ctx.classify(ElemType);
// Special case, since the bools here are packed.
bool PackedBools = FieldDesc->getType()->isExtVectorBoolType();
unsigned NumElems = FieldDesc->getNumElems();
bool Ok = true;
for (unsigned I = P.getIndex(); I != NumElems; ++I) {
- Ok = Ok && F(P.atIndex(I), ElemT, Offset, PackedBools);
- Offset += PackedBools ? 1 : ElemSizeInBits;
+ Ok = Ok && F(P.atIndex(I), ElemT, Offset, ElemSize, PackedBools);
+ Offset += PackedBools ? Bits(1) : ElemSize;
if (Offset >= BitsToRead)
break;
}
@@ -109,10 +114,10 @@ static bool enumerateData(const Pointer &P, const Context
&Ctx, Bits Offset,
// Composite arrays.
if (FieldDesc->isCompositeArray()) {
QualType ElemType = FieldDesc->getElemQualType();
-size_t ElemSizeInBits = Ctx.getASTContext().getTypeSize(ElemType);
+Bits ElemSize = Bits(Ctx.getASTContext().getTypeSize(ElemType));
for (unsigned I = 0; I != FieldDesc->getNumElems(); ++I) {
enumerateData(P.atIndex(I).narrow(), Ctx, Offset, BitsToRead, F);
- Offset += ElemSizeInBits;
+ Offset += ElemSize;
if (Offset >= BitsToRead)
break;
}
@@ -233,16 +238,14 @@ static bool readPointerToBuffer(const Context &Ctx, const
Pointer &FromPtr,
return enumeratePointerFields(
FromPtr, Ctx, Buffer.size(),
- [&](const Pointer &P, PrimType T, Bits BitOffset,
+ [&](const Pointer &P, PrimType T, Bits BitOffset, Bits FullBitWidth,
bool PackedBools) -> bool {
-CharUnits ObjectReprChars = ASTCtx.getTypeSizeInChars(P.getType());
-Bits BitWidth = Bits(ASTCtx.toBits(ObjectReprChars));
-Bits FullBitWidth = BitWidth;
+Bits BitWidth = FullBitWidth;
-if (const FieldDecl *FD = P.getField(); FD && FD->isBitField()) {
+if (const FieldDecl *FD = P.getField(); FD && FD->isBitField())
BitWidth = Bits(std::min(FD->getBitWidthValue(ASTCtx),
(unsigned)FullBitWidth.getQuantity()));
-} else if (T == PT_Bool && PackedBools)
+else if (T == PT_Bool && PackedBools)
BitWidth = Bits(1);
if (BitWidth.isZero())
@@ -261,8 +264,7 @@ static bool readPointerToBuffer(const Context &Ctx, const
Pointer &FromPtr,
}
assert(P.isInitialized());
-auto Buff =
-std::make_unique(ObjectReprChars.getQuantity());
+auto Buff = std::make_unique(FullBitWidth.roundToBytes());
// Work around floating point types that contain unused padding bytes.
// This is really just `long double` on x86, which is the only
// fundamental type with padding bytes.
@@ -357,11 +359,9 @@ bool clang::interp::DoBitCastPtr(InterpState &S, CodePtr
OpPC,
ASTCtx.getTargetInfo().isLittleEndian() ? Endian::Little : Endian::Big;
bool Success = enumeratePointerFields(
ToPtr, S.getContext(), Buffer.size(),
- [&](const Pointer &P, PrimType T, Bits BitOffset,
+ [&](const Pointer &P, PrimType T, Bits BitOffset, Bits FullBitWidth,
bool PackedBools) -> bool {
QualType PtrType = P.getType();
-CharUnits ObjectReprChars = ASTCtx.getTypeSizeInChars(P
[clang] [llvm] [RISCV] Add stack clash protection (PR #117612)
lenary wrote: > > @rzinsly do you need someone to commit this? > > Yes, please. Please turn off [Keep my email addresses private](https://github.com/settings/emails) setting in your account. Otherwise the commit will be from the github hidden email which we don't use in LLVM. See [LLVM Discourse](https://discourse.llvm.org/t/hidden-emails-on-github-should-we-do-something-about-it) for more information. https://github.com/llvm/llvm-project/pull/117612 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [clang][ExprConst] Move vector diagnostics to checkBitCastConstexprEl… (PR #119366)
https://github.com/tbaederr closed https://github.com/llvm/llvm-project/pull/119366 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [clang][CodeGen] `sret` args should always point to the `alloca` AS, so use that (PR #114062)
@@ -800,7 +800,9 @@ static ABIArgInfo classifyExpandedType(SwiftAggLowering
&lowering,
if (lowering.empty()) {
return ABIArgInfo::getIgnore();
} else if (lowering.shouldPassIndirectly(forReturn)) {
-return ABIArgInfo::getIndirect(alignmentForIndirect, /*byval*/ false);
+return ABIArgInfo::getIndirect(alignmentForIndirect,
+ /*AddrSpace*/ 0,
+ /*byval*/ false);
AlexVlx wrote:
I did not intend to alter behaviour here, but merely to adopt the changed
interface. There's no easy way to use she LangAS::Default -> Target mapping
here in this helper function, hence the hardcoding to 0.
https://github.com/llvm/llvm-project/pull/114062
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [clang] Fix a crash issue that caused by handling of fields with initializers in nested anonymous unions (PR #113049)
yronglin wrote: > LGTM > > I'm a little concerned this is going to lead to other crashes due to exposing > more codepaths to "invalid" decls. But it seems like it's doing the right > thing in the given cases. Yeah, if the PR breaks clang, I'll revert it quickly. https://github.com/llvm/llvm-project/pull/113049 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] e6ba345 - [X86][AVX10.2] Add comments for the avx10_2copyintrin.h file (#119238)
Author: Mikołaj Piróg
Date: 2024-12-10T21:49:28+08:00
New Revision: e6ba3452ab086cf24725f3587709150b686d4b05
URL:
https://github.com/llvm/llvm-project/commit/e6ba3452ab086cf24725f3587709150b686d4b05
DIFF:
https://github.com/llvm/llvm-project/commit/e6ba3452ab086cf24725f3587709150b686d4b05.diff
LOG: [X86][AVX10.2] Add comments for the avx10_2copyintrin.h file (#119238)
As in title
Added:
Modified:
clang/lib/Headers/avx10_2copyintrin.h
Removed:
diff --git a/clang/lib/Headers/avx10_2copyintrin.h
b/clang/lib/Headers/avx10_2copyintrin.h
index 7fc31190781d91..76b8f8ced540d3 100644
--- a/clang/lib/Headers/avx10_2copyintrin.h
+++ b/clang/lib/Headers/avx10_2copyintrin.h
@@ -19,11 +19,43 @@
__attribute__((__always_inline__, __nodebug__, __target__("avx10.2-256"),
\
__min_vector_width__(128)))
+/// Constructs a 128-bit integer vector, setting the lower 32 bits to the
+///lower 32 bits of the parameter \a __A; the upper bits are zeoroed.
+///
+/// \code{.operation}
+/// result[31:0] := __A[31:0]
+/// result[MAX:32] := 0
+/// \endcode
+///
+/// \headerfile
+///
+/// This intrinsic corresponds to the VMOVD instruction.
+///
+/// \param __A
+///A 128-bit integer vector.
+/// \returns A 128-bit integer vector. The lower 32 bits are copied from the
+///parameter \a __A; the upper bits are zeroed.
static __inline__ __m128i __DEFAULT_FN_ATTRS128 _mm_move_epi32(__m128i __A) {
return (__m128i)__builtin_shufflevector(
(__v4si)__A, (__v4si)_mm_setzero_si128(), 0, 4, 4, 4);
}
+/// Constructs a 128-bit integer vector, setting the lower 16 bits to the
+///lower 16 bits of the parameter \a __A; the upper bits are zeoroed.
+///
+/// \code{.operation}
+/// result[15:0] := __A[15:0]
+/// result[MAX:16] := 0
+/// \endcode
+///
+/// \headerfile
+///
+/// This intrinsic corresponds to the VMOVW instruction.
+///
+/// \param __A
+///A 128-bit integer vector.
+/// \returns A 128-bit integer vector. The lower 16 bits are copied from the
+///parameter \a __A; the upper bits are zeroed.
static __inline__ __m128i __DEFAULT_FN_ATTRS128 _mm_move_epi16(__m128i __A) {
return (__m128i)__builtin_shufflevector(
(__v8hi)__A, (__v8hi)_mm_setzero_si128(), 0, 8, 8, 8, 8, 8, 8, 8);
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [X86][AVX10.2] Add comments for the avx10_2copyintrin.h file (PR #119238)
github-actions[bot] wrote: @mikolaj-pirog Congratulations on having your first Pull Request (PR) merged into the LLVM Project! Your changes will be combined with recent changes from other authors, then tested by our [build bots](https://lab.llvm.org/buildbot/). If there is a problem with a build, you may receive a report in an email or a comment on this PR. Please check whether problems have been caused by your change specifically, as the builds can include changes from many authors. It is not uncommon for your change to be included in a build that fails due to someone else's changes, or infrastructure issues. How to do this, and the rest of the post-merge process, is covered in detail [here](https://llvm.org/docs/MyFirstTypoFix.html#myfirsttypofix-issues-after-landing-your-pr). If your change does cause a problem, it may be reverted, or you can revert it yourself. This is a normal part of [LLVM development](https://llvm.org/docs/DeveloperPolicy.html#patch-reversion-policy). You can fix your changes and open a new PR to merge them again. If you don't get any reports, no action is required from you. Your changes are working as expected, well done! https://github.com/llvm/llvm-project/pull/119238 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [X86][AVX10.2] Add comments for the avx10_2copyintrin.h file (PR #119238)
https://github.com/phoebewang closed https://github.com/llvm/llvm-project/pull/119238 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] c5ab70c - [WebAssembly] Add `-i128:128` to the `datalayout` string. (#119204)
Author: Dan Gohman
Date: 2024-12-10T09:21:58-08:00
New Revision: c5ab70c508457eaece5d7ff4ab79a2f90bc67f06
URL:
https://github.com/llvm/llvm-project/commit/c5ab70c508457eaece5d7ff4ab79a2f90bc67f06
DIFF:
https://github.com/llvm/llvm-project/commit/c5ab70c508457eaece5d7ff4ab79a2f90bc67f06.diff
LOG: [WebAssembly] Add `-i128:128` to the `datalayout` string. (#119204)
Clang [defaults to aligning `__int128_t` to 16 bytes], while LLVM
`datalayout` strings [default to aligning `i128` to 8 bytes]. Wasm is
currently using the defaults for both, so it's inconsistent. Fix this by
adding `-i128:128` to Wasm's `datalayout` string so that it aligns
`i128` to 16 bytes too.
This is similar to
[llvm/llvm-project@dbad963](https://github.com/llvm/llvm-project/commit/dbad963a69fd7b16c6838f81b61167fbf00a413c)
for SPARC.
This fixes rust-lang/rust#133991; see that issue for further discussion.
[defaults to aligning `__int128_t` to 16 bytes]:
https://github.com/llvm/llvm-project/blob/f8b4182f076f8fe55f9d5f617b5a25008a77b22f/clang/lib/Basic/TargetInfo.cpp#L77
[default to aligning `i128` to 8 bytes]:
https://llvm.org/docs/LangRef.html#langref-datalayout
Added:
llvm/test/CodeGen/WebAssembly/data-align.ll
Modified:
clang/lib/Basic/Targets/WebAssembly.h
clang/test/CodeGen/target-data.c
llvm/lib/IR/AutoUpgrade.cpp
llvm/lib/Target/WebAssembly/WebAssemblyTargetMachine.cpp
llvm/test/CodeGen/WebAssembly/multi-return.ll
llvm/unittests/Bitcode/DataLayoutUpgradeTest.cpp
Removed:
diff --git a/clang/lib/Basic/Targets/WebAssembly.h
b/clang/lib/Basic/Targets/WebAssembly.h
index 1cae72e58e08ba..6ce2bb00e5f2a6 100644
--- a/clang/lib/Basic/Targets/WebAssembly.h
+++ b/clang/lib/Basic/Targets/WebAssembly.h
@@ -183,11 +183,12 @@ class LLVM_LIBRARY_VISIBILITY WebAssembly32TargetInfo
const TargetOptions &Opts)
: WebAssemblyTargetInfo(T, Opts) {
if (T.isOSEmscripten())
- resetDataLayout("e-m:e-p:32:32-p10:8:8-p20:8:8-i64:64-f128:64-n32:64-"
- "S128-ni:1:10:20");
-else
resetDataLayout(
- "e-m:e-p:32:32-p10:8:8-p20:8:8-i64:64-n32:64-S128-ni:1:10:20");
+ "e-m:e-p:32:32-p10:8:8-p20:8:8-i64:64-i128:128-f128:64-n32:64-"
+ "S128-ni:1:10:20");
+else
+ resetDataLayout("e-m:e-p:32:32-p10:8:8-p20:8:8-i64:64-i128:128-n32:64-"
+ "S128-ni:1:10:20");
}
protected:
@@ -207,11 +208,12 @@ class LLVM_LIBRARY_VISIBILITY WebAssembly64TargetInfo
PtrDiffType = SignedLong;
IntPtrType = SignedLong;
if (T.isOSEmscripten())
- resetDataLayout("e-m:e-p:64:64-p10:8:8-p20:8:8-i64:64-f128:64-n32:64-"
- "S128-ni:1:10:20");
-else
resetDataLayout(
- "e-m:e-p:64:64-p10:8:8-p20:8:8-i64:64-n32:64-S128-ni:1:10:20");
+ "e-m:e-p:64:64-p10:8:8-p20:8:8-i64:64-i128:128-f128:64-n32:64-"
+ "S128-ni:1:10:20");
+else
+ resetDataLayout("e-m:e-p:64:64-p10:8:8-p20:8:8-i64:64-i128:128-n32:64-"
+ "S128-ni:1:10:20");
}
protected:
diff --git a/clang/test/CodeGen/target-data.c
b/clang/test/CodeGen/target-data.c
index 2c758567faa1f6..71eb849433ed40 100644
--- a/clang/test/CodeGen/target-data.c
+++ b/clang/test/CodeGen/target-data.c
@@ -108,11 +108,11 @@
// RUN: %clang_cc1 -triple wasm32-unknown-unknown -o - -emit-llvm %s | \
// RUN: FileCheck %s -check-prefix=WEBASSEMBLY32
-// WEBASSEMBLY32: target datalayout =
"e-m:e-p:32:32-p10:8:8-p20:8:8-i64:64-n32:64-S128-ni:1:10:20"
+// WEBASSEMBLY32: target datalayout =
"e-m:e-p:32:32-p10:8:8-p20:8:8-i64:64-i128:128-n32:64-S128-ni:1:10:20"
// RUN: %clang_cc1 -triple wasm64-unknown-unknown -o - -emit-llvm %s | \
// RUN: FileCheck %s -check-prefix=WEBASSEMBLY64
-// WEBASSEMBLY64: target datalayout =
"e-m:e-p:64:64-p10:8:8-p20:8:8-i64:64-n32:64-S128-ni:1:10:20"
+// WEBASSEMBLY64: target datalayout =
"e-m:e-p:64:64-p10:8:8-p20:8:8-i64:64-i128:128-n32:64-S128-ni:1:10:20"
// RUN: %clang_cc1 -triple lanai-unknown-unknown -o - -emit-llvm %s | \
// RUN: FileCheck %s -check-prefix=LANAI
diff --git a/llvm/lib/IR/AutoUpgrade.cpp b/llvm/lib/IR/AutoUpgrade.cpp
index d54176a7878b5a..06e62bf7f9f757 100644
--- a/llvm/lib/IR/AutoUpgrade.cpp
+++ b/llvm/lib/IR/AutoUpgrade.cpp
@@ -5559,7 +5559,8 @@ std::string llvm::UpgradeDataLayoutString(StringRef DL,
StringRef TT) {
return Res;
}
- if (T.isSPARC() || (T.isMIPS64() && !DL.contains("m:m")) || T.isPPC64()) {
+ if (T.isSPARC() || (T.isMIPS64() && !DL.contains("m:m")) || T.isPPC64() ||
+ T.isWasm()) {
// Mips64 with o32 ABI did not add "-i128:128".
// Add "-i128:128"
std::string I64 = "-i64:64";
diff --git a/llvm/lib/Target/WebAssembly/WebAssemblyTargetMachine.cpp
b/llvm/lib/Target/WebAssembly/WebAssemblyTargetMachine.cpp
index 8ec72d5c47833a..9c95d730480f91 100644
--- a/llvm/lib/Target/WebAssembly/Web
[clang] 0fb0617 - [clang][bytecode] Check vector element types for eligibility (#119385)
Author: Timm Baeder
Date: 2024-12-10T17:56:48+01:00
New Revision: 0fb06172f14110daa45bed8ae4c153967c9365dc
URL:
https://github.com/llvm/llvm-project/commit/0fb06172f14110daa45bed8ae4c153967c9365dc
DIFF:
https://github.com/llvm/llvm-project/commit/0fb06172f14110daa45bed8ae4c153967c9365dc.diff
LOG: [clang][bytecode] Check vector element types for eligibility (#119385)
Like we do in ExprConstant.cpp.
Added:
Modified:
clang/lib/AST/ByteCode/InterpBuiltinBitCast.cpp
clang/test/AST/ByteCode/builtin-bit-cast.cpp
Removed:
diff --git a/clang/lib/AST/ByteCode/InterpBuiltinBitCast.cpp
b/clang/lib/AST/ByteCode/InterpBuiltinBitCast.cpp
index c9cd113287557b..7ba0faff252530 100644
--- a/clang/lib/AST/ByteCode/InterpBuiltinBitCast.cpp
+++ b/clang/lib/AST/ByteCode/InterpBuiltinBitCast.cpp
@@ -222,6 +222,35 @@ static bool CheckBitcastType(InterpState &S, CodePtr OpPC,
QualType T,
IsToType))
return false;
+ if (const auto *VT = T->getAs()) {
+const ASTContext &ASTCtx = S.getASTContext();
+QualType EltTy = VT->getElementType();
+unsigned NElts = VT->getNumElements();
+unsigned EltSize =
+VT->isExtVectorBoolType() ? 1 : ASTCtx.getTypeSize(EltTy);
+
+if ((NElts * EltSize) % ASTCtx.getCharWidth() != 0) {
+ // The vector's size in bits is not a multiple of the target's byte size,
+ // so its layout is unspecified. For now, we'll simply treat these cases
+ // as unsupported (this should only be possible with OpenCL bool vectors
+ // whose element count isn't a multiple of the byte size).
+ const Expr *E = S.Current->getExpr(OpPC);
+ S.FFDiag(E, diag::note_constexpr_bit_cast_invalid_vector)
+ << QualType(VT, 0) << EltSize << NElts << ASTCtx.getCharWidth();
+ return false;
+}
+
+if (EltTy->isRealFloatingType() &&
+&ASTCtx.getFloatTypeSemantics(EltTy) == &APFloat::x87DoubleExtended())
{
+ // The layout for x86_fp80 vectors seems to be handled very
inconsistently
+ // by both clang and LLVM, so for now we won't allow bit_casts involving
+ // it in a constexpr context.
+ const Expr *E = S.Current->getExpr(OpPC);
+ S.FFDiag(E, diag::note_constexpr_bit_cast_unsupported_type) << EltTy;
+ return false;
+}
+ }
+
return true;
}
diff --git a/clang/test/AST/ByteCode/builtin-bit-cast.cpp
b/clang/test/AST/ByteCode/builtin-bit-cast.cpp
index e99ab3904c339c..8a5bef635b8fde 100644
--- a/clang/test/AST/ByteCode/builtin-bit-cast.cpp
+++ b/clang/test/AST/ByteCode/builtin-bit-cast.cpp
@@ -502,3 +502,8 @@ namespace OversizedBitField {
// ref-note {{constexpr bit_cast involving bit-field is not yet supported}}
#endif
}
+
+typedef bool bool9 __attribute__((ext_vector_type(9)));
+// both-error@+2 {{constexpr variable 'bad_bool9_to_short' must be initialized
by a constant expression}}
+// both-note@+1 {{bit_cast involving type 'bool
__attribute__((ext_vector_type(9)))' (vector of 9 'bool' values) is not allowed
in a constant expression; element size 1 * element count 9 is not a multiple of
the byte size 8}}
+constexpr unsigned short bad_bool9_to_short = __builtin_bit_cast(unsigned
short, bool9{1,1,0,1,0,1,0,1,0});
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang-tools-extra] [clang-tidy][NFC] reorder cmake source file (PR #119374)
EugeneZelenko wrote: Did you look on `CMakeLists.txt` files in other directories? Just in case :-) https://github.com/llvm/llvm-project/pull/119374 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [clang][perf-training] Fix profiling with -DCLANG_BOLT=perf (PR #119117)
https://github.com/tstellar updated
https://github.com/llvm/llvm-project/pull/119117
>From 5d13b69039fab7c5960288cead18dc76f5d01f4f Mon Sep 17 00:00:00 2001
From: Tom Stellard
Date: Thu, 5 Dec 2024 15:01:27 +
Subject: [PATCH 1/5] [clang][perf-training] Fix profiling with
-DCLANG_BOLT=perf
This fixes the llvm-support build that generates the profile data.
However, I'm wondering if maybe we should disable llvm-suppot and
only run hello-world with -DCLANG_BOLT=perf. The bolt optimizations
with perf only give about a 3% performance increase (although maybe
with hw counters this would be better) and it takes a very long
time to convert all the perf profiles to the fdata format.
---
clang/utils/perf-training/bolt.lit.cfg| 22 +++
.../perf-training/llvm-support/build.test | 4 ++--
2 files changed, 20 insertions(+), 6 deletions(-)
diff --git a/clang/utils/perf-training/bolt.lit.cfg
b/clang/utils/perf-training/bolt.lit.cfg
index 1d0cf9a8a17a8e..04b18975275495 100644
--- a/clang/utils/perf-training/bolt.lit.cfg
+++ b/clang/utils/perf-training/bolt.lit.cfg
@@ -8,21 +8,32 @@ import subprocess
clang_bolt_mode = config.clang_bolt_mode.lower()
clang_binary = "clang"
-perf_wrapper = f"{config.python_exe} {config.perf_helper_dir}/perf-helper.py
perf "
+perf_wrapper = f"{config.python_exe} {config.perf_helper_dir}/perf-helper.py
perf"
if clang_bolt_mode == "instrument":
perf_wrapper = ""
clang_binary = config.clang_bolt_name
elif clang_bolt_mode == "lbr":
-perf_wrapper += " --lbr -- "
+perf_wrapper += " --lbr --"
elif clang_bolt_mode == "perf":
-perf_wrapper += " -- "
+perf_wrapper += " --"
else:
assert 0, "Unsupported CLANG_BOLT_MODE variable"
-config.clang = perf_wrapper + os.path.realpath(
+clang_nowrapper = os.path.realpath(
lit.util.which(clang_binary, config.clang_tools_dir)
).replace("\\", "/")
+config.clang = f'{perf_wrapper} {clang_nowrapper}'
+
+# We need to limit the number of build jobs with perf in order to avoid this
+# error:
+#
+# | Permission error mapping pages.
+# | Consider increasing /proc/sys/kernel/perf_event_mlock_kb,
+# | or try again with a smaller value of -m/--mmap_pages.
+ninja_args = ""
+if clang_bolt_mode != "instrument":
+ninja_args = "-j1"
config.name = "Clang Perf Training"
config.suffixes = [
@@ -52,3 +63,6 @@ config.substitutions.append(("%test_root",
config.test_exec_root))
config.substitutions.append(('%cmake_generator', config.cmake_generator))
config.substitutions.append(('%cmake', config.cmake_exe))
config.substitutions.append(('%llvm_src_dir', config.llvm_src_dir))
+config.substitutions.append(('%perf_cmake_compiler_launcher',
perf_wrapper.replace(' ', ';')))
+config.substitutions.append(('%nowrapper_clang', clang_nowrapper))
+config.substitutions.append(('%ninja_args', ninja_args))
diff --git a/clang/utils/perf-training/llvm-support/build.test
b/clang/utils/perf-training/llvm-support/build.test
index f29a594c846869..1f4d76502a3757 100644
--- a/clang/utils/perf-training/llvm-support/build.test
+++ b/clang/utils/perf-training/llvm-support/build.test
@@ -1,2 +1,2 @@
-RUN: %cmake -G %cmake_generator -B %t -S %llvm_src_dir
-DCMAKE_C_COMPILER=%clang -DCMAKE_CXX_COMPILER=%clang
-DCMAKE_CXX_FLAGS="--driver-mode=g++" -DCMAKE_BUILD_TYPE=Release
-RUN: %cmake --build %t -v --target LLVMSupport
+RUN: %cmake -G %cmake_generator -B %t -S %llvm_src_dir
-DCMAKE_CXX_COMPILER_LAUNCHER="%perf_cmake_compiler_launcher"
-DCMAKE_C_COMPILER="%nowrapper_clang" -DCMAKE_CXX_COMPILER="%nowrapper_clang"
-DCMAKE_CXX_FLAGS="--driver-mode=g++" -DCMAKE_BUILD_TYPE=Release
+RUN: %cmake --build %t %ninja_args -v --target LLVMSupport
>From 9aa48ac20e931d8192cecfec6ef789ea936fa6ff Mon Sep 17 00:00:00 2001
From: Tom Stellard
Date: Mon, 9 Dec 2024 19:23:40 +
Subject: [PATCH 2/5] Revert "[clang][perf-training] Fix profiling with
-DCLANG_BOLT=perf"
This reverts commit 5d13b69039fab7c5960288cead18dc76f5d01f4f.
---
clang/utils/perf-training/bolt.lit.cfg| 22 ---
.../perf-training/llvm-support/build.test | 4 ++--
2 files changed, 6 insertions(+), 20 deletions(-)
diff --git a/clang/utils/perf-training/bolt.lit.cfg
b/clang/utils/perf-training/bolt.lit.cfg
index 04b18975275495..1d0cf9a8a17a8e 100644
--- a/clang/utils/perf-training/bolt.lit.cfg
+++ b/clang/utils/perf-training/bolt.lit.cfg
@@ -8,32 +8,21 @@ import subprocess
clang_bolt_mode = config.clang_bolt_mode.lower()
clang_binary = "clang"
-perf_wrapper = f"{config.python_exe} {config.perf_helper_dir}/perf-helper.py
perf"
+perf_wrapper = f"{config.python_exe} {config.perf_helper_dir}/perf-helper.py
perf "
if clang_bolt_mode == "instrument":
perf_wrapper = ""
clang_binary = config.clang_bolt_name
elif clang_bolt_mode == "lbr":
-perf_wrapper += " --lbr --"
+perf_wrapper += " --lbr -- "
elif clang_bolt_mode == "perf":
-perf_wrapper += " --"
+perf_wrapper += " -- "
else:
assert 0, "Un
[clang] [llvm] [WebAssembly] Add `-i128:128` to the `datalayout` string. (PR #119204)
https://github.com/sunfishcode closed https://github.com/llvm/llvm-project/pull/119204 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [Darwin][Driver][clang] Prioritise command line args over `DEFAULT_SYSROOT` (PR #115993)
https://github.com/cachemeifyoucan approved this pull request. LGTM https://github.com/llvm/llvm-project/pull/115993 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [HLSL] Implement SV_GroupThreadId semantic (PR #117781)
https://github.com/llvm-beanz approved this pull request. https://github.com/llvm/llvm-project/pull/117781 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [Clang][OpenMP] Support for dispatch construct (Sema & Codegen) support (PR #117904)
https://github.com/SunilKuravinakop updated
https://github.com/llvm/llvm-project/pull/117904
>From 1703aa62cfe35538aedbacb28e907535e838248c Mon Sep 17 00:00:00 2001
From: Sunil Kuravinakop
Date: Fri, 20 Sep 2024 01:41:29 -0500
Subject: [PATCH 1/5] Support for dispatch construct (Sema & Codegen) support.
Support for clauses depend, novariants & nocontext.
---
.../clang/Basic/DiagnosticSemaKinds.td| 3 +
clang/include/clang/Basic/OpenMPKinds.h | 6 +
clang/include/clang/Sema/SemaOpenMP.h | 7 +
clang/lib/Basic/OpenMPKinds.cpp | 5 +
clang/lib/CodeGen/CGStmt.cpp | 2 +-
clang/lib/CodeGen/CGStmtOpenMP.cpp| 4 +
clang/lib/CodeGen/CodeGenFunction.h | 1 +
clang/lib/Sema/SemaOpenMP.cpp | 301 ++-
clang/test/OpenMP/dispatch_codegen.cpp| 359 ++
clang/test/OpenMP/dispatch_unsupported.c | 7 -
.../include/llvm/Frontend/OpenMP/OMPContext.h | 6 +
11 files changed, 687 insertions(+), 14 deletions(-)
create mode 100644 clang/test/OpenMP/dispatch_codegen.cpp
delete mode 100644 clang/test/OpenMP/dispatch_unsupported.c
diff --git a/clang/include/clang/Basic/DiagnosticSemaKinds.td
b/clang/include/clang/Basic/DiagnosticSemaKinds.td
index 8495884dcd058f..81b876f9fd85c5 100644
--- a/clang/include/clang/Basic/DiagnosticSemaKinds.td
+++ b/clang/include/clang/Basic/DiagnosticSemaKinds.td
@@ -11774,6 +11774,9 @@ def err_omp_clause_requires_dispatch_construct : Error<
"'%0' clause requires 'dispatch' context selector">;
def err_omp_append_args_with_varargs : Error<
"'append_args' is not allowed with varargs functions">;
+def warn_omp_dispatch_clause_novariants_nocontext : Warning<
+ "only 'novariants' clause is supported when 'novariants' & 'nocontext'
clauses occur on the same dispatch construct">,
+ InGroup;
def err_openmp_vla_in_task_untied : Error<
"variable length arrays are not supported in OpenMP tasking regions with
'untied' clause">;
def warn_omp_unterminated_declare_target : Warning<
diff --git a/clang/include/clang/Basic/OpenMPKinds.h
b/clang/include/clang/Basic/OpenMPKinds.h
index 900ad6ca6d66f6..7579fab43dbb19 100644
--- a/clang/include/clang/Basic/OpenMPKinds.h
+++ b/clang/include/clang/Basic/OpenMPKinds.h
@@ -269,6 +269,12 @@ bool isOpenMPTaskLoopDirective(OpenMPDirectiveKind DKind);
/// parallel', otherwise - false.
bool isOpenMPParallelDirective(OpenMPDirectiveKind DKind);
+/// Checks if the specified directive is a dispatch-kind directive.
+/// \param DKind Specified directive.
+/// \return true - the directive is a dispatch-like directive like 'omp
+/// dispatch', otherwise - false.
+bool isOpenMPDispatchDirective(OpenMPDirectiveKind DKind);
+
/// Checks if the specified directive is a target code offload directive.
/// \param DKind Specified directive.
/// \return true - the directive is a target code offload directive like
diff --git a/clang/include/clang/Sema/SemaOpenMP.h
b/clang/include/clang/Sema/SemaOpenMP.h
index 3d1cc4fab1c10f..80cee9e7583051 100644
--- a/clang/include/clang/Sema/SemaOpenMP.h
+++ b/clang/include/clang/Sema/SemaOpenMP.h
@@ -1462,6 +1462,13 @@ class SemaOpenMP : public SemaBase {
: OMPDeclareVariantScopes.back().TI;
}
+ StmtResult transformDispatchDirective(OpenMPDirectiveKind Kind,
+const DeclarationNameInfo &DirName,
+OpenMPDirectiveKind CancelRegion,
+ArrayRef Clauses,
+Stmt *AStmt, SourceLocation StartLoc,
+SourceLocation EndLoc);
+
/// The current `omp begin/end declare variant` scopes.
SmallVector OMPDeclareVariantScopes;
diff --git a/clang/lib/Basic/OpenMPKinds.cpp b/clang/lib/Basic/OpenMPKinds.cpp
index 62a13f01481b28..44ee63df46adb5 100644
--- a/clang/lib/Basic/OpenMPKinds.cpp
+++ b/clang/lib/Basic/OpenMPKinds.cpp
@@ -621,6 +621,11 @@ bool clang::isOpenMPParallelDirective(OpenMPDirectiveKind
DKind) {
llvm::is_contained(getLeafConstructs(DKind), OMPD_parallel);
}
+bool clang::isOpenMPDispatchDirective(OpenMPDirectiveKind DKind) {
+ return DKind == OMPD_dispatch ||
+ llvm::is_contained(getLeafConstructs(DKind), OMPD_target);
+}
+
bool clang::isOpenMPTargetExecutionDirective(OpenMPDirectiveKind DKind) {
return DKind == OMPD_target ||
llvm::is_contained(getLeafConstructs(DKind), OMPD_target);
diff --git a/clang/lib/CodeGen/CGStmt.cpp b/clang/lib/CodeGen/CGStmt.cpp
index 698baf853507f4..f3eedb79844378 100644
--- a/clang/lib/CodeGen/CGStmt.cpp
+++ b/clang/lib/CodeGen/CGStmt.cpp
@@ -417,7 +417,7 @@ void CodeGenFunction::EmitStmt(const Stmt *S,
ArrayRef Attrs) {
EmitOMPInteropDirective(cast(*S));
break;
case Stmt::OMPDispatchDirectiveClass:
-CGM.ErrorUnsupported(S, "OpenMP dispatch directive");
+
[clang] [llvm] [RISCV] Add stack clash protection (PR #117612)
topperc wrote: I should have a fix for the buildbot failure in a few minutes https://github.com/llvm/llvm-project/pull/117612 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [NVPTX] Add NVVMUpgradeAnnotations pass to cleanup legacy annotations (PR #119261)
@@ -1270,77 +1270,21 @@ exit:
; MODULE: attributes #[[ATTR1:[0-9]+]] = { convergent nocallback nounwind }
; MODULE: attributes #[[ATTR2:[0-9]+]] = { convergent nocallback nofree
nounwind willreturn }
; MODULE: attributes #[[ATTR3:[0-9]+]] = { nocallback nofree nosync nounwind
willreturn memory(inaccessiblemem: write) }
-; MODULE: attributes #[[ATTR4]] = { "kernel" }
-; MODULE: attributes #[[ATTR5]] = { nosync memory(none) }
+; MODULE: attributes #[[ATTR4]] = { "kernel" "nvvm.kernel" }
+; MODULE: attributes #[[ATTR5]] = { "kernel" }
+; MODULE: attributes #[[ATTR6]] = { nosync memory(none) }
;.
; CGSCC: attributes #[[ATTR0]] = { "llvm.assume"="ompx_aligned_barrier" }
; CGSCC: attributes #[[ATTR1:[0-9]+]] = { convergent nocallback nounwind }
; CGSCC: attributes #[[ATTR2:[0-9]+]] = { convergent nocallback nofree
nounwind willreturn }
; CGSCC: attributes #[[ATTR3:[0-9]+]] = { nocallback nofree nosync nounwind
willreturn memory(inaccessiblemem: write) }
-; CGSCC: attributes #[[ATTR4]] = { "kernel" }
-; CGSCC: attributes #[[ATTR5]] = { nosync memory(none) }
+; CGSCC: attributes #[[ATTR4]] = { "kernel" "nvvm.kernel" }
AlexMaclean wrote:
The problem is that OpenMP seems to need to be able to draw a distinction
between OpenMP kernels and nvvm kernels. For example here it seems like OpenMP
only wants to look at "kernel" not "nvvm.kernel". As a result it seems like
these attributes cannot be easily unified.
https://github.com/llvm/llvm-project/blob/c835b48a4d72227b174bcd86f071238a1583803a/llvm/lib/Transforms/IPO/OpenMPOpt.cpp#L5932-L5938
https://github.com/llvm/llvm-project/pull/119261
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [CIR] Cleanup: mlirContext and astContext (PR #119450)
https://github.com/smeenai approved this pull request. LGTM https://github.com/llvm/llvm-project/pull/119450 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [test][AArch64] Fix test in non-writeable dir (PR #119448)
llvmbot wrote: @llvm/pr-subscribers-clang Author: Jordan Rupprecht (rupprecht) Changes cc1a2ea61e3f8e790125b10d9ec4e7d179156ddf adds a test which can fail if the output directory is not writeable. Following the pattern of other tests in this package, use `-o -` to print the IR to stdout instead of to a file. --- Full diff: https://github.com/llvm/llvm-project/pull/119448.diff 1 Files Affected: - (modified) clang/test/Sema/aarch64-sve2-intrinsics/acle_sve2_fp8.c (+1-1) ``diff diff --git a/clang/test/Sema/aarch64-sve2-intrinsics/acle_sve2_fp8.c b/clang/test/Sema/aarch64-sve2-intrinsics/acle_sve2_fp8.c index aafd42f798d935..2e94c2314f1824 100644 --- a/clang/test/Sema/aarch64-sve2-intrinsics/acle_sve2_fp8.c +++ b/clang/test/Sema/aarch64-sve2-intrinsics/acle_sve2_fp8.c @@ -1,6 +1,6 @@ // REQUIRES: aarch64-registered-target -// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -verify -emit-llvm %s +// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -verify -emit-llvm -o - %s #include `` https://github.com/llvm/llvm-project/pull/119448 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [NFC] Updating Debug Info generation for 'this' (PR #119445)
dwblaikie wrote: any chance you could A/B test this on a bootstrap of clang, for instance? To help validate that this really is NFC/dead code? https://github.com/llvm/llvm-project/pull/119445 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [CIR] Cleanup: mlirContext and astContext (PR #119450)
llvmbot wrote:
@llvm/pr-subscribers-clangir
Author: David Olsen (dkolsen-pgi)
Changes
ClangIR CodeGen code uses both `mlir::MLIRContext` and `clang::ASTContext`
objects extensively. Refering to either of those as just "context" can be
confusing.
Change the names of all variables, parameters, and fields in
`clang/lib/CIR/CodeGen` that refer to `MLIRContext` or an `ASTContext` to be
either `mlirContext` or `astContext`.
This change is only the renaming of variables/parameters/fields. There are no
behavior changes. So there are no new tests or changes to existing tests.
This change mimics a recent change in the ClangIR incubator repository.
---
Full diff: https://github.com/llvm/llvm-project/pull/119450.diff
6 Files Affected:
- (modified) clang/include/clang/CIR/CIRGenerator.h (+3-3)
- (modified) clang/lib/CIR/CodeGen/CIRGenModule.cpp (+7-6)
- (modified) clang/lib/CIR/CodeGen/CIRGenModule.h (+3-3)
- (modified) clang/lib/CIR/CodeGen/CIRGenTypes.cpp (+8-6)
- (modified) clang/lib/CIR/CodeGen/CIRGenTypes.h (+1-1)
- (modified) clang/lib/CIR/CodeGen/CIRGenerator.cpp (+6-6)
``diff
diff --git a/clang/include/clang/CIR/CIRGenerator.h
b/clang/include/clang/CIR/CIRGenerator.h
index c8ca7e4bfa7285..414eba80b88b8f 100644
--- a/clang/include/clang/CIR/CIRGenerator.h
+++ b/clang/include/clang/CIR/CIRGenerator.h
@@ -37,14 +37,14 @@ namespace cir {
class CIRGenerator : public clang::ASTConsumer {
virtual void anchor();
clang::DiagnosticsEngine &diags;
- clang::ASTContext *astCtx;
+ clang::ASTContext *astContext;
// Only used for debug info.
llvm::IntrusiveRefCntPtr fs;
const clang::CodeGenOptions &codeGenOpts;
protected:
- std::unique_ptr mlirCtx;
+ std::unique_ptr mlirContext;
std::unique_ptr cgm;
public:
@@ -52,7 +52,7 @@ class CIRGenerator : public clang::ASTConsumer {
llvm::IntrusiveRefCntPtr fs,
const clang::CodeGenOptions &cgo);
~CIRGenerator() override;
- void Initialize(clang::ASTContext &astCtx) override;
+ void Initialize(clang::ASTContext &astContext) override;
bool HandleTopLevelDecl(clang::DeclGroupRef group) override;
mlir::ModuleOp getModule() const;
};
diff --git a/clang/lib/CIR/CodeGen/CIRGenModule.cpp
b/clang/lib/CIR/CodeGen/CIRGenModule.cpp
index e7c9512dcd3de8..0db24c3b41d18d 100644
--- a/clang/lib/CIR/CodeGen/CIRGenModule.cpp
+++ b/clang/lib/CIR/CodeGen/CIRGenModule.cpp
@@ -25,13 +25,14 @@
using namespace clang;
using namespace clang::CIRGen;
-CIRGenModule::CIRGenModule(mlir::MLIRContext &context,
- clang::ASTContext &astctx,
+CIRGenModule::CIRGenModule(mlir::MLIRContext &mlirContext,
+ clang::ASTContext &astContext,
const clang::CodeGenOptions &cgo,
DiagnosticsEngine &diags)
-: builder(context, *this), astCtx(astctx), langOpts(astctx.getLangOpts()),
- theModule{mlir::ModuleOp::create(mlir::UnknownLoc::get(&context))},
- diags(diags), target(astctx.getTargetInfo()), genTypes(*this) {
+: builder(mlirContext, *this), astContext(astContext),
+ langOpts(astContext.getLangOpts()),
+ theModule{mlir::ModuleOp::create(mlir::UnknownLoc::get(&mlirContext))},
+ diags(diags), target(astContext.getTargetInfo()), genTypes(*this) {
// Initialize cached types
SInt8Ty = cir::IntType::get(&getMLIRContext(), 8, /*isSigned=*/true);
@@ -48,7 +49,7 @@ CIRGenModule::CIRGenModule(mlir::MLIRContext &context,
mlir::Location CIRGenModule::getLoc(SourceLocation cLoc) {
assert(cLoc.isValid() && "expected valid source location");
- const SourceManager &sm = astCtx.getSourceManager();
+ const SourceManager &sm = astContext.getSourceManager();
PresumedLoc pLoc = sm.getPresumedLoc(cLoc);
StringRef filename = pLoc.getFilename();
return mlir::FileLineColLoc::get(builder.getStringAttr(filename),
diff --git a/clang/lib/CIR/CodeGen/CIRGenModule.h
b/clang/lib/CIR/CodeGen/CIRGenModule.h
index 397e501fd4e873..1c7ed63773900a 100644
--- a/clang/lib/CIR/CodeGen/CIRGenModule.h
+++ b/clang/lib/CIR/CodeGen/CIRGenModule.h
@@ -41,7 +41,7 @@ class CIRGenModule : public CIRGenTypeCache {
CIRGenModule &operator=(CIRGenModule &) = delete;
public:
- CIRGenModule(mlir::MLIRContext &context, clang::ASTContext &astctx,
+ CIRGenModule(mlir::MLIRContext &mlirContext, clang::ASTContext &astContext,
const clang::CodeGenOptions &cgo,
clang::DiagnosticsEngine &diags);
@@ -51,7 +51,7 @@ class CIRGenModule : public CIRGenTypeCache {
CIRGenBuilderTy builder;
/// Hold Clang AST information.
- clang::ASTContext &astCtx;
+ clang::ASTContext &astContext;
const clang::LangOptions &langOpts;
@@ -67,7 +67,7 @@ class CIRGenModule : public CIRGenTypeCache {
public:
mlir::ModuleOp getModule() const { return theModule; }
CIRGenBuilderTy &getBuilder() { return builder; }
- clang::ASTContext &getASTContext() const { return astCtx; }
[clang] 7eb73b9 - [CIR] Cleanup: mlirContext and astContext (#119450)
Author: David Olsen
Date: 2024-12-10T13:46:07-08:00
New Revision: 7eb73b95cb336cde14d5c755a09cd880bd3d5df9
URL:
https://github.com/llvm/llvm-project/commit/7eb73b95cb336cde14d5c755a09cd880bd3d5df9
DIFF:
https://github.com/llvm/llvm-project/commit/7eb73b95cb336cde14d5c755a09cd880bd3d5df9.diff
LOG: [CIR] Cleanup: mlirContext and astContext (#119450)
ClangIR CodeGen code uses both `mlir::MLIRContext` and
`clang::ASTContext` objects extensively. Refering to either of those as
just "context" can be confusing.
Change the names of all variables, parameters, and fields in
`clang/lib/CIR/CodeGen` that refer to `MLIRContext` or an `ASTContext`
to be either `mlirContext` or `astContext`.
This change is only the renaming of variables/parameters/fields. There
are no behavior changes. So there are no new tests or changes to
existing tests. This change mimics a recent change in the ClangIR
incubator repository.
Added:
Modified:
clang/include/clang/CIR/CIRGenerator.h
clang/lib/CIR/CodeGen/CIRGenModule.cpp
clang/lib/CIR/CodeGen/CIRGenModule.h
clang/lib/CIR/CodeGen/CIRGenTypes.cpp
clang/lib/CIR/CodeGen/CIRGenTypes.h
clang/lib/CIR/CodeGen/CIRGenerator.cpp
Removed:
diff --git a/clang/include/clang/CIR/CIRGenerator.h
b/clang/include/clang/CIR/CIRGenerator.h
index c8ca7e4bfa7285..414eba80b88b8f 100644
--- a/clang/include/clang/CIR/CIRGenerator.h
+++ b/clang/include/clang/CIR/CIRGenerator.h
@@ -37,14 +37,14 @@ namespace cir {
class CIRGenerator : public clang::ASTConsumer {
virtual void anchor();
clang::DiagnosticsEngine &diags;
- clang::ASTContext *astCtx;
+ clang::ASTContext *astContext;
// Only used for debug info.
llvm::IntrusiveRefCntPtr fs;
const clang::CodeGenOptions &codeGenOpts;
protected:
- std::unique_ptr mlirCtx;
+ std::unique_ptr mlirContext;
std::unique_ptr cgm;
public:
@@ -52,7 +52,7 @@ class CIRGenerator : public clang::ASTConsumer {
llvm::IntrusiveRefCntPtr fs,
const clang::CodeGenOptions &cgo);
~CIRGenerator() override;
- void Initialize(clang::ASTContext &astCtx) override;
+ void Initialize(clang::ASTContext &astContext) override;
bool HandleTopLevelDecl(clang::DeclGroupRef group) override;
mlir::ModuleOp getModule() const;
};
diff --git a/clang/lib/CIR/CodeGen/CIRGenModule.cpp
b/clang/lib/CIR/CodeGen/CIRGenModule.cpp
index e7c9512dcd3de8..0db24c3b41d18d 100644
--- a/clang/lib/CIR/CodeGen/CIRGenModule.cpp
+++ b/clang/lib/CIR/CodeGen/CIRGenModule.cpp
@@ -25,13 +25,14 @@
using namespace clang;
using namespace clang::CIRGen;
-CIRGenModule::CIRGenModule(mlir::MLIRContext &context,
- clang::ASTContext &astctx,
+CIRGenModule::CIRGenModule(mlir::MLIRContext &mlirContext,
+ clang::ASTContext &astContext,
const clang::CodeGenOptions &cgo,
DiagnosticsEngine &diags)
-: builder(context, *this), astCtx(astctx), langOpts(astctx.getLangOpts()),
- theModule{mlir::ModuleOp::create(mlir::UnknownLoc::get(&context))},
- diags(diags), target(astctx.getTargetInfo()), genTypes(*this) {
+: builder(mlirContext, *this), astContext(astContext),
+ langOpts(astContext.getLangOpts()),
+ theModule{mlir::ModuleOp::create(mlir::UnknownLoc::get(&mlirContext))},
+ diags(diags), target(astContext.getTargetInfo()), genTypes(*this) {
// Initialize cached types
SInt8Ty = cir::IntType::get(&getMLIRContext(), 8, /*isSigned=*/true);
@@ -48,7 +49,7 @@ CIRGenModule::CIRGenModule(mlir::MLIRContext &context,
mlir::Location CIRGenModule::getLoc(SourceLocation cLoc) {
assert(cLoc.isValid() && "expected valid source location");
- const SourceManager &sm = astCtx.getSourceManager();
+ const SourceManager &sm = astContext.getSourceManager();
PresumedLoc pLoc = sm.getPresumedLoc(cLoc);
StringRef filename = pLoc.getFilename();
return mlir::FileLineColLoc::get(builder.getStringAttr(filename),
diff --git a/clang/lib/CIR/CodeGen/CIRGenModule.h
b/clang/lib/CIR/CodeGen/CIRGenModule.h
index 397e501fd4e873..1c7ed63773900a 100644
--- a/clang/lib/CIR/CodeGen/CIRGenModule.h
+++ b/clang/lib/CIR/CodeGen/CIRGenModule.h
@@ -41,7 +41,7 @@ class CIRGenModule : public CIRGenTypeCache {
CIRGenModule &operator=(CIRGenModule &) = delete;
public:
- CIRGenModule(mlir::MLIRContext &context, clang::ASTContext &astctx,
+ CIRGenModule(mlir::MLIRContext &mlirContext, clang::ASTContext &astContext,
const clang::CodeGenOptions &cgo,
clang::DiagnosticsEngine &diags);
@@ -51,7 +51,7 @@ class CIRGenModule : public CIRGenTypeCache {
CIRGenBuilderTy builder;
/// Hold Clang AST information.
- clang::ASTContext &astCtx;
+ clang::ASTContext &astContext;
const clang::LangOptions &langOpts;
@@ -67,7 +67,7 @@ class CIRGenModule : public CIRGenTypeCache {
public:
[clang-tools-extra] Add new tool: clang-read-diagnostics (PR #118522)
https://github.com/yuxuanchen1997 closed https://github.com/llvm/llvm-project/pull/118522 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [Clang] C++ Templates: Refactor and fix `TransformLambdaExpr`'s mishandling of TypeLocs (PR #78801)
https://github.com/yuxuanchen1997 closed https://github.com/llvm/llvm-project/pull/78801 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] 54ca1c4 - [clang-format] Fix idempotent format of hash in macro body (#118513)
Author: Owen Pan
Date: 2024-12-10T16:47:21-08:00
New Revision: 54ca1c4212e7ff3df880adb1a04dc3d41c033681
URL:
https://github.com/llvm/llvm-project/commit/54ca1c4212e7ff3df880adb1a04dc3d41c033681
DIFF:
https://github.com/llvm/llvm-project/commit/54ca1c4212e7ff3df880adb1a04dc3d41c033681.diff
LOG: [clang-format] Fix idempotent format of hash in macro body (#118513)
Fixes #118334.
Added:
Modified:
clang/lib/Format/UnwrappedLineParser.cpp
clang/unittests/Format/FormatTest.cpp
clang/unittests/Format/TokenAnnotatorTest.cpp
Removed:
diff --git a/clang/lib/Format/UnwrappedLineParser.cpp
b/clang/lib/Format/UnwrappedLineParser.cpp
index c182aaf0876d1b..de7e261b21d303 100644
--- a/clang/lib/Format/UnwrappedLineParser.cpp
+++ b/clang/lib/Format/UnwrappedLineParser.cpp
@@ -512,7 +512,7 @@ void UnwrappedLineParser::calculateBraceTypes(bool
ExpectClassBody) {
break;
do {
NextTok = Tokens->getNextToken();
-} while (NextTok->NewlinesBefore == 0 && NextTok->isNot(tok::eof));
+} while (!NextTok->HasUnescapedNewline && NextTok->isNot(tok::eof));
while (NextTok->is(tok::comment))
NextTok = Tokens->getNextToken();
diff --git a/clang/unittests/Format/FormatTest.cpp
b/clang/unittests/Format/FormatTest.cpp
index 93e1bbffe806fa..eeb857a914b1a2 100644
--- a/clang/unittests/Format/FormatTest.cpp
+++ b/clang/unittests/Format/FormatTest.cpp
@@ -5724,6 +5724,24 @@ TEST_F(FormatTest, HashInMacroDefinition) {
getLLVMStyleWithColumns(22));
verifyFormat("#define A void # ## #", getLLVMStyleWithColumns(22));
+
+#if 0
+ // FIXME: The correct format is:
+ verifyFormat("{\n"
+ " {\n"
+ "#define GEN_ID(_x) char *_x{#_x}\n"
+ "GEN_ID(one);\n"
+ " }\n"
+ "}");
+#endif
+ verifyFormat("{\n"
+ " {\n"
+ "#define GEN_ID(_x) \\\n"
+ " char *_x { #_x }\n"
+ "GEN_ID(one);\n"
+ " }\n"
+ "}",
+ getGoogleStyle());
}
TEST_F(FormatTest, RespectWhitespaceInMacroDefinitions) {
diff --git a/clang/unittests/Format/TokenAnnotatorTest.cpp
b/clang/unittests/Format/TokenAnnotatorTest.cpp
index 9db3187ac44e70..38658fcb0e9990 100644
--- a/clang/unittests/Format/TokenAnnotatorTest.cpp
+++ b/clang/unittests/Format/TokenAnnotatorTest.cpp
@@ -3346,6 +3346,25 @@ TEST_F(TokenAnnotatorTest, BraceKind) {
EXPECT_BRACE_KIND(Tokens[11], BK_BracedInit);
EXPECT_BRACE_KIND(Tokens[13], BK_Block);
+ Tokens = annotate("{\n"
+" {\n"
+"#define GEN_ID(_x) char *_x{#_x}\n"
+"GEN_ID(one);\n"
+" }\n"
+"}");
+ ASSERT_EQ(Tokens.size(), 23u) << Tokens;
+ EXPECT_TOKEN(Tokens[0], tok::l_brace, TT_BlockLBrace);
+ EXPECT_BRACE_KIND(Tokens[0], BK_Block);
+ EXPECT_TOKEN(Tokens[1], tok::l_brace, TT_BlockLBrace);
+ EXPECT_BRACE_KIND(Tokens[1], BK_Block);
+#if 0
+ // FIXME:
+ EXPECT_BRACE_KIND(Tokens[11], BK_BracedInit);
+ EXPECT_BRACE_KIND(Tokens[14], BK_BracedInit);
+#endif
+ EXPECT_BRACE_KIND(Tokens[20], BK_Block);
+ EXPECT_BRACE_KIND(Tokens[21], BK_Block);
+
Tokens = annotate("a = class extends goog.a {};",
getGoogleStyle(FormatStyle::LK_JavaScript));
ASSERT_EQ(Tokens.size(), 11u) << Tokens;
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [clang-format] Fix idempotent format of hash in macro body (PR #118513)
https://github.com/owenca closed https://github.com/llvm/llvm-project/pull/118513 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [Clang][NFC] Refactor suspend emit logic in coroutine codegen (PR #73564)
https://github.com/yuxuanchen1997 closed https://github.com/llvm/llvm-project/pull/73564 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [Clang] Coroutines: Properly Check if `await_suspend` return type is a `std::coroutine_handle` (PR #85684)
@@ -418,39 +448,60 @@ static ReadySuspendResumeResult buildCoawaitCalls(Sema
&S, VarDecl *CoroPromise,
return Calls;
}
Expr *CoroHandle = CoroHandleRes.get();
- CallExpr *AwaitSuspend = cast_or_null(
- BuildSubExpr(ACT::ACT_Suspend, "await_suspend", CoroHandle));
+ auto *AwaitSuspend = [&]() -> CallExpr * {
+auto *SubExpr = BuildSubExpr(ACT::ACT_Suspend, "await_suspend",
CoroHandle);
+if (!SubExpr)
+ return nullptr;
+if (auto *E = dyn_cast(SubExpr)) {
+ // This happens when await_suspend return type is not trivially
+ // destructible. This doesn't happen for the permitted return types of
+ // such function. Diagnose it later.
+ return cast_or_null(E->getSubExpr());
+} else {
+ return cast_or_null(SubExpr);
+}
yuxuanchen1997 wrote:
I still want to fix this. Let me rebase and revive it.
https://github.com/llvm/llvm-project/pull/85684
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [Clang] Coroutines: Properly Check if `await_suspend` return type is a `std::coroutine_handle` (PR #85684)
https://github.com/yuxuanchen1997 updated
https://github.com/llvm/llvm-project/pull/85684
>From b843c2f71a1a43cb897b557f783d60c6bf26f687 Mon Sep 17 00:00:00 2001
From: Yuxuan Chen
Date: Mon, 18 Mar 2024 10:45:20 -0700
Subject: [PATCH] Check if Coroutine await_suspend type returns the right type
---
clang/docs/ReleaseNotes.rst | 3 +
.../clang/Basic/DiagnosticSemaKinds.td| 2 +-
clang/lib/Sema/SemaCoroutine.cpp | 119 +-
clang/test/SemaCXX/coroutines.cpp | 28 -
4 files changed, 111 insertions(+), 41 deletions(-)
diff --git a/clang/docs/ReleaseNotes.rst b/clang/docs/ReleaseNotes.rst
index d6e179ca9d6904..f7b44e5e65641c 100644
--- a/clang/docs/ReleaseNotes.rst
+++ b/clang/docs/ReleaseNotes.rst
@@ -48,6 +48,9 @@ C++ Specific Potentially Breaking Changes
- Clang now diagnoses function/variable templates that shadow their own
template parameters, e.g. ``template void T();``.
This error can be disabled via `-Wno-strict-primary-template-shadow` for
compatibility with previous versions of clang.
+- Clang now emits errors for coroutine `await_suspend` functions whose return
type is not
+ one of `void`, `bool`, or `std::coroutine_handle`.
+
ABI Changes in This Version
---
- Fixed Microsoft name mangling of implicitly defined variables used for thread
diff --git a/clang/include/clang/Basic/DiagnosticSemaKinds.td
b/clang/include/clang/Basic/DiagnosticSemaKinds.td
index fc727cef9cd835..796b3d9d5e1190 100644
--- a/clang/include/clang/Basic/DiagnosticSemaKinds.td
+++ b/clang/include/clang/Basic/DiagnosticSemaKinds.td
@@ -11707,7 +11707,7 @@ def err_coroutine_promise_new_requires_nothrow : Error<
def note_coroutine_promise_call_implicitly_required : Note<
"call to %0 implicitly required by coroutine function here">;
def err_await_suspend_invalid_return_type : Error<
- "return type of 'await_suspend' is required to be 'void' or 'bool' (have %0)"
+ "return type of 'await_suspend' is required to be 'void', 'bool', or
'std::coroutine_handle' (have %0)"
>;
def note_await_ready_no_bool_conversion : Note<
"return type of 'await_ready' is required to be contextually convertible to
'bool'"
diff --git a/clang/lib/Sema/SemaCoroutine.cpp b/clang/lib/Sema/SemaCoroutine.cpp
index 736632857efc36..2e81a83b62df51 100644
--- a/clang/lib/Sema/SemaCoroutine.cpp
+++ b/clang/lib/Sema/SemaCoroutine.cpp
@@ -137,12 +137,8 @@ static QualType lookupPromiseType(Sema &S, const
FunctionDecl *FD,
return PromiseType;
}
-/// Look up the std::coroutine_handle.
-static QualType lookupCoroutineHandleType(Sema &S, QualType PromiseType,
- SourceLocation Loc) {
- if (PromiseType.isNull())
-return QualType();
-
+static ClassTemplateDecl *lookupCoroutineHandleTemplate(Sema &S,
+SourceLocation Loc) {
NamespaceDecl *CoroNamespace = S.getStdNamespace();
assert(CoroNamespace && "Should already be diagnosed");
@@ -151,18 +147,32 @@ static QualType lookupCoroutineHandleType(Sema &S,
QualType PromiseType,
if (!S.LookupQualifiedName(Result, CoroNamespace)) {
S.Diag(Loc, diag::err_implied_coroutine_type_not_found)
<< "std::coroutine_handle";
-return QualType();
+return nullptr;
}
- ClassTemplateDecl *CoroHandle = Result.getAsSingle();
+ auto *CoroHandle = Result.getAsSingle();
+
if (!CoroHandle) {
Result.suppressDiagnostics();
// We found something weird. Complain about the first thing we found.
NamedDecl *Found = *Result.begin();
S.Diag(Found->getLocation(), diag::err_malformed_std_coroutine_handle);
-return QualType();
+return nullptr;
}
+ return CoroHandle;
+}
+
+/// Look up the std::coroutine_handle.
+static QualType lookupCoroutineHandleType(Sema &S, QualType PromiseType,
+ SourceLocation Loc) {
+ if (PromiseType.isNull())
+return QualType();
+
+ ClassTemplateDecl *CoroHandle = lookupCoroutineHandleTemplate(S, Loc);
+ if (!CoroHandle)
+return QualType();
+
// Form template argument list for coroutine_handle.
TemplateArgumentListInfo Args(Loc, Loc);
Args.addArgument(TemplateArgumentLoc(
@@ -331,16 +341,12 @@ static ExprResult buildMemberCall(Sema &S, Expr *Base,
SourceLocation Loc,
// coroutine.
static Expr *maybeTailCall(Sema &S, QualType RetType, Expr *E,
SourceLocation Loc) {
- if (RetType->isReferenceType())
-return nullptr;
+ assert(!RetType->isReferenceType() &&
+ "Should have diagnosed reference types.");
Type const *T = RetType.getTypePtr();
if (!T->isClassType() && !T->isStructureType())
return nullptr;
- // FIXME: Add convertability check to coroutine_handle<>. Possibly via
- // EvaluateBinaryTypeTrait(BTT_IsConvertible, ...) which is at the moment
- // a private function in SemaExprCXX.cpp
-
ExprResult AddressExpr = b
[clang] [llvm] [NVPTX] Add NVVMUpgradeAnnotations pass to cleanup legacy annotations (PR #119261)
@@ -1270,77 +1270,21 @@ exit:
; MODULE: attributes #[[ATTR1:[0-9]+]] = { convergent nocallback nounwind }
; MODULE: attributes #[[ATTR2:[0-9]+]] = { convergent nocallback nofree
nounwind willreturn }
; MODULE: attributes #[[ATTR3:[0-9]+]] = { nocallback nofree nosync nounwind
willreturn memory(inaccessiblemem: write) }
-; MODULE: attributes #[[ATTR4]] = { "kernel" }
-; MODULE: attributes #[[ATTR5]] = { nosync memory(none) }
+; MODULE: attributes #[[ATTR4]] = { "kernel" "nvvm.kernel" }
+; MODULE: attributes #[[ATTR5]] = { "kernel" }
+; MODULE: attributes #[[ATTR6]] = { nosync memory(none) }
;.
; CGSCC: attributes #[[ATTR0]] = { "llvm.assume"="ompx_aligned_barrier" }
; CGSCC: attributes #[[ATTR1:[0-9]+]] = { convergent nocallback nounwind }
; CGSCC: attributes #[[ATTR2:[0-9]+]] = { convergent nocallback nofree
nounwind willreturn }
; CGSCC: attributes #[[ATTR3:[0-9]+]] = { nocallback nofree nosync nounwind
willreturn memory(inaccessiblemem: write) }
-; CGSCC: attributes #[[ATTR4]] = { "kernel" }
-; CGSCC: attributes #[[ATTR5]] = { nosync memory(none) }
+; CGSCC: attributes #[[ATTR4]] = { "kernel" "nvvm.kernel" }
shiltian wrote:
Yes, as the comment pointed out. Only OpenMP target regions are marked as
`kernel`. Nothing else.
https://github.com/llvm/llvm-project/pull/119261
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [NVPTX] Add NVVMUpgradeAnnotations pass to cleanup legacy annotations (PR #119261)
@@ -324,14 +326,15 @@ MaybeAlign getAlign(const Function &F, unsigned Index) {
F.getAttributes().getAttributes(Index).getStackAlignment())
return StackAlign;
- // If that is missing, check the legacy nvvm metadata
- std::vector Vs;
- bool retval = findAllNVVMAnnotation(&F, "align", Vs);
- if (!retval)
-return std::nullopt;
- for (unsigned V : Vs)
-if ((V >> 16) == Index)
- return Align(V & 0x);
+ // check the legacy nvvm metadata only for the return value since llvm does
+ // not support stackalign attribute for this.
+ if (Index == 0) {
+std::vector Vs;
+if (findAllNVVMAnnotation(&F, "align", Vs))
AlexMaclean wrote:
Yea, I agree the NVVM annotation APIs could be cleaned up significantly,
hopefully this work will remove the need for them altogether though.
https://github.com/llvm/llvm-project/pull/119261
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [NVPTX] Add NVVMUpgradeAnnotations pass to cleanup legacy annotations (PR #119261)
@@ -1270,77 +1270,21 @@ exit:
; MODULE: attributes #[[ATTR1:[0-9]+]] = { convergent nocallback nounwind }
; MODULE: attributes #[[ATTR2:[0-9]+]] = { convergent nocallback nofree
nounwind willreturn }
; MODULE: attributes #[[ATTR3:[0-9]+]] = { nocallback nofree nosync nounwind
willreturn memory(inaccessiblemem: write) }
-; MODULE: attributes #[[ATTR4]] = { "kernel" }
-; MODULE: attributes #[[ATTR5]] = { nosync memory(none) }
+; MODULE: attributes #[[ATTR4]] = { "kernel" "nvvm.kernel" }
+; MODULE: attributes #[[ATTR5]] = { "kernel" }
+; MODULE: attributes #[[ATTR6]] = { nosync memory(none) }
;.
; CGSCC: attributes #[[ATTR0]] = { "llvm.assume"="ompx_aligned_barrier" }
; CGSCC: attributes #[[ATTR1:[0-9]+]] = { convergent nocallback nounwind }
; CGSCC: attributes #[[ATTR2:[0-9]+]] = { convergent nocallback nofree
nounwind willreturn }
; CGSCC: attributes #[[ATTR3:[0-9]+]] = { nocallback nofree nosync nounwind
willreturn memory(inaccessiblemem: write) }
-; CGSCC: attributes #[[ATTR4]] = { "kernel" }
-; CGSCC: attributes #[[ATTR5]] = { nosync memory(none) }
+; CGSCC: attributes #[[ATTR4]] = { "kernel" "nvvm.kernel" }
shiltian wrote:
Theoretically yes, but I don't see how that works in the current compilation
pipeline. However, both CUDA and OpenMP kernels could reside in the same LLVM
module at link time.
https://github.com/llvm/llvm-project/pull/119261
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [ubsan] Don't merge non-trap handlers if -ubsan-unique-traps or not optimized (PR #119302)
https://github.com/thurstond closed https://github.com/llvm/llvm-project/pull/119302 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [compiler-rt] [llvm] [XRay][RISCV] RISCV support for XRay (PR #117368)
https://github.com/topperc approved this pull request. LGTM https://github.com/llvm/llvm-project/pull/117368 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] UncountedLocalVarsChecker and UncheckedLocalVarsChecker should recognize signletons. (PR #119339)
https://github.com/t-rasmud approved this pull request. https://github.com/llvm/llvm-project/pull/119339 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [clang][Modules] Fixing Build Breaks When -DLLVM_ENABLE_MODULES=ON (PR #119473)
llvmbot wrote:
@llvm/pr-subscribers-clang
Author: Qiongsi Wu (qiongsiwu)
Changes
A few recent changes are causing build breaks when `-DLLVM_ENABLE_MODULES=ON`
(such as 834dfd23155351c9885eddf7b9664f7697326946 and
7dfdca1961aadc75ca397818bfb9bd32f1879248).
This PR makes the required updates so that clang/llvm builds when
`-DLLVM_ENABLE_MODULES=ON`.
---
Full diff: https://github.com/llvm/llvm-project/pull/119473.diff
9 Files Affected:
- (modified) clang/include/clang/Support/Compiler.h (-2)
- (modified) clang/include/module.modulemap (+4-1)
- (modified) llvm/include/llvm/DebugInfo/DWARF/DWARFTypePrinter.h (+1)
- (modified) llvm/include/llvm/IR/NVVMIntrinsicFlags.h (+2)
- (modified) llvm/include/llvm/Support/Memory.h (+1)
- (modified) llvm/include/llvm/TargetParser/AArch64TargetParser.h (+1)
- (modified) llvm/include/module.modulemap (+1)
- (modified) llvm/tools/llvm-jitlink/llvm-jitlink-executor/CMakeLists.txt (+1)
- (modified) llvm/unittests/tools/llvm-profgen/CMakeLists.txt (+1)
``diff
diff --git a/clang/include/clang/Support/Compiler.h
b/clang/include/clang/Support/Compiler.h
index 13582b899dc2a6..0cba1ad7331682 100644
--- a/clang/include/clang/Support/Compiler.h
+++ b/clang/include/clang/Support/Compiler.h
@@ -14,8 +14,6 @@
#ifndef CLANG_SUPPORT_COMPILER_H
#define CLANG_SUPPORT_COMPILER_H
-#include "llvm/Support/Compiler.h"
-
/// CLANG_ABI is the main export/visibility macro to mark something as
/// explicitly exported when clang is built as a shared library with everything
/// else that is unannotated having hidden visibility.
diff --git a/clang/include/module.modulemap b/clang/include/module.modulemap
index b399f0beee59a1..e9218c715147b0 100644
--- a/clang/include/module.modulemap
+++ b/clang/include/module.modulemap
@@ -115,7 +115,7 @@ module Clang_Diagnostics {
module Driver { header "clang/Driver/DriverDiagnostic.h" export * }
module Frontend { header "clang/Frontend/FrontendDiagnostic.h" export * }
module Lex { header "clang/Lex/LexDiagnostic.h" export * }
- module Parse { header "clang/Parse/ParseDiagnostic.h" export * }
+ module Parse { header "clang/Basic/DiagnosticParse.h" export * }
module Serialization { header
"clang/Serialization/SerializationDiagnostic.h" export * }
module Refactoring { header
"clang/Tooling/Refactoring/RefactoringDiagnostic.h" export * }
}
@@ -186,6 +186,9 @@ module Clang_StaticAnalyzer_Frontend {
module Clang_Testing {
requires cplusplus
umbrella "clang/Testing"
+
+ textual header "clang/Testing/TestLanguage.def"
+
module * { export * }
}
diff --git a/llvm/include/llvm/DebugInfo/DWARF/DWARFTypePrinter.h
b/llvm/include/llvm/DebugInfo/DWARF/DWARFTypePrinter.h
index 3c936b93865045..bd25f6c30ebf1f 100644
--- a/llvm/include/llvm/DebugInfo/DWARF/DWARFTypePrinter.h
+++ b/llvm/include/llvm/DebugInfo/DWARF/DWARFTypePrinter.h
@@ -9,6 +9,7 @@
#ifndef LLVM_DEBUGINFO_DWARF_DWARFTYPEPRINTER_H
#define LLVM_DEBUGINFO_DWARF_DWARFTYPEPRINTER_H
+#include "llvm/ADT/StringExtras.h"
#include "llvm/ADT/StringRef.h"
#include "llvm/BinaryFormat/Dwarf.h"
#include "llvm/Support/Error.h"
diff --git a/llvm/include/llvm/IR/NVVMIntrinsicFlags.h
b/llvm/include/llvm/IR/NVVMIntrinsicFlags.h
index 43dde42bbbd620..dfb6e857b3a6ad 100644
--- a/llvm/include/llvm/IR/NVVMIntrinsicFlags.h
+++ b/llvm/include/llvm/IR/NVVMIntrinsicFlags.h
@@ -15,6 +15,8 @@
#ifndef LLVM_IR_NVVMINTRINSICFLAGS_H
#define LLVM_IR_NVVMINTRINSICFLAGS_H
+#include
+
namespace llvm {
namespace nvvm {
diff --git a/llvm/include/llvm/Support/Memory.h
b/llvm/include/llvm/Support/Memory.h
index c02a3cc14dc7de..a587f2a8542b2e 100644
--- a/llvm/include/llvm/Support/Memory.h
+++ b/llvm/include/llvm/Support/Memory.h
@@ -15,6 +15,7 @@
#include "llvm/Support/DataTypes.h"
#include
+#include
namespace llvm {
diff --git a/llvm/include/llvm/TargetParser/AArch64TargetParser.h
b/llvm/include/llvm/TargetParser/AArch64TargetParser.h
index 1311329821828f..11836e87d1b208 100644
--- a/llvm/include/llvm/TargetParser/AArch64TargetParser.h
+++ b/llvm/include/llvm/TargetParser/AArch64TargetParser.h
@@ -16,6 +16,7 @@
#include "llvm/ADT/ArrayRef.h"
#include "llvm/ADT/Bitset.h"
+#include "llvm/ADT/StringExtras.h"
#include "llvm/ADT/StringMap.h"
#include "llvm/ADT/StringRef.h"
#include "llvm/Support/VersionTuple.h"
diff --git a/llvm/include/module.modulemap b/llvm/include/module.modulemap
index b00da6d7cd28c7..ac7bd257cbb422 100644
--- a/llvm/include/module.modulemap
+++ b/llvm/include/module.modulemap
@@ -346,6 +346,7 @@ extern module LLVM_Extern_Utils_DataTypes
"module.extern.modulemap"
// TargetParser module before building the TargetParser module itself.
module TargetParserGen {
module AArch64TargetParserDef {
+textual header "llvm/TargetParser/AArch64CPUFeatures.inc"
header "llvm/TargetParser/AArch64TargetParser.h"
extern module LLVM_Extern_TargetParser_Gen "module.extern.modulemap"
export *
diff --git a/llvm/tools/llvm-jitli
[clang] [llvm] [clang][Modules] Fixing Build Breaks When -DLLVM_ENABLE_MODULES=ON (PR #119473)
https://github.com/qiongsiwu converted_to_draft https://github.com/llvm/llvm-project/pull/119473 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [clang][Modules] Fixing Build Breaks When -DLLVM_ENABLE_MODULES=ON (PR #119473)
llvmbot wrote:
@llvm/pr-subscribers-llvm-support
Author: Qiongsi Wu (qiongsiwu)
Changes
A few recent changes are causing build breaks when `-DLLVM_ENABLE_MODULES=ON`
(such as 834dfd23155351c9885eddf7b9664f7697326946 and
7dfdca1961aadc75ca397818bfb9bd32f1879248).
This PR makes the required updates so that clang/llvm builds when
`-DLLVM_ENABLE_MODULES=ON`.
---
Full diff: https://github.com/llvm/llvm-project/pull/119473.diff
9 Files Affected:
- (modified) clang/include/clang/Support/Compiler.h (-2)
- (modified) clang/include/module.modulemap (+4-1)
- (modified) llvm/include/llvm/DebugInfo/DWARF/DWARFTypePrinter.h (+1)
- (modified) llvm/include/llvm/IR/NVVMIntrinsicFlags.h (+2)
- (modified) llvm/include/llvm/Support/Memory.h (+1)
- (modified) llvm/include/llvm/TargetParser/AArch64TargetParser.h (+1)
- (modified) llvm/include/module.modulemap (+1)
- (modified) llvm/tools/llvm-jitlink/llvm-jitlink-executor/CMakeLists.txt (+1)
- (modified) llvm/unittests/tools/llvm-profgen/CMakeLists.txt (+1)
``diff
diff --git a/clang/include/clang/Support/Compiler.h
b/clang/include/clang/Support/Compiler.h
index 13582b899dc2a6..0cba1ad7331682 100644
--- a/clang/include/clang/Support/Compiler.h
+++ b/clang/include/clang/Support/Compiler.h
@@ -14,8 +14,6 @@
#ifndef CLANG_SUPPORT_COMPILER_H
#define CLANG_SUPPORT_COMPILER_H
-#include "llvm/Support/Compiler.h"
-
/// CLANG_ABI is the main export/visibility macro to mark something as
/// explicitly exported when clang is built as a shared library with everything
/// else that is unannotated having hidden visibility.
diff --git a/clang/include/module.modulemap b/clang/include/module.modulemap
index b399f0beee59a1..e9218c715147b0 100644
--- a/clang/include/module.modulemap
+++ b/clang/include/module.modulemap
@@ -115,7 +115,7 @@ module Clang_Diagnostics {
module Driver { header "clang/Driver/DriverDiagnostic.h" export * }
module Frontend { header "clang/Frontend/FrontendDiagnostic.h" export * }
module Lex { header "clang/Lex/LexDiagnostic.h" export * }
- module Parse { header "clang/Parse/ParseDiagnostic.h" export * }
+ module Parse { header "clang/Basic/DiagnosticParse.h" export * }
module Serialization { header
"clang/Serialization/SerializationDiagnostic.h" export * }
module Refactoring { header
"clang/Tooling/Refactoring/RefactoringDiagnostic.h" export * }
}
@@ -186,6 +186,9 @@ module Clang_StaticAnalyzer_Frontend {
module Clang_Testing {
requires cplusplus
umbrella "clang/Testing"
+
+ textual header "clang/Testing/TestLanguage.def"
+
module * { export * }
}
diff --git a/llvm/include/llvm/DebugInfo/DWARF/DWARFTypePrinter.h
b/llvm/include/llvm/DebugInfo/DWARF/DWARFTypePrinter.h
index 3c936b93865045..bd25f6c30ebf1f 100644
--- a/llvm/include/llvm/DebugInfo/DWARF/DWARFTypePrinter.h
+++ b/llvm/include/llvm/DebugInfo/DWARF/DWARFTypePrinter.h
@@ -9,6 +9,7 @@
#ifndef LLVM_DEBUGINFO_DWARF_DWARFTYPEPRINTER_H
#define LLVM_DEBUGINFO_DWARF_DWARFTYPEPRINTER_H
+#include "llvm/ADT/StringExtras.h"
#include "llvm/ADT/StringRef.h"
#include "llvm/BinaryFormat/Dwarf.h"
#include "llvm/Support/Error.h"
diff --git a/llvm/include/llvm/IR/NVVMIntrinsicFlags.h
b/llvm/include/llvm/IR/NVVMIntrinsicFlags.h
index 43dde42bbbd620..dfb6e857b3a6ad 100644
--- a/llvm/include/llvm/IR/NVVMIntrinsicFlags.h
+++ b/llvm/include/llvm/IR/NVVMIntrinsicFlags.h
@@ -15,6 +15,8 @@
#ifndef LLVM_IR_NVVMINTRINSICFLAGS_H
#define LLVM_IR_NVVMINTRINSICFLAGS_H
+#include
+
namespace llvm {
namespace nvvm {
diff --git a/llvm/include/llvm/Support/Memory.h
b/llvm/include/llvm/Support/Memory.h
index c02a3cc14dc7de..a587f2a8542b2e 100644
--- a/llvm/include/llvm/Support/Memory.h
+++ b/llvm/include/llvm/Support/Memory.h
@@ -15,6 +15,7 @@
#include "llvm/Support/DataTypes.h"
#include
+#include
namespace llvm {
diff --git a/llvm/include/llvm/TargetParser/AArch64TargetParser.h
b/llvm/include/llvm/TargetParser/AArch64TargetParser.h
index 1311329821828f..11836e87d1b208 100644
--- a/llvm/include/llvm/TargetParser/AArch64TargetParser.h
+++ b/llvm/include/llvm/TargetParser/AArch64TargetParser.h
@@ -16,6 +16,7 @@
#include "llvm/ADT/ArrayRef.h"
#include "llvm/ADT/Bitset.h"
+#include "llvm/ADT/StringExtras.h"
#include "llvm/ADT/StringMap.h"
#include "llvm/ADT/StringRef.h"
#include "llvm/Support/VersionTuple.h"
diff --git a/llvm/include/module.modulemap b/llvm/include/module.modulemap
index b00da6d7cd28c7..ac7bd257cbb422 100644
--- a/llvm/include/module.modulemap
+++ b/llvm/include/module.modulemap
@@ -346,6 +346,7 @@ extern module LLVM_Extern_Utils_DataTypes
"module.extern.modulemap"
// TargetParser module before building the TargetParser module itself.
module TargetParserGen {
module AArch64TargetParserDef {
+textual header "llvm/TargetParser/AArch64CPUFeatures.inc"
header "llvm/TargetParser/AArch64TargetParser.h"
extern module LLVM_Extern_TargetParser_Gen "module.extern.modulemap"
export *
diff --git a/llvm/tools/llv
[clang-tools-extra] [clang-tidy]detecting conversion directly by `make_unique` and `make_shared` in bugprone-optional-value-conversion (PR #119371)
HerrCai0907 wrote: > Can the logic for implementing this also be used to address > https://github.com/llvm/llvm-project/issues/86447#issuecomment-2016943524? I think this kind model of stl can be reused in other check. but i have no idea to make it more generic. https://github.com/llvm/llvm-project/pull/119371 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [NVPTX] Add NVVMUpgradeAnnotations pass to cleanup legacy annotations (PR #119261)
@@ -5022,6 +5022,69 @@ bool llvm::UpgradeDebugInfo(Module &M) {
return Modified;
}
+bool static upgradeSingleNVVMAnnotation(GlobalValue *GV, StringRef K,
+const Metadata *V) {
+ if (K == "kernel") {
+assert(mdconst::extract(V)->getZExtValue() == 1);
+cast(GV)->addFnAttr("nvvm.kernel");
+return true;
+ }
+ if (K == "align") {
+const uint64_t AlignBits =
mdconst::extract(V)->getZExtValue();
+const unsigned Idx = (AlignBits >> 16);
+const Align StackAlign = Align(AlignBits & 0x);
AlexMaclean wrote:
Fixed
https://github.com/llvm/llvm-project/pull/119261
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [AArch64] Implement intrinsics for FP8 FCVT/FCVTN/BFCVT (PR #118025)
https://github.com/SpencerAbson updated
https://github.com/llvm/llvm-project/pull/118025
>From 65b749f27ddf7d7987423effe8c7b15b702c59f4 Mon Sep 17 00:00:00 2001
From: Spencer Abson
Date: Sun, 24 Nov 2024 18:38:36 +
Subject: [PATCH 1/2] [AArch64] Implement intrinsics for FP8 FCVT/FCVTN/BFCVT
---
clang/include/clang/Basic/arm_sve.td | 6 ++
.../fp8-intrinsics/acle_sme2_fp8_cvt.c| 64 +++
.../acle_sme2_fp8_cvt.c | 12 +++-
llvm/include/llvm/IR/IntrinsicsAArch64.td | 17 +
.../lib/Target/AArch64/AArch64SMEInstrInfo.td | 8 +--
llvm/lib/Target/AArch64/SMEInstrFormats.td| 15 -
.../AArch64/sme2-fp8-intrinsics-cvt.ll| 52 +++
7 files changed, 166 insertions(+), 8 deletions(-)
diff --git a/clang/include/clang/Basic/arm_sve.td
b/clang/include/clang/Basic/arm_sve.td
index 7b8ecf29a9de6e..30dc5c55d411ab 100644
--- a/clang/include/clang/Basic/arm_sve.td
+++ b/clang/include/clang/Basic/arm_sve.td
@@ -2436,6 +2436,12 @@ let SVETargetGuard = InvalidMode, SMETargetGuard =
"sme2,fp8" in {
// Convert from FP8 to deinterleaved half-precision/BFloat16 multi-vector
def SVF1CVTL_X2 : Inst<"svcvtl1_{d}[_mf8]_x2_fpm", "2~>", "bh", MergeNone,
"aarch64_sve_fp8_cvtl1_x2", [IsStreaming, SetsFPMR], []>;
def SVF2CVTL_X2 : Inst<"svcvtl2_{d}[_mf8]_x2_fpm", "2~>", "bh", MergeNone,
"aarch64_sve_fp8_cvtl2_x2", [IsStreaming, SetsFPMR], []>;
+
+ // Convert from single/half/bfloat multivector to FP8
+ def SVFCVT_X2 : Inst<"svcvt_mf8[_{d}_x2]_fpm", "~2>", "bh", MergeNone,
"aarch64_sve_fp8_cvt_x2", [IsStreaming, SetsFPMR], []>;
+ def SVFCVT_X4 : Inst<"svcvt_mf8[_{d}_x4]_fpm", "~4>", "f", MergeNone,
"aarch64_sve_fp8_cvt_x4", [IsOverloadNone, IsStreaming, SetsFPMR], []>;
+ // interleaved
+ def SVFCVTN_X4 : Inst<"svcvtn_mf8[_{d}_x4]_fpm", "~4>", "f", MergeNone,
"aarch64_sve_fp8_cvtn_x4", [IsOverloadNone, IsStreaming, SetsFPMR], []>;
}
let SVETargetGuard = "sve2p1", SMETargetGuard = "sme2" in {
diff --git a/clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sme2_fp8_cvt.c
b/clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sme2_fp8_cvt.c
index 13609f034da336..ae2e780f84cfe3 100644
--- a/clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sme2_fp8_cvt.c
+++ b/clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sme2_fp8_cvt.c
@@ -16,6 +16,70 @@
#define SVE_ACLE_FUNC(A1,A2,A3) A1##A2##A3
#endif
+// CHECK-LABEL: @test_cvt_f16_x2(
+// CHECK-NEXT: entry:
+// CHECK-NEXT:tail call void @llvm.aarch64.set.fpmr(i64 [[FPMR:%.*]])
+// CHECK-NEXT:[[TMP0:%.*]] = tail call
@llvm.aarch64.sve.fp8.cvt.x2.nxv8f16( [[ZN_COERCE0:%.*]],
[[ZN_COERCE1:%.*]])
+// CHECK-NEXT:ret [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_cvt_f16_x213svfloat16x2_tm(
+// CPP-CHECK-NEXT: entry:
+// CPP-CHECK-NEXT:tail call void @llvm.aarch64.set.fpmr(i64 [[FPMR:%.*]])
+// CPP-CHECK-NEXT:[[TMP0:%.*]] = tail call
@llvm.aarch64.sve.fp8.cvt.x2.nxv8f16( [[ZN_COERCE0:%.*]],
[[ZN_COERCE1:%.*]])
+// CPP-CHECK-NEXT:ret [[TMP0]]
+//
+svmfloat8_t test_cvt_f16_x2(svfloat16x2_t zn, fpm_t fpmr) __arm_streaming {
+ return SVE_ACLE_FUNC(svcvt_mf8,_f16_x2,_fpm)(zn, fpmr);
+}
+
+// CHECK-LABEL: @test_cvt_f32_x4(
+// CHECK-NEXT: entry:
+// CHECK-NEXT:tail call void @llvm.aarch64.set.fpmr(i64 [[FPMR:%.*]])
+// CHECK-NEXT:[[TMP0:%.*]] = tail call
@llvm.aarch64.sve.fp8.cvt.x4( [[ZN_COERCE0:%.*]], [[ZN_COERCE1:%.*]], [[ZN_COERCE2:%.*]], [[ZN_COERCE3:%.*]])
+// CHECK-NEXT:ret [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_cvt_f32_x413svfloat32x4_tm(
+// CPP-CHECK-NEXT: entry:
+// CPP-CHECK-NEXT:tail call void @llvm.aarch64.set.fpmr(i64 [[FPMR:%.*]])
+// CPP-CHECK-NEXT:[[TMP0:%.*]] = tail call
@llvm.aarch64.sve.fp8.cvt.x4( [[ZN_COERCE0:%.*]], [[ZN_COERCE1:%.*]], [[ZN_COERCE2:%.*]], [[ZN_COERCE3:%.*]])
+// CPP-CHECK-NEXT:ret [[TMP0]]
+//
+svmfloat8_t test_cvt_f32_x4(svfloat32x4_t zn, fpm_t fpmr) __arm_streaming {
+ return SVE_ACLE_FUNC(svcvt_mf8,_f32_x4,_fpm)(zn, fpmr);
+}
+
+// CHECK-LABEL: @test_cvtn_f32_x4(
+// CHECK-NEXT: entry:
+// CHECK-NEXT:tail call void @llvm.aarch64.set.fpmr(i64 [[FPMR:%.*]])
+// CHECK-NEXT:[[TMP0:%.*]] = tail call
@llvm.aarch64.sve.fp8.cvtn.x4( [[ZN_COERCE0:%.*]], [[ZN_COERCE1:%.*]], [[ZN_COERCE2:%.*]],
[[ZN_COERCE3:%.*]])
+// CHECK-NEXT:ret [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_cvtn_f32_x413svfloat32x4_tm(
+// CPP-CHECK-NEXT: entry:
+// CPP-CHECK-NEXT:tail call void @llvm.aarch64.set.fpmr(i64 [[FPMR:%.*]])
+// CPP-CHECK-NEXT:[[TMP0:%.*]] = tail call
@llvm.aarch64.sve.fp8.cvtn.x4( [[ZN_COERCE0:%.*]], [[ZN_COERCE1:%.*]], [[ZN_COERCE2:%.*]],
[[ZN_COERCE3:%.*]])
+// CPP-CHECK-NEXT:ret [[TMP0]]
+//
+svmfloat8_t test_cvtn_f32_x4(svfloat32x4_t zn, fpm_t fpmr) __arm_streaming {
+ return SVE_ACLE_FUNC(svcvtn_mf8,_f32_x4,_fpm)(zn, fpmr);
+}
+
+// CHECK-LABEL: @test_cvt_bf16_x2(
+// CHECK-NEXT: entry:
+// CHECK-NEXT:tail call void @llvm.aarch64.set.fpmr(i64
[clang] [llvm] [clang][Modules] Fixing Build Breaks When -DLLVM_ENABLE_MODULES=ON (PR #119473)
@@ -14,8 +14,6 @@
#ifndef CLANG_SUPPORT_COMPILER_H
#define CLANG_SUPPORT_COMPILER_H
-#include "llvm/Support/Compiler.h"
Bigcheese wrote:
The include of this file needs to move to Attr.h, module imports can't be in a
namespace, so putting the include in Attrs.inc is wrong.
You should also add a new top level module for the Support directory.
`module Clang_Support { requires cplusplus umbrella "clang/Support" module * {
export * } }`
https://github.com/llvm/llvm-project/pull/119473
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [Clang] Match MSVC handling of duplicate header search paths in Microsoft compatibility modes. (PR #105738)
https://github.com/tahonermann updated https://github.com/llvm/llvm-project/pull/105738 >From d5c7035fd0c007e9833150a3d1b0a86d8744aa5d Mon Sep 17 00:00:00 2001 From: Tom Honermann Date: Thu, 22 Aug 2024 09:44:56 -0700 Subject: [PATCH 1/2] [Clang] Match MSVC handling of duplicate header search paths in Microsoft compatibility modes. Clang has historically matched GCC's behavior for header search path order and pruning of duplicate paths. That traditional behavior is to order user search paths before system search paths, to ignore user search paths that duplicate a (later) system search path, and to ignore search paths that duplicate an earlier search path of the same user/system kind. This differs from MSVC and can result in inconsistent header file resolution for `#include` directives. MSVC orders header search paths as follows: 1) Paths specified by the `/I` and `/external:I` options are processed in the order that they appear. Paths specified by `/I` that duplicate a path specified by `/external:I` are ignored regardless of the order of the options. Paths specified by `/I` that duplicate a path from a prior `/I` option are ignored. Paths specified by `/external:I` that duplicate a path from a later `/external:I` option are ignored. 2) Paths specified by `/external:env` are processed in the order that they appear. Paths that duplicate a path from a `/I` or `/external:I` option are ignored regardless of the order of the options. Paths that duplicate a path from a prior `/external:env` option or an earlier path from the same `/external:env` option are ignored. 3) Paths specified by the `INCLUDE` environment variable are processed in the order they appear. Paths that duplicate a path from a `/I`, `/external:I`, or `/external:env` option are ignored. Paths that duplicate an earlier path in the `INCLUDE` environment variable are ignored. 4) Paths specified by the `EXTERNAL_INCLUDE` environment variable are processed in the order they appear. Paths that duplicate a path from a `/I`, `/external:I`, or `/external:env` option are ignored. Paths that duplicate a path from the `INCLUDE` environment variable are ignored. Paths that duplicate an earlier path in the `EXTERNAL_INCLUDE environment variable are ignored. Prior to this change, Clang handled the `/external:I` and `/external:env` options and the paths present in the `INCLUDE` and `EXTERNAL_INCLUDE` environment variables as though they were specified with the `-isystem` option. The GCC behavior described above then lead to a command line such as `/external:I dir1 /Idir2` having a header search order of `dir2` followed by `dir1`; contrary to MSVC behavior. This change adds support for the MSVC external path concept for both the `clang` and `clang-cl` drivers with the following option syntax. These options match the MSVC behavior described above for both drivers. clangclang-cl --- -iexternal /external:I -iexternal-env= /external:env: Paths specified by these options are still treated as system paths. That is, whether warnings are issued in header files found via these paths remains subject to use of the `-Wsystem-headers` and `-Wno-system-headers` options. In the future, it would make sense to add a separate option that matches the MSVC `/external:Wn` option to control such warnings. The MSVC behavior described above implies that (system) paths present in the `INCLUDE` and `EXTERNAL_INCLUDE` environment variables do not suppress matching user paths specified via `/I`. This contrasts with GCC's behavior of suppressing user paths that match a system path regardless of how each is specified. Since the `clang-cl` driver maps paths from the `INCLUDE` and `EXTERNAL_INCLUDE` environment variable to `-internal-isystem`, matching MSVC behavior requires suppressing that aspect of the GCC behavior. With this change, system paths will no longer suppress user paths when the `-fms-compatibility` option is explicitly or implicitly enabled. This will affect header search path ordering for options like `-isystem` when duplicate user paths are present. Should motivation arise for preserving such suppression of user paths when compiling with `-fms-compatibility` enabled, it would make sense to introduce a new option for the `clang-cl` driver to map paths in these environment variabless to that would match MSVC behavior without impacting other system path options. --- clang/docs/ReleaseNotes.rst | 38 ++ clang/include/clang/Driver/Options.td | 17 +- clang/include/clang/Driver/ToolChain.h| 11 +- clang/include/clang/Lex/HeaderSearchOptions.h | 10 + clang/lib/Driver/Driver.cpp | 4 +- clang/lib/Driver/ToolChain.cpp| 45 +++ clang/lib/Driver/ToolChains/Clang.cpp | 15 +- clang/lib/Driver/ToolChai
[clang] Fix a cmake error when using the Xcode generator. (PR #119403)
https://github.com/paulhuggett edited https://github.com/llvm/llvm-project/pull/119403 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [clang] Fix missing check for nullptr in CallExpr::getUnusedResultAttr (PR #118636)
https://github.com/cor3ntin approved this pull request. LGTM https://github.com/llvm/llvm-project/pull/118636 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [AArch64] Add Neon FP8 conversion intrinsics (PR #119033)
https://github.com/momchil-velikov edited https://github.com/llvm/llvm-project/pull/119033 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [AArch64] Add Neon FP8 conversion intrinsics (PR #119033)
https://github.com/momchil-velikov updated
https://github.com/llvm/llvm-project/pull/119033
>From a83b190b25f01843922530d9e409cfb9c0a86c18 Mon Sep 17 00:00:00 2001
From: Momchil Velikov
Date: Fri, 6 Dec 2024 13:09:23 +
Subject: [PATCH 1/4] [AArch64] Refactor implementation of FP8 types (NFC)
* The FP8 scalar type (`__mfp8`) was described as a vector type
* The FP8 vector types were described/assumed to have
integer element type (the element type ought to be `__mfp8`),
* Add support for `m` type specifier (denoting `__mfp8`)
in `DecodeTypeFromStr` and create SVE builtin prototypes using
the specifier, instead of `int8_t`.
---
clang/include/clang/AST/Type.h| 5 +++
.../clang/Basic/AArch64SVEACLETypes.def | 24 +---
clang/lib/AST/ASTContext.cpp | 37 +++
clang/lib/AST/ItaniumMangle.cpp | 5 +++
clang/lib/AST/Type.cpp| 4 +-
clang/lib/CodeGen/CodeGenTypes.cpp| 13 +--
clang/lib/CodeGen/Targets/AArch64.cpp | 7 +++-
clang/utils/TableGen/SveEmitter.cpp | 4 +-
8 files changed, 76 insertions(+), 23 deletions(-)
diff --git a/clang/include/clang/AST/Type.h b/clang/include/clang/AST/Type.h
index 6fd6c73a516f08..626cfea56a47db 100644
--- a/clang/include/clang/AST/Type.h
+++ b/clang/include/clang/AST/Type.h
@@ -2518,6 +2518,7 @@ class alignas(TypeAlignment) Type : public
ExtQualsTypeCommonBase {
bool isFloat32Type() const;
bool isDoubleType() const;
bool isBFloat16Type() const;
+ bool isMFloat8Type() const;
bool isFloat128Type() const;
bool isIbm128Type() const;
bool isRealType() const; // C99 6.2.5p17 (real floating + integer)
@@ -8532,6 +8533,10 @@ inline bool Type::isBFloat16Type() const {
return isSpecificBuiltinType(BuiltinType::BFloat16);
}
+inline bool Type::isMFloat8Type() const {
+ return isSpecificBuiltinType(BuiltinType::MFloat8);
+}
+
inline bool Type::isFloat128Type() const {
return isSpecificBuiltinType(BuiltinType::Float128);
}
diff --git a/clang/include/clang/Basic/AArch64SVEACLETypes.def
b/clang/include/clang/Basic/AArch64SVEACLETypes.def
index 063cac1f4a58ee..6b704b386536c9 100644
--- a/clang/include/clang/Basic/AArch64SVEACLETypes.def
+++ b/clang/include/clang/Basic/AArch64SVEACLETypes.def
@@ -57,6 +57,11 @@
// - IsBF true for vector of brain float elements.
//===--===//
+#ifndef SVE_SCALAR_TYPE
+#define SVE_SCALAR_TYPE(Name, MangledName, Id, SingletonId, Bits) \
+ SVE_TYPE(Name, Id, SingletonId)
+#endif
+
#ifndef SVE_VECTOR_TYPE
#define SVE_VECTOR_TYPE(Name, MangledName, Id, SingletonId) \
SVE_TYPE(Name, Id, SingletonId)
@@ -72,6 +77,11 @@
SVE_VECTOR_TYPE_DETAILS(Name, MangledName, Id, SingletonId, NumEls, ElBits,
NF, false, false, true)
#endif
+#ifndef SVE_VECTOR_TYPE_MFLOAT
+#define SVE_VECTOR_TYPE_MFLOAT(Name, MangledName, Id, SingletonId, NumEls,
ElBits, NF) \
+ SVE_VECTOR_TYPE_DETAILS(Name, MangledName, Id, SingletonId, NumEls, ElBits,
NF, false, false, false)
+#endif
+
#ifndef SVE_VECTOR_TYPE_FLOAT
#define SVE_VECTOR_TYPE_FLOAT(Name, MangledName, Id, SingletonId, NumEls,
ElBits, NF) \
SVE_VECTOR_TYPE_DETAILS(Name, MangledName, Id, SingletonId, NumEls, ElBits,
NF, false, true, false)
@@ -125,8 +135,7 @@ SVE_VECTOR_TYPE_FLOAT("__SVFloat64_t", "__SVFloat64_t",
SveFloat64, SveFloat64Ty
SVE_VECTOR_TYPE_BFLOAT("__SVBfloat16_t", "__SVBfloat16_t", SveBFloat16,
SveBFloat16Ty, 8, 16, 1)
-// This is a 8 bits opaque type.
-SVE_VECTOR_TYPE_INT("__SVMfloat8_t", "__SVMfloat8_t", SveMFloat8,
SveMFloat8Ty, 16, 8, 1, false)
+SVE_VECTOR_TYPE_MFLOAT("__SVMfloat8_t", "__SVMfloat8_t", SveMFloat8,
SveMFloat8Ty, 16, 8, 1)
//
// x2
@@ -148,7 +157,7 @@ SVE_VECTOR_TYPE_FLOAT("__clang_svfloat64x2_t",
"svfloat64x2_t", SveFloat64x2, Sv
SVE_VECTOR_TYPE_BFLOAT("__clang_svbfloat16x2_t", "svbfloat16x2_t",
SveBFloat16x2, SveBFloat16x2Ty, 8, 16, 2)
-SVE_VECTOR_TYPE_INT("__clang_svmfloat8x2_t", "svmfloat8x2_t", SveMFloat8x2,
SveMFloat8x2Ty, 16, 8, 2, false)
+SVE_VECTOR_TYPE_MFLOAT("__clang_svmfloat8x2_t", "svmfloat8x2_t", SveMFloat8x2,
SveMFloat8x2Ty, 16, 8, 2)
//
// x3
@@ -170,7 +179,7 @@ SVE_VECTOR_TYPE_FLOAT("__clang_svfloat64x3_t",
"svfloat64x3_t", SveFloat64x3, Sv
SVE_VECTOR_TYPE_BFLOAT("__clang_svbfloat16x3_t", "svbfloat16x3_t",
SveBFloat16x3, SveBFloat16x3Ty, 8, 16, 3)
-SVE_VECTOR_TYPE_INT("__clang_svmfloat8x3_t", "svmfloat8x3_t", SveMFloat8x3,
SveMFloat8x3Ty, 16, 8, 3, false)
+SVE_VECTOR_TYPE_MFLOAT("__clang_svmfloat8x3_t", "svmfloat8x3_t", SveMFloat8x3,
SveMFloat8x3Ty, 16, 8, 3)
//
// x4
@@ -192,7 +201,7 @@ SVE_VECTOR_TYPE_FLOAT("__clang_svfloat64x4_t",
"svfloat64x4_t", SveFloat64x4, Sv
SVE_VECTOR_TYPE_BFLOAT("__clang_svbfloat16x4_t", "svbfloat16x4_t",
SveBFloat16x4, SveBFloat16x4Ty, 8, 16, 4)
-SVE_VECTOR_TYPE_INT("__clang_svmfloat8x4_t", "svmfloat8x4_t", SveMFloat8x4,
SveMFloat8x4T
[clang] [clang-format] Fix crashes when the macro expansion is empty (PR #119428)
llvmbot wrote:
@llvm/pr-subscribers-clang-format
Author: 天音あめ (amane-ame)
Changes
An empty expansion should be valid, like `echo 'A()' | clang-format
--style='{Macros: [A(x)=x]}'`
See #119258.
---
Full diff: https://github.com/llvm/llvm-project/pull/119428.diff
1 Files Affected:
- (modified) clang/lib/Format/MacroExpander.cpp (+4)
``diff
diff --git a/clang/lib/Format/MacroExpander.cpp
b/clang/lib/Format/MacroExpander.cpp
index fd2a16894d643d..ed9e51dfbfef1f 100644
--- a/clang/lib/Format/MacroExpander.cpp
+++ b/clang/lib/Format/MacroExpander.cpp
@@ -233,6 +233,10 @@ MacroExpander::expand(FormatToken *ID,
if (Result.size() > 1) {
++Result[0]->MacroCtx->StartOfExpansion;
++Result[Result.size() - 2]->MacroCtx->EndOfExpansion;
+ } else {
+// If the macro expansion is empty, mark the start and end
+Result[0]->MacroCtx->StartOfExpansion = 1;
+Result[0]->MacroCtx->EndOfExpansion = 1;
}
return Result;
}
``
https://github.com/llvm/llvm-project/pull/119428
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] 1a5e18a - [HLSL] Do not print details in IR for target extension types (#115971)
Author: Joshua Batista
Date: 2024-12-10T10:07:30-08:00
New Revision: 1a5e18a492481ff48ca764c8f7d08d231d922382
URL:
https://github.com/llvm/llvm-project/commit/1a5e18a492481ff48ca764c8f7d08d231d922382
DIFF:
https://github.com/llvm/llvm-project/commit/1a5e18a492481ff48ca764c8f7d08d231d922382.diff
LOG: [HLSL] Do not print details in IR for target extension types (#115971)
This PR changes how target extension types are printed when they are
emitted as IR.
This prevents repetitive phrases like "struct = type {...}" from being
repeated over and over in the outputted IR.
Additionally, it should allow opt to not crash when parsing the DXIL
output.
Fixes [#114131](https://github.com/llvm/llvm-project/issues/114131)
Added:
Modified:
clang/test/CodeGenHLSL/builtins/AppendStructuredBuffer-elementtype.hlsl
clang/test/CodeGenHLSL/builtins/ConsumeStructuredBuffer-elementtype.hlsl
clang/test/CodeGenHLSL/builtins/RasterizerOrderedStructuredBuffer-elementtype.hlsl
clang/test/CodeGenHLSL/builtins/hlsl_resource_t.hlsl
clang/test/CodeGenHLSL/resource-bindings.hlsl
llvm/lib/IR/AsmWriter.cpp
llvm/unittests/IR/TypesTest.cpp
Removed:
diff --git
a/clang/test/CodeGenHLSL/builtins/AppendStructuredBuffer-elementtype.hlsl
b/clang/test/CodeGenHLSL/builtins/AppendStructuredBuffer-elementtype.hlsl
index 1e8aae588fc33d..85face8eaeb6c8 100644
--- a/clang/test/CodeGenHLSL/builtins/AppendStructuredBuffer-elementtype.hlsl
+++ b/clang/test/CodeGenHLSL/builtins/AppendStructuredBuffer-elementtype.hlsl
@@ -18,7 +18,8 @@ struct MyStruct {
// DXIL: %"class.hlsl::AppendStructuredBuffer.9" = type {
target("dx.RawBuffer", <3 x i32>, 1, 0)
// DXIL: %"class.hlsl::AppendStructuredBuffer.10" = type {
target("dx.RawBuffer", <2 x half>, 1, 0)
// DXIL: %"class.hlsl::AppendStructuredBuffer.11" = type {
target("dx.RawBuffer", <3 x float>, 1, 0)
-// DXIL: %"class.hlsl::AppendStructuredBuffer.12" = type {
target("dx.RawBuffer", %struct.MyStruct = type { <4 x float>, <2 x i32>, [8 x
i8] }, 1, 0)
+// DXIL: %"class.hlsl::AppendStructuredBuffer.12" = type {
target("dx.RawBuffer", %struct.MyStruct, 1, 0)
+// DXIL: %struct.MyStruct = type { <4 x float>, <2 x i32>, [8 x i8] }
AppendStructuredBuffer BufI16;
AppendStructuredBuffer BufU16;
diff --git
a/clang/test/CodeGenHLSL/builtins/ConsumeStructuredBuffer-elementtype.hlsl
b/clang/test/CodeGenHLSL/builtins/ConsumeStructuredBuffer-elementtype.hlsl
index f8574c6460d4e1..5ed9e9ad8160ff 100644
--- a/clang/test/CodeGenHLSL/builtins/ConsumeStructuredBuffer-elementtype.hlsl
+++ b/clang/test/CodeGenHLSL/builtins/ConsumeStructuredBuffer-elementtype.hlsl
@@ -18,7 +18,8 @@ struct MyStruct {
// DXIL: %"class.hlsl::ConsumeStructuredBuffer.9" = type {
target("dx.RawBuffer", <3 x i32>, 1, 0)
// DXIL: %"class.hlsl::ConsumeStructuredBuffer.10" = type {
target("dx.RawBuffer", <2 x half>, 1, 0)
// DXIL: %"class.hlsl::ConsumeStructuredBuffer.11" = type {
target("dx.RawBuffer", <3 x float>, 1, 0)
-// DXIL: %"class.hlsl::ConsumeStructuredBuffer.12" = type {
target("dx.RawBuffer", %struct.MyStruct = type { <4 x float>, <2 x i32>, [8 x
i8] }, 1, 0)
+// DXIL: %"class.hlsl::ConsumeStructuredBuffer.12" = type {
target("dx.RawBuffer", %struct.MyStruct, 1, 0)
+// DXIL: %struct.MyStruct = type { <4 x float>, <2 x i32>, [8 x i8] }
ConsumeStructuredBuffer BufI16;
ConsumeStructuredBuffer BufU16;
diff --git
a/clang/test/CodeGenHLSL/builtins/RasterizerOrderedStructuredBuffer-elementtype.hlsl
b/clang/test/CodeGenHLSL/builtins/RasterizerOrderedStructuredBuffer-elementtype.hlsl
index 7057988de24b34..68d626de689f2e 100644
---
a/clang/test/CodeGenHLSL/builtins/RasterizerOrderedStructuredBuffer-elementtype.hlsl
+++
b/clang/test/CodeGenHLSL/builtins/RasterizerOrderedStructuredBuffer-elementtype.hlsl
@@ -18,7 +18,7 @@ struct MyStruct {
// DXIL: %"class.hlsl::RasterizerOrderedStructuredBuffer.9" = type {
target("dx.RawBuffer", <3 x i32>, 1, 1) }
// DXIL: %"class.hlsl::RasterizerOrderedStructuredBuffer.10" = type {
target("dx.RawBuffer", <2 x half>, 1, 1) }
// DXIL: %"class.hlsl::RasterizerOrderedStructuredBuffer.11" = type {
target("dx.RawBuffer", <3 x float>, 1, 1) }
-// DXIL: %"class.hlsl::RasterizerOrderedStructuredBuffer.12" = type {
target("dx.RawBuffer", %struct.MyStruct = type { <4 x float>, <2 x i32>, [8 x
i8] }, 1, 1) }
+// DXIL: %struct.MyStruct = type { <4 x float>, <2 x i32>, [8 x i8] }
RasterizerOrderedStructuredBuffer BufI16;
RasterizerOrderedStructuredBuffer BufU16;
diff --git a/clang/test/CodeGenHLSL/builtins/hlsl_resource_t.hlsl
b/clang/test/CodeGenHLSL/builtins/hlsl_resource_t.hlsl
index 7de62a363eedb5..fda034ce7f2034 100644
--- a/clang/test/CodeGenHLSL/builtins/hlsl_resource_t.hlsl
+++ b/clang/test/CodeGenHLSL/builtins/hlsl_resource_t.hlsl
@@ -3,7 +3,8 @@
using handle_float_t = __hlsl_resource_t [[hlsl::resource_class(UAV)]]
[[hlsl::contained_type(float)]];
[clang] [clang-format] Fix crashes when the macro expansion is empty (PR #119428)
https://github.com/amane-ame created
https://github.com/llvm/llvm-project/pull/119428
An empty expansion should be valid, like `echo 'A()' | clang-format
--style='{Macros: [A(x)=x]}'`
See #119258.
From 3a4c1a924faef3a7a09126694fcb943bd7083451 Mon Sep 17 00:00:00 2001
From: amane-ame
Date: Wed, 11 Dec 2024 02:13:43 +0800
Subject: [PATCH] Fix crashes when the macro expansion is empty
---
clang/lib/Format/MacroExpander.cpp | 4
1 file changed, 4 insertions(+)
diff --git a/clang/lib/Format/MacroExpander.cpp
b/clang/lib/Format/MacroExpander.cpp
index fd2a16894d643d..ed9e51dfbfef1f 100644
--- a/clang/lib/Format/MacroExpander.cpp
+++ b/clang/lib/Format/MacroExpander.cpp
@@ -233,6 +233,10 @@ MacroExpander::expand(FormatToken *ID,
if (Result.size() > 1) {
++Result[0]->MacroCtx->StartOfExpansion;
++Result[Result.size() - 2]->MacroCtx->EndOfExpansion;
+ } else {
+// If the macro expansion is empty, mark the start and end
+Result[0]->MacroCtx->StartOfExpansion = 1;
+Result[0]->MacroCtx->EndOfExpansion = 1;
}
return Result;
}
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [AArch64] Add Neon FP8 conversion intrinsics (PR #119033)
momchil-velikov wrote: - added the tests - fixed calls to the wrong LLVM intrinsic https://github.com/llvm/llvm-project/pull/119033 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [AArch64] Implement FP8 SVE Intrinsics for narrowing conversions (PR #118124)
https://github.com/momchil-velikov updated
https://github.com/llvm/llvm-project/pull/118124
>From 6f268d4a80a8994855c99679ce1c66c11be8e357 Mon Sep 17 00:00:00 2001
From: Momchil Velikov
Date: Mon, 25 Nov 2024 09:47:41 +
Subject: [PATCH] [AArch64] Implement FP8 SVE Intrinsics for narrowing
conversions
* Half-precision and BFloat16 convert, narrow, and interleave to 8-bit
floating-point.
// Variant is also available for: _bf16_x2
svmfloat8_t svcvtn_mf8[_f16_x2]_fpm(svfloat16x2_t zn, fpm_t fpm);
* Single-precision convert, narrow, and interleave to 8-bit floating-point (top
and bottom).
svmfloat8_t svcvtnt_mf8[_f32_x2]_fpm(svmfloat8_t zd, svfloat32x2_t zn, fpm_t
fpm);
svmfloat8_t svcvtnb_mf8[_f32_x2]_fpm(svfloat32x2_t zn, fpm_t fpm);
---
clang/include/clang/Basic/arm_sve.td | 7 ++
.../fp8-intrinsics/acle_sve2_fp8_cvtn.c | 101 ++
.../aarch64-sve2-intrinsics/acle_sve2_fp8.c | 11 +-
llvm/include/llvm/IR/IntrinsicsAArch64.td | 14 +++
.../lib/Target/AArch64/AArch64SVEInstrInfo.td | 9 +-
llvm/lib/Target/AArch64/SVEInstrFormats.td| 37 ++-
llvm/test/CodeGen/AArch64/fp8-sve-cvtn.ll | 49 +
7 files changed, 222 insertions(+), 6 deletions(-)
create mode 100644
clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sve2_fp8_cvtn.c
create mode 100644 llvm/test/CodeGen/AArch64/fp8-sve-cvtn.ll
diff --git a/clang/include/clang/Basic/arm_sve.td
b/clang/include/clang/Basic/arm_sve.td
index 7b8ecf29a9de6e..d467720fc5c61f 100644
--- a/clang/include/clang/Basic/arm_sve.td
+++ b/clang/include/clang/Basic/arm_sve.td
@@ -2462,4 +2462,11 @@ let SVETargetGuard = "sve2,fp8", SMETargetGuard =
"sme2,fp8" in {
// 8-bit floating-point convert to BFloat16/Float16 (top)
def SVF1CVTLT : SInst<"svcvtlt1_{d}[_mf8]_fpm", "d~>", "bh", MergeNone,
"aarch64_sve_fp8_cvtlt1", [VerifyRuntimeMode, SetsFPMR]>;
def SVF2CVTLT : SInst<"svcvtlt2_{d}[_mf8]_fpm", "d~>", "bh", MergeNone,
"aarch64_sve_fp8_cvtlt2", [VerifyRuntimeMode, SetsFPMR]>;
+
+ // BFloat16/Float16 convert, narrow and interleave to 8-bit floating-point
+ def SVFCVTN : SInst<"svcvtn_mf8[_{d}_x2]_fpm", "~2>", "bh", MergeNone,
"aarch64_sve_fp8_cvtn", [VerifyRuntimeMode, SetsFPMR]>;
+
+ // Single-precision convert, narrow and interleave to 8-bit floating-point
(top and bottom)
+ def SVFCVTNB : SInst<"svcvtnb_mf8[_f32_x2]_fpm", "~2>", "f", MergeNone,
"aarch64_sve_fp8_cvtnb", [VerifyRuntimeMode, SetsFPMR]>;
+ def SVFCVTNT : SInst<"svcvtnt_mf8[_f32_x2]_fpm", "~~2>", "f", MergeNone,
"aarch64_sve_fp8_cvtnt", [VerifyRuntimeMode, SetsFPMR]>;
}
diff --git a/clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sve2_fp8_cvtn.c
b/clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sve2_fp8_cvtn.c
new file mode 100644
index 00..ed5b0ce02af4bd
--- /dev/null
+++ b/clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sve2_fp8_cvtn.c
@@ -0,0 +1,101 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
UTC_ARGS: --version 5
+// RUN: %clang_cc1-triple aarch64-none-linux-gnu -target-feature +sve
-target-feature +sve2 -target-feature +fp8 -disable-O0-optnone -Werror -Wall
-emit-llvm -o - %s | opt -S -p mem2reg,instcombine,tailcallelim | FileCheck %s
+// RUN: %clang_cc1 -x c++ -triple aarch64-none-linux-gnu -target-feature +sme
-target-feature +sme2 -target-feature +fp8 -disable-O0-optnone -Werror -Wall
-emit-llvm -o - %s | opt -S -p mem2reg,instcombine,tailcallelim | FileCheck %s
-check-prefix=CHECK-CXX
+
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS-triple
aarch64-none-linux-gnu -target-feature +sme -target-feature +sme2
-target-feature +fp8 -disable-O0-optnone -Werror -Wall -emit-llvm -o - %s | opt
-S -p mem2reg,instcombine,tailcallelim | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -x c++ -triple
aarch64-none-linux-gnu -target-feature +sve -target-feature +sve2
-target-feature +fp8 -disable-O0-optnone -Werror -Wall -emit-llvm -o - %s | opt
-S -p mem2reg,instcombine,tailcallelim | FileCheck %s -check-prefix=CHECK-CXX
+
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve
-target-feature +sve2 -target-feature +fp8 -S -disable-O0-optnone -Werror -Wall
-o /dev/null %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sme
-target-feature +sme2 -target-feature +fp8 -S -disable-O0-optnone -Werror -Wall
-o /dev/null %s
+
+// REQUIRES: aarch64-registered-target
+
+#ifdef __ARM_FEATURE_SME
+#include
+#else
+#include
+#endif
+
+#ifdef SVE_OVERLOADED_FORMS
+#define SVE_ACLE_FUNC(A1,A2_UNUSED,A3) A1##A3
+#else
+#define SVE_ACLE_FUNC(A1,A2,A3) A1##A2##A3
+#endif
+
+#ifdef __ARM_FEATURE_SME
+#define STREAMING __arm_streaming
+#else
+#define STREAMING
+#endif
+
+// CHECK-LABEL: define dso_local @test_svcvtn_f8_bf16(
+// CHECK-SAME: [[ZN_ZM_COERCE0:%.*]], [[ZN_ZM_COERCE1:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT:tail call void @llvm.aarch64.set.fpmr(i6
[clang] [llvm] Support for dispatch construct (Sema & Codegen) support. (PR #117904)
https://github.com/SunilKuravinakop updated
https://github.com/llvm/llvm-project/pull/117904
>From 1703aa62cfe35538aedbacb28e907535e838248c Mon Sep 17 00:00:00 2001
From: Sunil Kuravinakop
Date: Fri, 20 Sep 2024 01:41:29 -0500
Subject: [PATCH 1/3] Support for dispatch construct (Sema & Codegen) support.
Support for clauses depend, novariants & nocontext.
---
.../clang/Basic/DiagnosticSemaKinds.td| 3 +
clang/include/clang/Basic/OpenMPKinds.h | 6 +
clang/include/clang/Sema/SemaOpenMP.h | 7 +
clang/lib/Basic/OpenMPKinds.cpp | 5 +
clang/lib/CodeGen/CGStmt.cpp | 2 +-
clang/lib/CodeGen/CGStmtOpenMP.cpp| 4 +
clang/lib/CodeGen/CodeGenFunction.h | 1 +
clang/lib/Sema/SemaOpenMP.cpp | 301 ++-
clang/test/OpenMP/dispatch_codegen.cpp| 359 ++
clang/test/OpenMP/dispatch_unsupported.c | 7 -
.../include/llvm/Frontend/OpenMP/OMPContext.h | 6 +
11 files changed, 687 insertions(+), 14 deletions(-)
create mode 100644 clang/test/OpenMP/dispatch_codegen.cpp
delete mode 100644 clang/test/OpenMP/dispatch_unsupported.c
diff --git a/clang/include/clang/Basic/DiagnosticSemaKinds.td
b/clang/include/clang/Basic/DiagnosticSemaKinds.td
index 8495884dcd058f..81b876f9fd85c5 100644
--- a/clang/include/clang/Basic/DiagnosticSemaKinds.td
+++ b/clang/include/clang/Basic/DiagnosticSemaKinds.td
@@ -11774,6 +11774,9 @@ def err_omp_clause_requires_dispatch_construct : Error<
"'%0' clause requires 'dispatch' context selector">;
def err_omp_append_args_with_varargs : Error<
"'append_args' is not allowed with varargs functions">;
+def warn_omp_dispatch_clause_novariants_nocontext : Warning<
+ "only 'novariants' clause is supported when 'novariants' & 'nocontext'
clauses occur on the same dispatch construct">,
+ InGroup;
def err_openmp_vla_in_task_untied : Error<
"variable length arrays are not supported in OpenMP tasking regions with
'untied' clause">;
def warn_omp_unterminated_declare_target : Warning<
diff --git a/clang/include/clang/Basic/OpenMPKinds.h
b/clang/include/clang/Basic/OpenMPKinds.h
index 900ad6ca6d66f6..7579fab43dbb19 100644
--- a/clang/include/clang/Basic/OpenMPKinds.h
+++ b/clang/include/clang/Basic/OpenMPKinds.h
@@ -269,6 +269,12 @@ bool isOpenMPTaskLoopDirective(OpenMPDirectiveKind DKind);
/// parallel', otherwise - false.
bool isOpenMPParallelDirective(OpenMPDirectiveKind DKind);
+/// Checks if the specified directive is a dispatch-kind directive.
+/// \param DKind Specified directive.
+/// \return true - the directive is a dispatch-like directive like 'omp
+/// dispatch', otherwise - false.
+bool isOpenMPDispatchDirective(OpenMPDirectiveKind DKind);
+
/// Checks if the specified directive is a target code offload directive.
/// \param DKind Specified directive.
/// \return true - the directive is a target code offload directive like
diff --git a/clang/include/clang/Sema/SemaOpenMP.h
b/clang/include/clang/Sema/SemaOpenMP.h
index 3d1cc4fab1c10f..80cee9e7583051 100644
--- a/clang/include/clang/Sema/SemaOpenMP.h
+++ b/clang/include/clang/Sema/SemaOpenMP.h
@@ -1462,6 +1462,13 @@ class SemaOpenMP : public SemaBase {
: OMPDeclareVariantScopes.back().TI;
}
+ StmtResult transformDispatchDirective(OpenMPDirectiveKind Kind,
+const DeclarationNameInfo &DirName,
+OpenMPDirectiveKind CancelRegion,
+ArrayRef Clauses,
+Stmt *AStmt, SourceLocation StartLoc,
+SourceLocation EndLoc);
+
/// The current `omp begin/end declare variant` scopes.
SmallVector OMPDeclareVariantScopes;
diff --git a/clang/lib/Basic/OpenMPKinds.cpp b/clang/lib/Basic/OpenMPKinds.cpp
index 62a13f01481b28..44ee63df46adb5 100644
--- a/clang/lib/Basic/OpenMPKinds.cpp
+++ b/clang/lib/Basic/OpenMPKinds.cpp
@@ -621,6 +621,11 @@ bool clang::isOpenMPParallelDirective(OpenMPDirectiveKind
DKind) {
llvm::is_contained(getLeafConstructs(DKind), OMPD_parallel);
}
+bool clang::isOpenMPDispatchDirective(OpenMPDirectiveKind DKind) {
+ return DKind == OMPD_dispatch ||
+ llvm::is_contained(getLeafConstructs(DKind), OMPD_target);
+}
+
bool clang::isOpenMPTargetExecutionDirective(OpenMPDirectiveKind DKind) {
return DKind == OMPD_target ||
llvm::is_contained(getLeafConstructs(DKind), OMPD_target);
diff --git a/clang/lib/CodeGen/CGStmt.cpp b/clang/lib/CodeGen/CGStmt.cpp
index 698baf853507f4..f3eedb79844378 100644
--- a/clang/lib/CodeGen/CGStmt.cpp
+++ b/clang/lib/CodeGen/CGStmt.cpp
@@ -417,7 +417,7 @@ void CodeGenFunction::EmitStmt(const Stmt *S,
ArrayRef Attrs) {
EmitOMPInteropDirective(cast(*S));
break;
case Stmt::OMPDispatchDirectiveClass:
-CGM.ErrorUnsupported(S, "OpenMP dispatch directive");
+
[clang] [clang-format] Fix crashes when the macro expansion is empty (PR #119428)
github-actions[bot] wrote: Thank you for submitting a Pull Request (PR) to the LLVM Project! This PR will be automatically labeled and the relevant teams will be notified. If you wish to, you can add reviewers by using the "Reviewers" section on this page. If this is not working for you, it is probably because you do not have write permissions for the repository. In which case you can instead tag reviewers by name in a comment by using `@` followed by their GitHub username. If you have received no comments on your PR for a week, you can request a review by "ping"ing the PR by adding a comment “Ping”. The common courtesy "ping" rate is once a week. Please remember that you are asking for valuable time from other developers. If you have further questions, they may be answered by the [LLVM GitHub User Guide](https://llvm.org/docs/GitHub.html). You can also ask questions in a comment on this PR, on the [LLVM Discord](https://discord.com/invite/xS7Z362) or on the [forums](https://discourse.llvm.org/). https://github.com/llvm/llvm-project/pull/119428 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [Clang][OpenMP] Support for dispatch construct (Sema & Codegen) support (PR #117904)
@@ -5965,6 +5967,266 @@ static bool teamsLoopCanBeParallelFor(Stmt *AStmt, Sema
&SemaRef) {
return Checker.teamsLoopCanBeParallelFor();
}
+static Expr *getInitialExprFromCapturedExpr(Expr *Cond) {
+
+ Expr *SubExpr = Cond->IgnoreParenImpCasts();
+
+ if (auto *DeclRef = dyn_cast(SubExpr)) {
+if (auto *CapturedExprDecl =
+dyn_cast(DeclRef->getDecl())) {
+
+ // Retrieve the initial expression from the captured expression
+ return CapturedExprDecl->getInit();
+}
+ }
+ return nullptr;
+}
+
+static Expr *replaceWithNewTraitsOrDirectCall(const ASTContext &Context, Expr
*,
+ SemaOpenMP *, bool);
+
+/// cloneAssociatedStmt() function is for cloning the Associated Statement
+/// present with a Directive and then modifying it. By this we avoid modifying
+/// the original Associated Statement.
+static StmtResult cloneAssociatedStmt(const ASTContext &Context, Stmt *StmtP,
+ SemaOpenMP *SemaPtr, bool NoContext) {
+ if (auto *AssocStmt = dyn_cast(StmtP)) {
+CapturedDecl *CDecl = AssocStmt->getCapturedDecl();
+Stmt *AssocExprStmt = AssocStmt->getCapturedStmt();
+auto *AssocExpr = dyn_cast(AssocExprStmt);
+Expr *NewCallOrPseudoObjOrBinExpr = replaceWithNewTraitsOrDirectCall(
+Context, AssocExpr, SemaPtr, NoContext);
+
+// Copy Current Captured Decl to a New Captured Decl for noting the
+// Annotation
+CapturedDecl *NewDecl =
+CapturedDecl::Create(const_cast(Context),
+ CDecl->getDeclContext(), CDecl->getNumParams());
+NewDecl->setBody(static_cast(NewCallOrPseudoObjOrBinExpr));
+for (unsigned I : llvm::seq(CDecl->getNumParams())) {
+ if (I != CDecl->getContextParamPosition())
+NewDecl->setParam(I, CDecl->getParam(I));
+ else
+NewDecl->setContextParam(I, CDecl->getContextParam());
+}
+
+// Create a New Captured Stmt containing the New Captured Decl
+SmallVector Captures;
+SmallVector CaptureInits;
+for (const CapturedStmt::Capture &Capture : AssocStmt->captures())
+ Captures.push_back(Capture);
+for (Expr *CaptureInit : AssocStmt->capture_inits())
+ CaptureInits.push_back(CaptureInit);
+auto *NewStmt = CapturedStmt::Create(
+Context, AssocStmt->getCapturedStmt(),
+AssocStmt->getCapturedRegionKind(), Captures, CaptureInits, NewDecl,
+const_cast(AssocStmt->getCapturedRecordDecl()));
+
+return NewStmt;
+ }
+ return static_cast(nullptr);
+}
+
+/// replaceWithNewTraitsOrDirectCall() is for transforming the call traits.
+/// Call traits associated with a function call are removed and replaced with
+/// a direct call. For clause "nocontext" only, the direct call is then
+/// modified to have call traits for a non-dispatch variant.
+static Expr *replaceWithNewTraitsOrDirectCall(const ASTContext &Context,
+ Expr *AssocExpr,
+ SemaOpenMP *SemaPtr,
+ bool NoContext) {
+ BinaryOperator *BinaryCopyOpr = nullptr;
+ bool IsBinaryOp = false;
+ Expr *PseudoObjExprOrCall = AssocExpr;
+ if (auto *BinOprExpr = dyn_cast(AssocExpr)) {
+IsBinaryOp = true;
+BinaryCopyOpr = BinaryOperator::Create(
+Context, BinOprExpr->getLHS(), BinOprExpr->getRHS(),
+BinOprExpr->getOpcode(), BinOprExpr->getType(),
+BinOprExpr->getValueKind(), BinOprExpr->getObjectKind(),
+BinOprExpr->getOperatorLoc(), FPOptionsOverride());
+PseudoObjExprOrCall = BinaryCopyOpr->getRHS();
+ }
+
+ Expr *CallWithoutInvariants = PseudoObjExprOrCall;
+ // Change PseudoObjectExpr to a direct call
+ if (auto *PseudoObjExpr = dyn_cast(PseudoObjExprOrCall))
+CallWithoutInvariants = *((PseudoObjExpr->semantics_begin()) - 1);
+
+ Expr *FinalCall = CallWithoutInvariants; // For noinvariants clause
+ if (NoContext) {
+// example to explain the changes done for "nocontext" clause:
+//
+// #pragma omp declare variant(foo_variant_dispatch)
+// match(construct = {dispatch})
+// #pragma omp declare variant(foo_variant_allCond)
+// match(user = {condition(1)})
+// ...
+// #pragma omp dispatch nocontext(cond_true)
+// foo(i, j); // with traits: CodeGen call to
+// foo_variant_dispatch(i,j)
+// dispatch construct is changed to:
+// if (cond_true) {
+//foo(i,j) // with traits: CodeGen call to foo_variant_allCond(i,j)
+// } else {
+// #pragma omp dispatch
+// foo(i,j) // with traits: CodeGen call to foo_variant_dispatch(i,j)
+// }
+
+// Convert StmtResult to a CallExpr before calling ActOnOpenMPCall()
+auto *CallExprWithinStmt = cast(CallWithoutInvariants);
+int NumArgs = CallExprWithinStmt->getNumArgs();
+clang::Expr **Args = Cal
[clang] [llvm] [Clang][OpenMP] Support for dispatch construct (Sema & Codegen) support (PR #117904)
@@ -10520,11 +10794,22 @@ StmtResult
SemaOpenMP::ActOnOpenMPSectionDirective(Stmt *AStmt,
DSAStack->isCancelRegion());
}
+/// PseudoObjectExpr is a Trait for dispatch containing the
+/// function and its variant. Returning only the function.
+static Expr *RemovePseudoObjectExpr(Expr *PseudoObjExprOrDirectCall) {
shiltian wrote:
```suggestion
static Expr *removePseudoObjectExpr(Expr *PseudoObjExprOrDirectCall) {
```
https://github.com/llvm/llvm-project/pull/117904
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [Clang][OpenMP] Support for dispatch construct (Sema & Codegen) support (PR #117904)
@@ -5965,6 +5967,266 @@ static bool teamsLoopCanBeParallelFor(Stmt *AStmt, Sema
&SemaRef) {
return Checker.teamsLoopCanBeParallelFor();
}
+static Expr *getInitialExprFromCapturedExpr(Expr *Cond) {
+
+ Expr *SubExpr = Cond->IgnoreParenImpCasts();
+
+ if (auto *DeclRef = dyn_cast(SubExpr)) {
+if (auto *CapturedExprDecl =
+dyn_cast(DeclRef->getDecl())) {
+
+ // Retrieve the initial expression from the captured expression
+ return CapturedExprDecl->getInit();
+}
+ }
+ return nullptr;
+}
+
+static Expr *replaceWithNewTraitsOrDirectCall(const ASTContext &Context, Expr
*,
+ SemaOpenMP *, bool);
+
+/// cloneAssociatedStmt() function is for cloning the Associated Statement
+/// present with a Directive and then modifying it. By this we avoid modifying
+/// the original Associated Statement.
+static StmtResult cloneAssociatedStmt(const ASTContext &Context, Stmt *StmtP,
+ SemaOpenMP *SemaPtr, bool NoContext) {
+ if (auto *AssocStmt = dyn_cast(StmtP)) {
+CapturedDecl *CDecl = AssocStmt->getCapturedDecl();
+Stmt *AssocExprStmt = AssocStmt->getCapturedStmt();
+auto *AssocExpr = dyn_cast(AssocExprStmt);
+Expr *NewCallOrPseudoObjOrBinExpr = replaceWithNewTraitsOrDirectCall(
+Context, AssocExpr, SemaPtr, NoContext);
+
+// Copy Current Captured Decl to a New Captured Decl for noting the
+// Annotation
+CapturedDecl *NewDecl =
+CapturedDecl::Create(const_cast(Context),
+ CDecl->getDeclContext(), CDecl->getNumParams());
+NewDecl->setBody(static_cast(NewCallOrPseudoObjOrBinExpr));
+for (unsigned I : llvm::seq(CDecl->getNumParams())) {
+ if (I != CDecl->getContextParamPosition())
+NewDecl->setParam(I, CDecl->getParam(I));
+ else
+NewDecl->setContextParam(I, CDecl->getContextParam());
+}
+
+// Create a New Captured Stmt containing the New Captured Decl
+SmallVector Captures;
+SmallVector CaptureInits;
+for (const CapturedStmt::Capture &Capture : AssocStmt->captures())
+ Captures.push_back(Capture);
+for (Expr *CaptureInit : AssocStmt->capture_inits())
+ CaptureInits.push_back(CaptureInit);
+auto *NewStmt = CapturedStmt::Create(
+Context, AssocStmt->getCapturedStmt(),
+AssocStmt->getCapturedRegionKind(), Captures, CaptureInits, NewDecl,
+const_cast(AssocStmt->getCapturedRecordDecl()));
+
+return NewStmt;
+ }
+ return static_cast(nullptr);
+}
+
+/// replaceWithNewTraitsOrDirectCall() is for transforming the call traits.
+/// Call traits associated with a function call are removed and replaced with
+/// a direct call. For clause "nocontext" only, the direct call is then
+/// modified to have call traits for a non-dispatch variant.
+static Expr *replaceWithNewTraitsOrDirectCall(const ASTContext &Context,
+ Expr *AssocExpr,
+ SemaOpenMP *SemaPtr,
+ bool NoContext) {
+ BinaryOperator *BinaryCopyOpr = nullptr;
+ bool IsBinaryOp = false;
+ Expr *PseudoObjExprOrCall = AssocExpr;
+ if (auto *BinOprExpr = dyn_cast(AssocExpr)) {
+IsBinaryOp = true;
+BinaryCopyOpr = BinaryOperator::Create(
+Context, BinOprExpr->getLHS(), BinOprExpr->getRHS(),
+BinOprExpr->getOpcode(), BinOprExpr->getType(),
+BinOprExpr->getValueKind(), BinOprExpr->getObjectKind(),
+BinOprExpr->getOperatorLoc(), FPOptionsOverride());
+PseudoObjExprOrCall = BinaryCopyOpr->getRHS();
+ }
+
+ Expr *CallWithoutInvariants = PseudoObjExprOrCall;
+ // Change PseudoObjectExpr to a direct call
+ if (auto *PseudoObjExpr = dyn_cast(PseudoObjExprOrCall))
+CallWithoutInvariants = *((PseudoObjExpr->semantics_begin()) - 1);
+
+ Expr *FinalCall = CallWithoutInvariants; // For noinvariants clause
+ if (NoContext) {
+// example to explain the changes done for "nocontext" clause:
+//
+// #pragma omp declare variant(foo_variant_dispatch)
+// match(construct = {dispatch})
+// #pragma omp declare variant(foo_variant_allCond)
+// match(user = {condition(1)})
+// ...
+// #pragma omp dispatch nocontext(cond_true)
+// foo(i, j); // with traits: CodeGen call to
+// foo_variant_dispatch(i,j)
+// dispatch construct is changed to:
+// if (cond_true) {
+//foo(i,j) // with traits: CodeGen call to foo_variant_allCond(i,j)
+// } else {
+// #pragma omp dispatch
+// foo(i,j) // with traits: CodeGen call to foo_variant_dispatch(i,j)
+// }
+
+// Convert StmtResult to a CallExpr before calling ActOnOpenMPCall()
+auto *CallExprWithinStmt = cast(CallWithoutInvariants);
+int NumArgs = CallExprWithinStmt->getNumArgs();
+clang::Expr **Args = Cal
[clang] [llvm] [Clang][OpenMP] Support for dispatch construct (Sema & Codegen) support (PR #117904)
@@ -5965,6 +5967,266 @@ static bool teamsLoopCanBeParallelFor(Stmt *AStmt, Sema
&SemaRef) {
return Checker.teamsLoopCanBeParallelFor();
}
+static Expr *getInitialExprFromCapturedExpr(Expr *Cond) {
+
+ Expr *SubExpr = Cond->IgnoreParenImpCasts();
+
+ if (auto *DeclRef = dyn_cast(SubExpr)) {
+if (auto *CapturedExprDecl =
+dyn_cast(DeclRef->getDecl())) {
+
+ // Retrieve the initial expression from the captured expression
+ return CapturedExprDecl->getInit();
+}
+ }
+ return nullptr;
+}
+
+static Expr *replaceWithNewTraitsOrDirectCall(const ASTContext &Context, Expr
*,
+ SemaOpenMP *, bool);
+
+/// cloneAssociatedStmt() function is for cloning the Associated Statement
+/// present with a Directive and then modifying it. By this we avoid modifying
+/// the original Associated Statement.
+static StmtResult cloneAssociatedStmt(const ASTContext &Context, Stmt *StmtP,
+ SemaOpenMP *SemaPtr, bool NoContext) {
+ if (auto *AssocStmt = dyn_cast(StmtP)) {
+CapturedDecl *CDecl = AssocStmt->getCapturedDecl();
+Stmt *AssocExprStmt = AssocStmt->getCapturedStmt();
+auto *AssocExpr = dyn_cast(AssocExprStmt);
+Expr *NewCallOrPseudoObjOrBinExpr = replaceWithNewTraitsOrDirectCall(
+Context, AssocExpr, SemaPtr, NoContext);
+
+// Copy Current Captured Decl to a New Captured Decl for noting the
+// Annotation
+CapturedDecl *NewDecl =
+CapturedDecl::Create(const_cast(Context),
+ CDecl->getDeclContext(), CDecl->getNumParams());
+NewDecl->setBody(static_cast(NewCallOrPseudoObjOrBinExpr));
+for (unsigned I : llvm::seq(CDecl->getNumParams())) {
+ if (I != CDecl->getContextParamPosition())
+NewDecl->setParam(I, CDecl->getParam(I));
+ else
+NewDecl->setContextParam(I, CDecl->getContextParam());
+}
+
+// Create a New Captured Stmt containing the New Captured Decl
+SmallVector Captures;
+SmallVector CaptureInits;
+for (const CapturedStmt::Capture &Capture : AssocStmt->captures())
+ Captures.push_back(Capture);
+for (Expr *CaptureInit : AssocStmt->capture_inits())
+ CaptureInits.push_back(CaptureInit);
+auto *NewStmt = CapturedStmt::Create(
+Context, AssocStmt->getCapturedStmt(),
+AssocStmt->getCapturedRegionKind(), Captures, CaptureInits, NewDecl,
+const_cast(AssocStmt->getCapturedRecordDecl()));
+
+return NewStmt;
+ }
+ return static_cast(nullptr);
+}
+
+/// replaceWithNewTraitsOrDirectCall() is for transforming the call traits.
+/// Call traits associated with a function call are removed and replaced with
+/// a direct call. For clause "nocontext" only, the direct call is then
+/// modified to have call traits for a non-dispatch variant.
+static Expr *replaceWithNewTraitsOrDirectCall(const ASTContext &Context,
+ Expr *AssocExpr,
+ SemaOpenMP *SemaPtr,
+ bool NoContext) {
+ BinaryOperator *BinaryCopyOpr = nullptr;
+ bool IsBinaryOp = false;
+ Expr *PseudoObjExprOrCall = AssocExpr;
+ if (auto *BinOprExpr = dyn_cast(AssocExpr)) {
+IsBinaryOp = true;
+BinaryCopyOpr = BinaryOperator::Create(
+Context, BinOprExpr->getLHS(), BinOprExpr->getRHS(),
+BinOprExpr->getOpcode(), BinOprExpr->getType(),
+BinOprExpr->getValueKind(), BinOprExpr->getObjectKind(),
+BinOprExpr->getOperatorLoc(), FPOptionsOverride());
+PseudoObjExprOrCall = BinaryCopyOpr->getRHS();
+ }
+
+ Expr *CallWithoutInvariants = PseudoObjExprOrCall;
+ // Change PseudoObjectExpr to a direct call
+ if (auto *PseudoObjExpr = dyn_cast(PseudoObjExprOrCall))
+CallWithoutInvariants = *((PseudoObjExpr->semantics_begin()) - 1);
+
+ Expr *FinalCall = CallWithoutInvariants; // For noinvariants clause
+ if (NoContext) {
+// example to explain the changes done for "nocontext" clause:
+//
+// #pragma omp declare variant(foo_variant_dispatch)
+// match(construct = {dispatch})
+// #pragma omp declare variant(foo_variant_allCond)
+// match(user = {condition(1)})
+// ...
+// #pragma omp dispatch nocontext(cond_true)
+// foo(i, j); // with traits: CodeGen call to
+// foo_variant_dispatch(i,j)
+// dispatch construct is changed to:
+// if (cond_true) {
+//foo(i,j) // with traits: CodeGen call to foo_variant_allCond(i,j)
+// } else {
+// #pragma omp dispatch
+// foo(i,j) // with traits: CodeGen call to foo_variant_dispatch(i,j)
+// }
+
+// Convert StmtResult to a CallExpr before calling ActOnOpenMPCall()
+auto *CallExprWithinStmt = cast(CallWithoutInvariants);
+int NumArgs = CallExprWithinStmt->getNumArgs();
+clang::Expr **Args = Cal
[clang] [HLSL] Move `_init_resource_bindings` to entry point functions (PR #119311)
llvm-beanz wrote: Is this the right solution, or should we be doing something to force the constructor function to be generated. I'm a little worried that we have one path for initializing some types of globals and a different approach for other globals. https://github.com/llvm/llvm-project/pull/119311 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [Clang][OpenMP] Support for dispatch construct (Sema & Codegen) support (PR #117904)
https://github.com/shiltian edited https://github.com/llvm/llvm-project/pull/117904 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [NVPTX] Add NVVMUpgradeAnnotations pass to cleanup legacy annotations (PR #119261)
@@ -302,6 +299,19 @@ void NVPTXTargetCodeGenInfo::addNVVMMetadata(
llvm::ConstantAsMetadata::get(GV), llvm::MDString::get(Ctx, Name),
llvm::ConstantAsMetadata::get(
llvm::ConstantInt::get(llvm::Type::getInt32Ty(Ctx), Operand))};
+ // Append metadata to nvvm.annotations
+ MD->addOperand(llvm::MDNode::get(Ctx, MDVals));
+}
+
+void NVPTXTargetCodeGenInfo::addNVVMGridConstantMetadata(
+llvm::GlobalValue *GV, const SmallVectorImpl &GridConstantArgs) {
+ llvm::Module *M = GV->getParent();
+ llvm::LLVMContext &Ctx = M->getContext();
+
+ // Get "nvvm.annotations" metadata node
+ llvm::NamedMDNode *MD = M->getOrInsertNamedMetadata("nvvm.annotations");
Artem-B wrote:
> For this MR I want to lay down the framework
SGTM. We don't have to it all at once.
> Specifically for grid_constant can we just upgrade to the existing readonly
> parameter attribute?
That would be even better. No custom attributes *and* we can take advantage of
local copy avoidance without explicit user annotations. Sounds like a win-win
to me. I can't think of a reason not to do that.
https://github.com/llvm/llvm-project/pull/119261
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [Clang][OpenMP] Support for dispatch construct (Sema & Codegen) support (PR #117904)
@@ -7209,8 +7482,9 @@ ExprResult SemaOpenMP::ActOnOpenMPCall(ExprResult Call,
Scope *Scope,
Exprs.erase(Exprs.begin() + BestIdx);
} while (!VMIs.empty());
- if (!NewCall.isUsable())
+ if (!NewCall.isUsable()) {
return Call;
+ }
shiltian wrote:
unrelated
https://github.com/llvm/llvm-project/pull/117904
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [Clang][OpenMP] Support for dispatch construct (Sema & Codegen) support (PR #117904)
@@ -10556,15 +10841,18 @@
SemaOpenMP::ActOnOpenMPDispatchDirective(ArrayRef Clauses,
E = E->IgnoreParenCasts()->IgnoreImplicit();
if (auto *BO = dyn_cast(E)) {
- if (BO->getOpcode() == BO_Assign)
+ if (BO->getOpcode() == BO_Assign) {
TargetCall = getDirectCallExpr(BO->getRHS());
+ }
} else {
if (auto *COCE = dyn_cast(E))
if (COCE->getOperator() == OO_Equal)
TargetCall = getDirectCallExpr(COCE->getArg(1));
- if (!TargetCall)
+ if (!TargetCall) {
TargetCall = getDirectCallExpr(E);
+ }
}
+
shiltian wrote:
unrelated changes
https://github.com/llvm/llvm-project/pull/117904
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [Clang][OpenMP] Support for dispatch construct (Sema & Codegen) support (PR #117904)
@@ -5979,6 +6241,17 @@ StmtResult SemaOpenMP::ActOnOpenMPExecutableDirective(
OMPExecutableDirective::getSingleClause(Clauses))
BindKind = BC->getBindKind();
+ if ((Kind == OMPD_dispatch) && (Clauses.size() > 0)) {
shiltian wrote:
```suggestion
if ((Kind == OMPD_dispatch) && (!Clauses.empty())) {
```
https://github.com/llvm/llvm-project/pull/117904
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [Clang][OpenMP] Support for dispatch construct (Sema & Codegen) support (PR #117904)
@@ -5965,6 +5967,266 @@ static bool teamsLoopCanBeParallelFor(Stmt *AStmt, Sema
&SemaRef) {
return Checker.teamsLoopCanBeParallelFor();
}
+static Expr *getInitialExprFromCapturedExpr(Expr *Cond) {
+
+ Expr *SubExpr = Cond->IgnoreParenImpCasts();
+
+ if (auto *DeclRef = dyn_cast(SubExpr)) {
+if (auto *CapturedExprDecl =
+dyn_cast(DeclRef->getDecl())) {
+
+ // Retrieve the initial expression from the captured expression
+ return CapturedExprDecl->getInit();
+}
+ }
+ return nullptr;
+}
+
+static Expr *replaceWithNewTraitsOrDirectCall(const ASTContext &Context, Expr
*,
+ SemaOpenMP *, bool);
+
+/// cloneAssociatedStmt() function is for cloning the Associated Statement
+/// present with a Directive and then modifying it. By this we avoid modifying
+/// the original Associated Statement.
+static StmtResult cloneAssociatedStmt(const ASTContext &Context, Stmt *StmtP,
+ SemaOpenMP *SemaPtr, bool NoContext) {
+ if (auto *AssocStmt = dyn_cast(StmtP)) {
+CapturedDecl *CDecl = AssocStmt->getCapturedDecl();
+Stmt *AssocExprStmt = AssocStmt->getCapturedStmt();
+auto *AssocExpr = dyn_cast(AssocExprStmt);
+Expr *NewCallOrPseudoObjOrBinExpr = replaceWithNewTraitsOrDirectCall(
+Context, AssocExpr, SemaPtr, NoContext);
+
+// Copy Current Captured Decl to a New Captured Decl for noting the
+// Annotation
+CapturedDecl *NewDecl =
+CapturedDecl::Create(const_cast(Context),
+ CDecl->getDeclContext(), CDecl->getNumParams());
+NewDecl->setBody(static_cast(NewCallOrPseudoObjOrBinExpr));
+for (unsigned I : llvm::seq(CDecl->getNumParams())) {
+ if (I != CDecl->getContextParamPosition())
+NewDecl->setParam(I, CDecl->getParam(I));
+ else
+NewDecl->setContextParam(I, CDecl->getContextParam());
+}
+
+// Create a New Captured Stmt containing the New Captured Decl
+SmallVector Captures;
+SmallVector CaptureInits;
+for (const CapturedStmt::Capture &Capture : AssocStmt->captures())
+ Captures.push_back(Capture);
+for (Expr *CaptureInit : AssocStmt->capture_inits())
+ CaptureInits.push_back(CaptureInit);
+auto *NewStmt = CapturedStmt::Create(
+Context, AssocStmt->getCapturedStmt(),
+AssocStmt->getCapturedRegionKind(), Captures, CaptureInits, NewDecl,
+const_cast(AssocStmt->getCapturedRecordDecl()));
+
+return NewStmt;
+ }
+ return static_cast(nullptr);
+}
+
+/// replaceWithNewTraitsOrDirectCall() is for transforming the call traits.
+/// Call traits associated with a function call are removed and replaced with
+/// a direct call. For clause "nocontext" only, the direct call is then
+/// modified to have call traits for a non-dispatch variant.
+static Expr *replaceWithNewTraitsOrDirectCall(const ASTContext &Context,
+ Expr *AssocExpr,
+ SemaOpenMP *SemaPtr,
+ bool NoContext) {
+ BinaryOperator *BinaryCopyOpr = nullptr;
+ bool IsBinaryOp = false;
+ Expr *PseudoObjExprOrCall = AssocExpr;
+ if (auto *BinOprExpr = dyn_cast(AssocExpr)) {
+IsBinaryOp = true;
+BinaryCopyOpr = BinaryOperator::Create(
+Context, BinOprExpr->getLHS(), BinOprExpr->getRHS(),
+BinOprExpr->getOpcode(), BinOprExpr->getType(),
+BinOprExpr->getValueKind(), BinOprExpr->getObjectKind(),
+BinOprExpr->getOperatorLoc(), FPOptionsOverride());
+PseudoObjExprOrCall = BinaryCopyOpr->getRHS();
+ }
+
+ Expr *CallWithoutInvariants = PseudoObjExprOrCall;
+ // Change PseudoObjectExpr to a direct call
+ if (auto *PseudoObjExpr = dyn_cast(PseudoObjExprOrCall))
+CallWithoutInvariants = *((PseudoObjExpr->semantics_begin()) - 1);
+
+ Expr *FinalCall = CallWithoutInvariants; // For noinvariants clause
+ if (NoContext) {
+// example to explain the changes done for "nocontext" clause:
+//
+// #pragma omp declare variant(foo_variant_dispatch)
+// match(construct = {dispatch})
+// #pragma omp declare variant(foo_variant_allCond)
+// match(user = {condition(1)})
+// ...
+// #pragma omp dispatch nocontext(cond_true)
+// foo(i, j); // with traits: CodeGen call to
+// foo_variant_dispatch(i,j)
+// dispatch construct is changed to:
+// if (cond_true) {
+//foo(i,j) // with traits: CodeGen call to foo_variant_allCond(i,j)
+// } else {
+// #pragma omp dispatch
+// foo(i,j) // with traits: CodeGen call to foo_variant_dispatch(i,j)
+// }
+
+// Convert StmtResult to a CallExpr before calling ActOnOpenMPCall()
+auto *CallExprWithinStmt = cast(CallWithoutInvariants);
+int NumArgs = CallExprWithinStmt->getNumArgs();
+clang::Expr **Args = Cal
[clang] dc04d41 - SystemZ: Add support for __builtin_setjmp and __builtin_longjmp. (#119257)
Author: anoopkg6
Date: 2024-12-10T19:50:51+01:00
New Revision: dc04d414df9c243bb90d7cfc683a632a2c032c62
URL:
https://github.com/llvm/llvm-project/commit/dc04d414df9c243bb90d7cfc683a632a2c032c62
DIFF:
https://github.com/llvm/llvm-project/commit/dc04d414df9c243bb90d7cfc683a632a2c032c62.diff
LOG: SystemZ: Add support for __builtin_setjmp and __builtin_longjmp. (#119257)
This pr includes fixes for original pr##116642.
Implementation for __builtin_setjmp and __builtin_longjmp for SystemZ..
Added:
clang/test/CodeGen/SystemZ/builtin-setjmp-logjmp.c
llvm/test/CodeGen/SystemZ/builtin-longjmp.ll
llvm/test/CodeGen/SystemZ/builtin-setjmp-alloca.ll
llvm/test/CodeGen/SystemZ/builtin-setjmp-spills.ll
llvm/test/CodeGen/SystemZ/builtin-setjmp.ll
Modified:
clang/lib/Basic/Targets/SystemZ.h
clang/lib/CodeGen/CGBuiltin.cpp
llvm/docs/ExceptionHandling.rst
llvm/lib/Target/SystemZ/SystemZAsmPrinter.cpp
llvm/lib/Target/SystemZ/SystemZISelLowering.cpp
llvm/lib/Target/SystemZ/SystemZISelLowering.h
llvm/lib/Target/SystemZ/SystemZInstrInfo.td
llvm/lib/Target/SystemZ/SystemZLongBranch.cpp
llvm/lib/Target/SystemZ/SystemZOperators.td
llvm/lib/Target/SystemZ/SystemZRegisterInfo.cpp
llvm/lib/Target/SystemZ/SystemZRegisterInfo.h
Removed:
diff --git a/clang/lib/Basic/Targets/SystemZ.h
b/clang/lib/Basic/Targets/SystemZ.h
index 98637f230b69cc..04e00bac644765 100644
--- a/clang/lib/Basic/Targets/SystemZ.h
+++ b/clang/lib/Basic/Targets/SystemZ.h
@@ -248,6 +248,8 @@ class LLVM_LIBRARY_VISIBILITY SystemZTargetInfo : public
TargetInfo {
return RegNo < 4 ? 6 + RegNo : -1;
}
+ bool hasSjLjLowering() const override { return true; }
+
std::pair hardwareInterferenceSizes() const override {
return std::make_pair(256, 256);
}
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index c2e983eebebc10..49a4c1ecc825e7 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -4860,6 +4860,14 @@ RValue CodeGenFunction::EmitBuiltinExpr(const GlobalDecl
GD, unsigned BuiltinID,
// Buffer is a void**.
Address Buf = EmitPointerWithAlignment(E->getArg(0));
+if (getTarget().getTriple().getArch() == llvm::Triple::systemz) {
+ // On this target, the back end fills in the context buffer completely.
+ // It doesn't really matter if the frontend stores to the buffer before
+ // calling setjmp, the back-end is going to overwrite them anyway.
+ Function *F = CGM.getIntrinsic(Intrinsic::eh_sjlj_setjmp);
+ return RValue::get(Builder.CreateCall(F, Buf.emitRawPointer(*this)));
+}
+
// Store the frame pointer to the setjmp buffer.
Value *FrameAddr = Builder.CreateCall(
CGM.getIntrinsic(Intrinsic::frameaddress, AllocaInt8PtrTy),
diff --git a/clang/test/CodeGen/SystemZ/builtin-setjmp-logjmp.c
b/clang/test/CodeGen/SystemZ/builtin-setjmp-logjmp.c
new file mode 100644
index 00..898891fa182ea4
--- /dev/null
+++ b/clang/test/CodeGen/SystemZ/builtin-setjmp-logjmp.c
@@ -0,0 +1,26 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
UTC_ARGS: --version 5
+// RUN: %clang_cc1 -triple s390x-linux -emit-llvm -o - %s | FileCheck %s
+
+void *buf[20];
+
+// CHECK-LABEL: define dso_local void @foo(
+// CHECK-SAME: ) #[[ATTR0:[0-9]+]] {
+// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT:[[TMP0:%.*]] = call i32 @llvm.eh.sjlj.setjmp(ptr @buf)
+// CHECK-NEXT:ret void
+//
+void foo()
+{
+ __builtin_setjmp (buf);
+}
+
+// CHECK-LABEL: define dso_local void @foo1(
+// CHECK-SAME: ) #[[ATTR0]] {
+// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT:call void @llvm.eh.sjlj.longjmp(ptr @buf)
+// CHECK-NEXT:unreachable
+//
+void foo1()
+{
+ __builtin_longjmp (buf, 1);
+}
diff --git a/llvm/docs/ExceptionHandling.rst b/llvm/docs/ExceptionHandling.rst
index c2afb25c6ccd47..bb72e5a71a77bf 100644
--- a/llvm/docs/ExceptionHandling.rst
+++ b/llvm/docs/ExceptionHandling.rst
@@ -374,10 +374,12 @@ overall functioning of this intrinsic is compatible with
the GCC
to interoperate.
The single parameter is a pointer to a five word buffer in which the calling
-context is saved. The front end places the frame pointer in the first word, and
-the target implementation of this intrinsic should place the destination
address
-for a `llvm.eh.sjlj.longjmp`_ in the second word. The following three words are
-available for use in a target-specific manner.
+context is saved. The format and contents of the buffer are target-specific.
+On certain targets (ARM, PowerPC, VE, X86), the front end places the
+frame pointer in the first word and the stack pointer in the third word,
+while the target implementation of this intrinsic fills in the remaining
+words. On other targets (SystemZ), saving the calling context to the buffer
+is left completely to the target implementat
[clang] [llvm] SystemZ: Add support for __builtin_setjmp and __builtin_longjmp. (PR #119257)
https://github.com/uweigand closed https://github.com/llvm/llvm-project/pull/119257 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [Clang][OpenMP] Support for dispatch construct (Sema & Codegen) support (PR #117904)
https://github.com/SunilKuravinakop updated
https://github.com/llvm/llvm-project/pull/117904
>From 1703aa62cfe35538aedbacb28e907535e838248c Mon Sep 17 00:00:00 2001
From: Sunil Kuravinakop
Date: Fri, 20 Sep 2024 01:41:29 -0500
Subject: [PATCH 1/4] Support for dispatch construct (Sema & Codegen) support.
Support for clauses depend, novariants & nocontext.
---
.../clang/Basic/DiagnosticSemaKinds.td| 3 +
clang/include/clang/Basic/OpenMPKinds.h | 6 +
clang/include/clang/Sema/SemaOpenMP.h | 7 +
clang/lib/Basic/OpenMPKinds.cpp | 5 +
clang/lib/CodeGen/CGStmt.cpp | 2 +-
clang/lib/CodeGen/CGStmtOpenMP.cpp| 4 +
clang/lib/CodeGen/CodeGenFunction.h | 1 +
clang/lib/Sema/SemaOpenMP.cpp | 301 ++-
clang/test/OpenMP/dispatch_codegen.cpp| 359 ++
clang/test/OpenMP/dispatch_unsupported.c | 7 -
.../include/llvm/Frontend/OpenMP/OMPContext.h | 6 +
11 files changed, 687 insertions(+), 14 deletions(-)
create mode 100644 clang/test/OpenMP/dispatch_codegen.cpp
delete mode 100644 clang/test/OpenMP/dispatch_unsupported.c
diff --git a/clang/include/clang/Basic/DiagnosticSemaKinds.td
b/clang/include/clang/Basic/DiagnosticSemaKinds.td
index 8495884dcd058f..81b876f9fd85c5 100644
--- a/clang/include/clang/Basic/DiagnosticSemaKinds.td
+++ b/clang/include/clang/Basic/DiagnosticSemaKinds.td
@@ -11774,6 +11774,9 @@ def err_omp_clause_requires_dispatch_construct : Error<
"'%0' clause requires 'dispatch' context selector">;
def err_omp_append_args_with_varargs : Error<
"'append_args' is not allowed with varargs functions">;
+def warn_omp_dispatch_clause_novariants_nocontext : Warning<
+ "only 'novariants' clause is supported when 'novariants' & 'nocontext'
clauses occur on the same dispatch construct">,
+ InGroup;
def err_openmp_vla_in_task_untied : Error<
"variable length arrays are not supported in OpenMP tasking regions with
'untied' clause">;
def warn_omp_unterminated_declare_target : Warning<
diff --git a/clang/include/clang/Basic/OpenMPKinds.h
b/clang/include/clang/Basic/OpenMPKinds.h
index 900ad6ca6d66f6..7579fab43dbb19 100644
--- a/clang/include/clang/Basic/OpenMPKinds.h
+++ b/clang/include/clang/Basic/OpenMPKinds.h
@@ -269,6 +269,12 @@ bool isOpenMPTaskLoopDirective(OpenMPDirectiveKind DKind);
/// parallel', otherwise - false.
bool isOpenMPParallelDirective(OpenMPDirectiveKind DKind);
+/// Checks if the specified directive is a dispatch-kind directive.
+/// \param DKind Specified directive.
+/// \return true - the directive is a dispatch-like directive like 'omp
+/// dispatch', otherwise - false.
+bool isOpenMPDispatchDirective(OpenMPDirectiveKind DKind);
+
/// Checks if the specified directive is a target code offload directive.
/// \param DKind Specified directive.
/// \return true - the directive is a target code offload directive like
diff --git a/clang/include/clang/Sema/SemaOpenMP.h
b/clang/include/clang/Sema/SemaOpenMP.h
index 3d1cc4fab1c10f..80cee9e7583051 100644
--- a/clang/include/clang/Sema/SemaOpenMP.h
+++ b/clang/include/clang/Sema/SemaOpenMP.h
@@ -1462,6 +1462,13 @@ class SemaOpenMP : public SemaBase {
: OMPDeclareVariantScopes.back().TI;
}
+ StmtResult transformDispatchDirective(OpenMPDirectiveKind Kind,
+const DeclarationNameInfo &DirName,
+OpenMPDirectiveKind CancelRegion,
+ArrayRef Clauses,
+Stmt *AStmt, SourceLocation StartLoc,
+SourceLocation EndLoc);
+
/// The current `omp begin/end declare variant` scopes.
SmallVector OMPDeclareVariantScopes;
diff --git a/clang/lib/Basic/OpenMPKinds.cpp b/clang/lib/Basic/OpenMPKinds.cpp
index 62a13f01481b28..44ee63df46adb5 100644
--- a/clang/lib/Basic/OpenMPKinds.cpp
+++ b/clang/lib/Basic/OpenMPKinds.cpp
@@ -621,6 +621,11 @@ bool clang::isOpenMPParallelDirective(OpenMPDirectiveKind
DKind) {
llvm::is_contained(getLeafConstructs(DKind), OMPD_parallel);
}
+bool clang::isOpenMPDispatchDirective(OpenMPDirectiveKind DKind) {
+ return DKind == OMPD_dispatch ||
+ llvm::is_contained(getLeafConstructs(DKind), OMPD_target);
+}
+
bool clang::isOpenMPTargetExecutionDirective(OpenMPDirectiveKind DKind) {
return DKind == OMPD_target ||
llvm::is_contained(getLeafConstructs(DKind), OMPD_target);
diff --git a/clang/lib/CodeGen/CGStmt.cpp b/clang/lib/CodeGen/CGStmt.cpp
index 698baf853507f4..f3eedb79844378 100644
--- a/clang/lib/CodeGen/CGStmt.cpp
+++ b/clang/lib/CodeGen/CGStmt.cpp
@@ -417,7 +417,7 @@ void CodeGenFunction::EmitStmt(const Stmt *S,
ArrayRef Attrs) {
EmitOMPInteropDirective(cast(*S));
break;
case Stmt::OMPDispatchDirectiveClass:
-CGM.ErrorUnsupported(S, "OpenMP dispatch directive");
+
[clang] [HLSL] Move `_init_resource_bindings` to entry point functions (PR #119311)
https://github.com/hekota converted_to_draft https://github.com/llvm/llvm-project/pull/119311 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [HLSL][SPIRV][DXIL] Implement `WaveActiveSum` intrinsic (PR #118580)
@@ -9272,6 +9272,8 @@ def err_typecheck_expect_scalar_or_vector : Error< "a vector of such type is required">; def err_typecheck_expect_any_scalar_or_vector : Error< "invalid operand of type %0 where a scalar or vector is required">; +def err_typecheck_expect_scalar_or_vector_not_type : Error< llvm-beanz wrote: Do we need a new unique diagnostic for this or should we be using `err_typecheck_expect_any_scalar_or_vector` or `err_typecheck_expect_scalar_or_vector`? In general we should try not to add unique diagnostics unless there really isn't an existing one that works. In looking at this more closely I'm not sure we should have added `err_typecheck_expect_any_scalar_or_vector`, as we've basically added two _extremely similar_ diagnostics that probably should be combined. https://github.com/llvm/llvm-project/pull/118580 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [HLSL] Move `_init_resource_bindings` to entry point functions (PR #119311)
hekota wrote: > Is this the right solution, or should we be doing something to force the > constructor function to be generated? I'm a little worried that we have one > path for initializing some types of globals and a different approach for > other globals. I'll see if I can move creation of the `cbuffer` global variable right after it is declared and that should get captured by the module init. We might still need to generate the body of the resource init function later in codegen. https://github.com/llvm/llvm-project/pull/119311 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [AArch64] Implement intrinsics for SME FP8 FMLAL/FMLALL (Indexed) (PR #118549)
CarolineConcatto wrote: I will wait you merge the previous patch to have a look on this one again later! https://github.com/llvm/llvm-project/pull/118549 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [Clang][OpenMP] Support for dispatch construct (Sema & Codegen) support (PR #117904)
@@ -5965,6 +5967,266 @@ static bool teamsLoopCanBeParallelFor(Stmt *AStmt, Sema
&SemaRef) {
return Checker.teamsLoopCanBeParallelFor();
}
+static Expr *getInitialExprFromCapturedExpr(Expr *Cond) {
+
+ Expr *SubExpr = Cond->IgnoreParenImpCasts();
+
+ if (auto *DeclRef = dyn_cast(SubExpr)) {
+if (auto *CapturedExprDecl =
+dyn_cast(DeclRef->getDecl())) {
+
+ // Retrieve the initial expression from the captured expression
+ return CapturedExprDecl->getInit();
+}
+ }
+ return nullptr;
+}
+
+static Expr *replaceWithNewTraitsOrDirectCall(const ASTContext &Context, Expr
*,
+ SemaOpenMP *, bool);
+
+/// cloneAssociatedStmt() function is for cloning the Associated Statement
+/// present with a Directive and then modifying it. By this we avoid modifying
+/// the original Associated Statement.
+static StmtResult cloneAssociatedStmt(const ASTContext &Context, Stmt *StmtP,
+ SemaOpenMP *SemaPtr, bool NoContext) {
+ if (auto *AssocStmt = dyn_cast(StmtP)) {
+CapturedDecl *CDecl = AssocStmt->getCapturedDecl();
+Stmt *AssocExprStmt = AssocStmt->getCapturedStmt();
+auto *AssocExpr = dyn_cast(AssocExprStmt);
+Expr *NewCallOrPseudoObjOrBinExpr = replaceWithNewTraitsOrDirectCall(
+Context, AssocExpr, SemaPtr, NoContext);
+
+// Copy Current Captured Decl to a New Captured Decl for noting the
+// Annotation
+CapturedDecl *NewDecl =
+CapturedDecl::Create(const_cast(Context),
+ CDecl->getDeclContext(), CDecl->getNumParams());
+NewDecl->setBody(static_cast(NewCallOrPseudoObjOrBinExpr));
+for (unsigned I : llvm::seq(CDecl->getNumParams())) {
+ if (I != CDecl->getContextParamPosition())
+NewDecl->setParam(I, CDecl->getParam(I));
+ else
+NewDecl->setContextParam(I, CDecl->getContextParam());
+}
+
+// Create a New Captured Stmt containing the New Captured Decl
+SmallVector Captures;
+SmallVector CaptureInits;
+for (const CapturedStmt::Capture &Capture : AssocStmt->captures())
+ Captures.push_back(Capture);
+for (Expr *CaptureInit : AssocStmt->capture_inits())
+ CaptureInits.push_back(CaptureInit);
+auto *NewStmt = CapturedStmt::Create(
+Context, AssocStmt->getCapturedStmt(),
+AssocStmt->getCapturedRegionKind(), Captures, CaptureInits, NewDecl,
+const_cast(AssocStmt->getCapturedRecordDecl()));
+
+return NewStmt;
+ }
+ return static_cast(nullptr);
shiltian wrote:
I'm confused by this return. Why do you want to cast a `nullptr` to a `Stmt *`,
and then implicit cast it to `StmtResult`?
https://github.com/llvm/llvm-project/pull/117904
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [DebugInfo] Place local ODR-uniqued types in decl DISubprograms (PR #119001)
dwblaikie wrote: > > Maybe it's too big/complex a problem to try to think about, and incremental > > development should overrule here - but may be worth thinking about? > > Urgh. I'll admit that I hadn't thought that far ahead, I just wanted to get > the patch-series going again now that the root problem is identified. > > The DWARF in the existing patch series [0] seems to take a similar direction > to this patch, of putting information into declarations where possible. Have > a look at the check-lines for > llvm/test/DebugInfo/Generic/lexical-block-static-var.ll: > the static-local variables are defined and given a location inside > DW_TAG_lexical_blocks within the abstract definitions of functions, Ah, fair enough, because the subprogram will contian a list of these local types, then the local type's scope will be a (potentially series of) lexical scope - so the lexical scopes are preserved? That's OK-ish, at least not as problematic as I was imagining. (rest of the discussion below here is just discussion - this point ^ is enough to satisfy my concerns for this patch direction, at least) > and then there are some other DW_TAG_lexical_blocks inside the inlined > instances. It's not clear to me how the inlined-lexical-blocks refer back to > the abstract ones. Yeah, that's probably missing - the ideal would be the inlined/concrete non-inlined instance's lexical blocks should have abstract_origins that refer to the abstract lexical blocks. > Given that the ODR-uniquing issue was only found experimentally in-the-field, > I think we have to accept the incremental approach and discover problems by > experience. Not sure I follow this. > Stepping back slightly: I think the ODR-uniquing of these types is directly > opposed to scope-precision: reducing the number of type definitions down to > one _necessarily_ means discarding the extra locality information about where > it's scoped in different function instances. Not sure I follow this either - I think that if each concrete lexical scope references the abstract ones, we could emit correct abstract origins to get the local static variables and types to be scoped appropriately in the resulting DWARF/for the consumer. GCC does this: https://godbolt.org/z/Khrzdq1Wx https://github.com/llvm/llvm-project/pull/119001 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [RISCV] Add stack clash protection (PR #117612)
llvm-ci wrote: LLVM Buildbot has detected a new failure on builder `llvm-clang-x86_64-expensive-checks-debian` running on `gribozavr4` while building `clang,llvm` at step 6 "test-build-unified-tree-check-all". Full details are available at: https://lab.llvm.org/buildbot/#/builders/16/builds/10433 Here is the relevant piece of the build log for the reference ``` Step 6 (test-build-unified-tree-check-all) failure: test (failure) TEST 'LLVM :: CodeGen/RISCV/stack-clash-prologue-nounwind.ll' FAILED Exit Code: 2 Command Output (stderr): -- RUN: at line 2: /b/1/llvm-clang-x86_64-expensive-checks-debian/build/bin/llc -mtriple=riscv64 -mattr=+m -O2 < /b/1/llvm-clang-x86_64-expensive-checks-debian/llvm-project/llvm/test/CodeGen/RISCV/stack-clash-prologue-nounwind.ll | /b/1/llvm-clang-x86_64-expensive-checks-debian/build/bin/FileCheck /b/1/llvm-clang-x86_64-expensive-checks-debian/llvm-project/llvm/test/CodeGen/RISCV/stack-clash-prologue-nounwind.ll -check-prefix=RV64I + /b/1/llvm-clang-x86_64-expensive-checks-debian/build/bin/llc -mtriple=riscv64 -mattr=+m -O2 + /b/1/llvm-clang-x86_64-expensive-checks-debian/build/bin/FileCheck /b/1/llvm-clang-x86_64-expensive-checks-debian/llvm-project/llvm/test/CodeGen/RISCV/stack-clash-prologue-nounwind.ll -check-prefix=RV64I # After Prologue/Epilogue Insertion & Frame Finalization # Machine code for function f2: NoPHIs, TracksLiveness, NoVRegs, TiedOpsRewritten, TracksDebugUserValues Frame Objects: fi#0: size=65536, align=1, at location [SP-65536] fi#1: size=8, align=8, at location [SP-65552] bb.0.entry: successors: %bb.1(0x8000); %bb.1(100.00%) $x10 = frame-setup LUI 16 $x6 = frame-setup SUB $x2, killed $x10 frame-setup CFI_INSTRUCTION def_cfa $x6, 65536 $x7 = frame-setup LUI 1 bb.1.entry: ; predecessors: %bb.1, %bb.0 successors: %bb.2(0x4000), %bb.1(0x4000); %bb.2(50.00%), %bb.1(50.00%) $x2 = frame-setup SUB $x2, $x7 frame-setup SD $x0, $x2, 0 frame-setup BNE $x2, $x6, %bb.1 bb.2.entry: ; predecessors: %bb.1 frame-setup CFI_INSTRUCTION def_cfa_register $x2 $x2 = frame-setup ADDI $x2, -16 frame-setup CFI_INSTRUCTION def_cfa_offset 65552 renamable $x10 = ADDI $x0, 3 SB killed renamable $x10, $x2, 16 :: (volatile store (s8) into %ir.a) renamable $x10 = LBU $x2, 16 :: (volatile load (s8) from %ir.a) $x11 = frame-destroy LUI 16 $x11 = frame-destroy ADDIW killed $x11, 16 $x2 = frame-destroy ADD $x2, killed $x11 frame-destroy CFI_INSTRUCTION def_cfa_offset 0 PseudoRET implicit $x10 # End machine code for function f2. *** Bad machine code: Using an undefined physical register *** - function:f2 ... ``` https://github.com/llvm/llvm-project/pull/117612 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [HLSL] Implement SV_GroupThreadId semantic (PR #117781)
https://github.com/lizhengxing updated
https://github.com/llvm/llvm-project/pull/117781
>From e66f451e1db4c321acbfac578fa57f286af2fe27 Mon Sep 17 00:00:00 2001
From: Zhengxing Li
Date: Wed, 13 Nov 2024 10:54:16 -0800
Subject: [PATCH 1/4] [HLSL] Implement SV_GroupThreadId semantic
Support SV_GroupThreadId attribute.
Translate it into dx.thread.id.in.group in clang codeGen.
Fixes: #70122
---
clang/include/clang/Basic/Attr.td | 7
clang/include/clang/Basic/AttrDocs.td | 11 +++
clang/include/clang/Sema/SemaHLSL.h | 1 +
clang/lib/CodeGen/CGHLSLRuntime.cpp | 5 +++
clang/lib/Parse/ParseHLSL.cpp | 1 +
clang/lib/Sema/SemaDeclAttr.cpp | 3 ++
clang/lib/Sema/SemaHLSL.cpp | 10 ++
.../semantics/SV_GroupThreadID.hlsl | 32 +++
.../SemaHLSL/Semantics/entry_parameter.hlsl | 13 +---
.../Semantics/invalid_entry_parameter.hlsl| 22 +
.../Semantics/valid_entry_parameter.hlsl | 25 +++
11 files changed, 125 insertions(+), 5 deletions(-)
create mode 100644 clang/test/CodeGenHLSL/semantics/SV_GroupThreadID.hlsl
diff --git a/clang/include/clang/Basic/Attr.td
b/clang/include/clang/Basic/Attr.td
index 17fc36fbe2ac8c..90d2a2056fe1ba 100644
--- a/clang/include/clang/Basic/Attr.td
+++ b/clang/include/clang/Basic/Attr.td
@@ -4651,6 +4651,13 @@ def HLSLNumThreads: InheritableAttr {
let Documentation = [NumThreadsDocs];
}
+def HLSLSV_GroupThreadID: HLSLAnnotationAttr {
+ let Spellings = [HLSLAnnotation<"SV_GroupThreadID">];
+ let Subjects = SubjectList<[ParmVar, Field]>;
+ let LangOpts = [HLSL];
+ let Documentation = [HLSLSV_GroupThreadIDDocs];
+}
+
def HLSLSV_GroupID: HLSLAnnotationAttr {
let Spellings = [HLSLAnnotation<"SV_GroupID">];
let Subjects = SubjectList<[ParmVar, Field]>;
diff --git a/clang/include/clang/Basic/AttrDocs.td
b/clang/include/clang/Basic/AttrDocs.td
index 7a82b8fa320590..fdad4c9a3ea191 100644
--- a/clang/include/clang/Basic/AttrDocs.td
+++ b/clang/include/clang/Basic/AttrDocs.td
@@ -7941,6 +7941,17 @@ randomized.
}];
}
+def HLSLSV_GroupThreadIDDocs : Documentation {
+ let Category = DocCatFunction;
+ let Content = [{
+The ``SV_GroupThreadID`` semantic, when applied to an input parameter,
specifies which
+individual thread within a thread group is executing in. This attribute is
+only supported in compute shaders.
+
+The full documentation is available here:
https://docs.microsoft.com/en-us/windows/win32/direct3dhlsl/sv-groupthreadid
+ }];
+}
+
def HLSLSV_GroupIDDocs : Documentation {
let Category = DocCatFunction;
let Content = [{
diff --git a/clang/include/clang/Sema/SemaHLSL.h
b/clang/include/clang/Sema/SemaHLSL.h
index ee685d95c96154..f4cd11f423a84a 100644
--- a/clang/include/clang/Sema/SemaHLSL.h
+++ b/clang/include/clang/Sema/SemaHLSL.h
@@ -119,6 +119,7 @@ class SemaHLSL : public SemaBase {
void handleNumThreadsAttr(Decl *D, const ParsedAttr &AL);
void handleWaveSizeAttr(Decl *D, const ParsedAttr &AL);
void handleSV_DispatchThreadIDAttr(Decl *D, const ParsedAttr &AL);
+ void handleSV_GroupThreadIDAttr(Decl *D, const ParsedAttr &AL);
void handleSV_GroupIDAttr(Decl *D, const ParsedAttr &AL);
void handlePackOffsetAttr(Decl *D, const ParsedAttr &AL);
void handleShaderAttr(Decl *D, const ParsedAttr &AL);
diff --git a/clang/lib/CodeGen/CGHLSLRuntime.cpp
b/clang/lib/CodeGen/CGHLSLRuntime.cpp
index 2c293523fca8ca..19db7faddaeac0 100644
--- a/clang/lib/CodeGen/CGHLSLRuntime.cpp
+++ b/clang/lib/CodeGen/CGHLSLRuntime.cpp
@@ -389,6 +389,11 @@ llvm::Value *CGHLSLRuntime::emitInputSemantic(IRBuilder<>
&B,
CGM.getIntrinsic(getThreadIdIntrinsic());
return buildVectorInput(B, ThreadIDIntrinsic, Ty);
}
+ if (D.hasAttr()) {
+llvm::Function *GroupThreadIDIntrinsic =
+CGM.getIntrinsic(Intrinsic::dx_thread_id_in_group);
+return buildVectorInput(B, GroupThreadIDIntrinsic, Ty);
+ }
if (D.hasAttr()) {
llvm::Function *GroupIDIntrinsic =
CGM.getIntrinsic(Intrinsic::dx_group_id);
return buildVectorInput(B, GroupIDIntrinsic, Ty);
diff --git a/clang/lib/Parse/ParseHLSL.cpp b/clang/lib/Parse/ParseHLSL.cpp
index 4de342b63ed802..443bf2b9ec626a 100644
--- a/clang/lib/Parse/ParseHLSL.cpp
+++ b/clang/lib/Parse/ParseHLSL.cpp
@@ -280,6 +280,7 @@ void Parser::ParseHLSLAnnotations(ParsedAttributes &Attrs,
case ParsedAttr::UnknownAttribute:
Diag(Loc, diag::err_unknown_hlsl_semantic) << II;
return;
+ case ParsedAttr::AT_HLSLSV_GroupThreadID:
case ParsedAttr::AT_HLSLSV_GroupID:
case ParsedAttr::AT_HLSLSV_GroupIndex:
case ParsedAttr::AT_HLSLSV_DispatchThreadID:
diff --git a/clang/lib/Sema/SemaDeclAttr.cpp b/clang/lib/Sema/SemaDeclAttr.cpp
index 4fd8ef6dbebf84..5d7ee097383771 100644
--- a/clang/lib/Sema/SemaDeclAttr.cpp
+++ b/clang/lib/Sema/SemaDeclAttr.cpp
@@ -7114,6 +7114,9 @@ ProcessDeclAttribute(Sema &S, Scope *scope, Decl *D,
const Parsed
[clang] ffb19f4 - [CIR] Infrastructure: class CIRGenBuilderTy; cache CIR types (#119037)
Author: David Olsen
Date: 2024-12-10T11:29:48-08:00
New Revision: ffb19f4018e38ba7ff034b78914d5a8d2890a603
URL:
https://github.com/llvm/llvm-project/commit/ffb19f4018e38ba7ff034b78914d5a8d2890a603
DIFF:
https://github.com/llvm/llvm-project/commit/ffb19f4018e38ba7ff034b78914d5a8d2890a603.diff
LOG: [CIR] Infrastructure: class CIRGenBuilderTy; cache CIR types (#119037)
Small infrastructure and background changes to ClangIR.
Create class `CIRGenBuilderTy` and its base class `CIRBaseBuilderTy`.
These are mostly empty for now, except for what is inherited from
`mlir::OpBuilder`. But they will fill up quickly as more ClangIR code
gen is upstreamed. `CIRGenModule` and `CIRGenTypes` are changed to use
`CIRGenBuilderTy`.
Add cached types to struct `CIRGenTypeCache` for the integral types that
are currently supported. Initialize those cached types in the
`CIRGenModule` constructor. The first uses of those types (well, one of
them) is in `CIRGenTypes::convertType`.
Have `CIRGenTypes::convertType` cache its results in a map from
`clang::Type` to `mlir::Type`, saving it from having to convert the same
type again and again.
There are no new tests or changed tests in this commit. There are no
changes to behavior, just improvements to how the existing behavior is
implemented.
Added:
clang/include/clang/CIR/Dialect/Builder/CIRBaseBuilder.h
clang/lib/CIR/CodeGen/CIRGenBuilder.h
Modified:
clang/lib/CIR/CodeGen/CIRGenModule.cpp
clang/lib/CIR/CodeGen/CIRGenModule.h
clang/lib/CIR/CodeGen/CIRGenTypeCache.h
clang/lib/CIR/CodeGen/CIRGenTypes.cpp
clang/lib/CIR/CodeGen/CIRGenTypes.h
Removed:
diff --git a/clang/include/clang/CIR/Dialect/Builder/CIRBaseBuilder.h
b/clang/include/clang/CIR/Dialect/Builder/CIRBaseBuilder.h
new file mode 100644
index 00..75ae74e926fbc6
--- /dev/null
+++ b/clang/include/clang/CIR/Dialect/Builder/CIRBaseBuilder.h
@@ -0,0 +1,25 @@
+//===--===//
+//
+// 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
+//
+//===--===//
+
+#ifndef LLVM_CLANG_CIR_DIALECT_BUILDER_CIRBASEBUILDER_H
+#define LLVM_CLANG_CIR_DIALECT_BUILDER_CIRBASEBUILDER_H
+
+#include "mlir/IR/Builders.h"
+
+namespace cir {
+
+class CIRBaseBuilderTy : public mlir::OpBuilder {
+
+public:
+ CIRBaseBuilderTy(mlir::MLIRContext &mlirContext)
+ : mlir::OpBuilder(&mlirContext) {}
+};
+
+} // namespace cir
+
+#endif
diff --git a/clang/lib/CIR/CodeGen/CIRGenBuilder.h
b/clang/lib/CIR/CodeGen/CIRGenBuilder.h
new file mode 100644
index 00..92115778518d4a
--- /dev/null
+++ b/clang/lib/CIR/CodeGen/CIRGenBuilder.h
@@ -0,0 +1,28 @@
+//===--===//
+//
+// 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
+//
+//===--===//
+
+#ifndef LLVM_CLANG_LIB_CIR_CODEGEN_CIRGENBUILDER_H
+#define LLVM_CLANG_LIB_CIR_CODEGEN_CIRGENBUILDER_H
+
+#include "CIRGenTypeCache.h"
+
+#include "clang/CIR/Dialect/Builder/CIRBaseBuilder.h"
+
+namespace clang::CIRGen {
+
+class CIRGenBuilderTy : public cir::CIRBaseBuilderTy {
+ const CIRGenTypeCache &typeCache;
+
+public:
+ CIRGenBuilderTy(mlir::MLIRContext &mlirContext, const CIRGenTypeCache &tc)
+ : CIRBaseBuilderTy(mlirContext), typeCache(tc) {}
+};
+
+} // namespace clang::CIRGen
+
+#endif
diff --git a/clang/lib/CIR/CodeGen/CIRGenModule.cpp
b/clang/lib/CIR/CodeGen/CIRGenModule.cpp
index b44f66493254f2..e7c9512dcd3de8 100644
--- a/clang/lib/CIR/CodeGen/CIRGenModule.cpp
+++ b/clang/lib/CIR/CodeGen/CIRGenModule.cpp
@@ -29,9 +29,22 @@ CIRGenModule::CIRGenModule(mlir::MLIRContext &context,
clang::ASTContext &astctx,
const clang::CodeGenOptions &cgo,
DiagnosticsEngine &diags)
-: builder(&context), astCtx(astctx), langOpts(astctx.getLangOpts()),
+: builder(context, *this), astCtx(astctx), langOpts(astctx.getLangOpts()),
theModule{mlir::ModuleOp::create(mlir::UnknownLoc::get(&context))},
- diags(diags), target(astCtx.getTargetInfo()), genTypes(*this) {}
+ diags(diags), target(astctx.getTargetInfo()), genTypes(*this) {
+
+ // Initialize cached types
+ SInt8Ty = cir::IntType::get(&getMLIRContext(), 8, /*isSigned=*/true);
+ SInt16Ty = cir::IntType::get(&getMLIRContext(), 16, /*isSigned=*/true);
+ SInt32Ty = cir::IntType::get(&getMLIRContext(), 32, /*isSigned=*/true);
+ SInt64Ty = cir::IntType::get(&getM
[clang] [llvm] [NVPTX] Add NVVMUpgradeAnnotations pass to cleanup legacy annotations (PR #119261)
@@ -1270,77 +1270,21 @@ exit:
; MODULE: attributes #[[ATTR1:[0-9]+]] = { convergent nocallback nounwind }
; MODULE: attributes #[[ATTR2:[0-9]+]] = { convergent nocallback nofree
nounwind willreturn }
; MODULE: attributes #[[ATTR3:[0-9]+]] = { nocallback nofree nosync nounwind
willreturn memory(inaccessiblemem: write) }
-; MODULE: attributes #[[ATTR4]] = { "kernel" }
-; MODULE: attributes #[[ATTR5]] = { nosync memory(none) }
+; MODULE: attributes #[[ATTR4]] = { "kernel" "nvvm.kernel" }
+; MODULE: attributes #[[ATTR5]] = { "kernel" }
+; MODULE: attributes #[[ATTR6]] = { nosync memory(none) }
;.
; CGSCC: attributes #[[ATTR0]] = { "llvm.assume"="ompx_aligned_barrier" }
; CGSCC: attributes #[[ATTR1:[0-9]+]] = { convergent nocallback nounwind }
; CGSCC: attributes #[[ATTR2:[0-9]+]] = { convergent nocallback nofree
nounwind willreturn }
; CGSCC: attributes #[[ATTR3:[0-9]+]] = { nocallback nofree nosync nounwind
willreturn memory(inaccessiblemem: write) }
-; CGSCC: attributes #[[ATTR4]] = { "kernel" }
-; CGSCC: attributes #[[ATTR5]] = { nosync memory(none) }
+; CGSCC: attributes #[[ATTR4]] = { "kernel" "nvvm.kernel" }
AlexMaclean wrote:
Unfortunately, I think we do. "kernel" is really more like "OpenMP kernel" and
the semantics for this do not seem to be a perfect match for "nvvm.kernel". For
example, `@multiple_blocks_functions_non_kernel_effects_2` in this test has
"kernel" but is not an nvvm kernel. I'm vary unfamiliar with the OpenMP
semantics so I thought keeping it separate would be the safest approach, it
also may be clearest to have a common "nvvm.*" prefix for all attributes
currently represented as nvvm.annotations.
https://github.com/llvm/llvm-project/pull/119261
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [NVPTX] Add NVVMUpgradeAnnotations pass to cleanup legacy annotations (PR #119261)
@@ -5022,6 +5022,69 @@ bool llvm::UpgradeDebugInfo(Module &M) {
return Modified;
}
+bool static upgradeSingleNVVMAnnotation(GlobalValue *GV, StringRef K,
+const Metadata *V) {
+ if (K == "kernel") {
+assert(mdconst::extract(V)->getZExtValue() == 1);
Artem-B wrote:
Assert should not be used here as it's not an LLVM error. It's an input error
and should be diagnosed (or accepted/ignored). Considering that the value is a
flag, and not very useful one at that (I don't think `kernel 0` is ever
actually used, it would be OK to accept a non-zero value as "it's a kernel" and
zero as "not a kernel".
https://github.com/llvm/llvm-project/pull/119261
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] Revert "[PAC][ELF][AArch64] Support signed personality function pointer" (PR #119331)
llvm-ci wrote: LLVM Buildbot has detected a new failure on builder `lld-x86_64-win` running on `as-worker-93` while building `clang,llvm` at step 7 "test-build-unified-tree-check-all". Full details are available at: https://lab.llvm.org/buildbot/#/builders/146/builds/1803 Here is the relevant piece of the build log for the reference ``` Step 7 (test-build-unified-tree-check-all) failure: test (failure) TEST 'LLVM-Unit :: Support/./SupportTests.exe/38/87' FAILED Script(shard): -- GTEST_OUTPUT=json:C:\a\lld-x86_64-win\build\unittests\Support\.\SupportTests.exe-LLVM-Unit-21236-38-87.json GTEST_SHUFFLE=0 GTEST_TOTAL_SHARDS=87 GTEST_SHARD_INDEX=38 C:\a\lld-x86_64-win\build\unittests\Support\.\SupportTests.exe -- Script: -- C:\a\lld-x86_64-win\build\unittests\Support\.\SupportTests.exe --gtest_filter=ProgramEnvTest.CreateProcessLongPath -- C:\a\lld-x86_64-win\llvm-project\llvm\unittests\Support\ProgramTest.cpp(160): error: Expected equality of these values: 0 RC Which is: -2 C:\a\lld-x86_64-win\llvm-project\llvm\unittests\Support\ProgramTest.cpp(163): error: fs::remove(Twine(LongPath)): did not return errc::success. error number: 13 error message: permission denied C:\a\lld-x86_64-win\llvm-project\llvm\unittests\Support\ProgramTest.cpp:160 Expected equality of these values: 0 RC Which is: -2 C:\a\lld-x86_64-win\llvm-project\llvm\unittests\Support\ProgramTest.cpp:163 fs::remove(Twine(LongPath)): did not return errc::success. error number: 13 error message: permission denied ``` https://github.com/llvm/llvm-project/pull/119331 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] Cache SanitizerArgs in Clang driver (NFC) (PR #119442)
https://github.com/aaronpuchert created
https://github.com/llvm/llvm-project/pull/119442
The name getSanitizerArgs seems to mislead callers that this is a cheap
function, but it extracts the SanitizerArgs each time it is called.
So we try to reuse it a bit more.
>From fdcd8a22e13f0feb665b76a801dbdd31fe9354f5 Mon Sep 17 00:00:00 2001
From: Aaron Puchert
Date: Fri, 6 Dec 2024 23:13:17 +0100
Subject: [PATCH] Cache SanitizerArgs in Clang driver (NFC)
The name getSanitizerArgs seems to mislead callers that this is a cheap
function, but it extracts the SanitizerArgs each time it is called.
So we try to reuse it a bit more.
---
clang/lib/Driver/ToolChains/CommonArgs.cpp | 9 +
clang/lib/Driver/ToolChains/CommonArgs.h | 2 ++
clang/lib/Driver/ToolChains/FreeBSD.cpp| 6 --
clang/lib/Driver/ToolChains/Fuchsia.cpp| 10 +-
clang/lib/Driver/ToolChains/Gnu.cpp| 7 +--
clang/lib/Driver/ToolChains/Hexagon.cpp| 9 ++---
clang/lib/Driver/ToolChains/NetBSD.cpp | 7 ---
clang/lib/Driver/ToolChains/OpenBSD.cpp| 6 --
clang/lib/Driver/ToolChains/Solaris.cpp| 6 +++---
9 files changed, 38 insertions(+), 24 deletions(-)
diff --git a/clang/lib/Driver/ToolChains/CommonArgs.cpp
b/clang/lib/Driver/ToolChains/CommonArgs.cpp
index 03dbdc27975b42..6702032172ddd3 100644
--- a/clang/lib/Driver/ToolChains/CommonArgs.cpp
+++ b/clang/lib/Driver/ToolChains/CommonArgs.cpp
@@ -1384,6 +1384,7 @@ void tools::addAsNeededOption(const ToolChain &TC,
void tools::linkSanitizerRuntimeDeps(const ToolChain &TC,
const llvm::opt::ArgList &Args,
+ const SanitizerArgs &SanArgs,
ArgStringList &CmdArgs) {
// Force linking against the system libraries sanitizers depends on
// (see PR15823 why this is necessary).
@@ -1410,18 +1411,18 @@ void tools::linkSanitizerRuntimeDeps(const ToolChain
&TC,
// libresolv.a, even if exists, is an empty archive to satisfy POSIX -lresolv
// requirement.
if (TC.getTriple().isOSLinux() && !TC.getTriple().isAndroid() &&
- !TC.getTriple().isMusl() && TC.getSanitizerArgs(Args).needsMsanRt())
+ !TC.getTriple().isMusl() && SanArgs.needsMsanRt())
CmdArgs.push_back("-lresolv");
}
static void
collectSanitizerRuntimes(const ToolChain &TC, const ArgList &Args,
+ const SanitizerArgs &SanArgs,
SmallVectorImpl &SharedRuntimes,
SmallVectorImpl &StaticRuntimes,
SmallVectorImpl &NonWholeStaticRuntimes,
SmallVectorImpl &HelperStaticRuntimes,
SmallVectorImpl &RequiredSymbols) {
- const SanitizerArgs &SanArgs = TC.getSanitizerArgs(Args);
// Collect shared runtimes.
if (SanArgs.needsSharedRt()) {
if (SanArgs.needsAsanRt()) {
@@ -1555,12 +1556,12 @@ collectSanitizerRuntimes(const ToolChain &TC, const
ArgList &Args,
// Should be called before we add system libraries (C++ ABI, libstdc++/libc++,
// C runtime, etc). Returns true if sanitizer system deps need to be linked in.
bool tools::addSanitizerRuntimes(const ToolChain &TC, const ArgList &Args,
+ const SanitizerArgs &SanArgs,
ArgStringList &CmdArgs) {
- const SanitizerArgs &SanArgs = TC.getSanitizerArgs(Args);
SmallVector SharedRuntimes, StaticRuntimes,
NonWholeStaticRuntimes, HelperStaticRuntimes, RequiredSymbols;
if (SanArgs.linkRuntimes()) {
-collectSanitizerRuntimes(TC, Args, SharedRuntimes, StaticRuntimes,
+collectSanitizerRuntimes(TC, Args, SanArgs, SharedRuntimes, StaticRuntimes,
NonWholeStaticRuntimes, HelperStaticRuntimes,
RequiredSymbols);
}
diff --git a/clang/lib/Driver/ToolChains/CommonArgs.h
b/clang/lib/Driver/ToolChains/CommonArgs.h
index b6ddd99b872798..de2d143b904790 100644
--- a/clang/lib/Driver/ToolChains/CommonArgs.h
+++ b/clang/lib/Driver/ToolChains/CommonArgs.h
@@ -38,10 +38,12 @@ void addLinkerCompressDebugSectionsOption(const ToolChain
&TC,
void claimNoWarnArgs(const llvm::opt::ArgList &Args);
bool addSanitizerRuntimes(const ToolChain &TC, const llvm::opt::ArgList &Args,
+ const SanitizerArgs &SanArgs,
llvm::opt::ArgStringList &CmdArgs);
void linkSanitizerRuntimeDeps(const ToolChain &TC,
const llvm::opt::ArgList &Args,
+ const SanitizerArgs &SanArgs,
llvm::opt::ArgStringList &CmdArgs);
bool addXRayRuntime(const ToolChain &TC, const llvm::opt::ArgList &Args,
diff --git a/clang/lib/Driver/ToolChains/FreeBSD.cpp
b/clang/lib/Driver/ToolChains/FreeBSD.cpp
index 3d744bc087f467..c78f4e26a2f107 100644
--- a/clang/lib/Driver/ToolChains/FreeBSD.cpp
+++ b/clang/lib/Driver/ToolChains/Fre
[clang] Cache SanitizerArgs in Clang driver (NFC) (PR #119442)
llvmbot wrote:
@llvm/pr-subscribers-clang
Author: Aaron Puchert (aaronpuchert)
Changes
The name getSanitizerArgs seems to mislead callers that this is a cheap
function, but it extracts the SanitizerArgs each time it is called.
So we try to reuse it a bit more.
---
Full diff: https://github.com/llvm/llvm-project/pull/119442.diff
9 Files Affected:
- (modified) clang/lib/Driver/ToolChains/CommonArgs.cpp (+5-4)
- (modified) clang/lib/Driver/ToolChains/CommonArgs.h (+2)
- (modified) clang/lib/Driver/ToolChains/FreeBSD.cpp (+4-2)
- (modified) clang/lib/Driver/ToolChains/Fuchsia.cpp (+5-5)
- (modified) clang/lib/Driver/ToolChains/Gnu.cpp (+5-2)
- (modified) clang/lib/Driver/ToolChains/Hexagon.cpp (+6-3)
- (modified) clang/lib/Driver/ToolChains/NetBSD.cpp (+4-3)
- (modified) clang/lib/Driver/ToolChains/OpenBSD.cpp (+4-2)
- (modified) clang/lib/Driver/ToolChains/Solaris.cpp (+3-3)
``diff
diff --git a/clang/lib/Driver/ToolChains/CommonArgs.cpp
b/clang/lib/Driver/ToolChains/CommonArgs.cpp
index 03dbdc27975b42..6702032172ddd3 100644
--- a/clang/lib/Driver/ToolChains/CommonArgs.cpp
+++ b/clang/lib/Driver/ToolChains/CommonArgs.cpp
@@ -1384,6 +1384,7 @@ void tools::addAsNeededOption(const ToolChain &TC,
void tools::linkSanitizerRuntimeDeps(const ToolChain &TC,
const llvm::opt::ArgList &Args,
+ const SanitizerArgs &SanArgs,
ArgStringList &CmdArgs) {
// Force linking against the system libraries sanitizers depends on
// (see PR15823 why this is necessary).
@@ -1410,18 +1411,18 @@ void tools::linkSanitizerRuntimeDeps(const ToolChain
&TC,
// libresolv.a, even if exists, is an empty archive to satisfy POSIX -lresolv
// requirement.
if (TC.getTriple().isOSLinux() && !TC.getTriple().isAndroid() &&
- !TC.getTriple().isMusl() && TC.getSanitizerArgs(Args).needsMsanRt())
+ !TC.getTriple().isMusl() && SanArgs.needsMsanRt())
CmdArgs.push_back("-lresolv");
}
static void
collectSanitizerRuntimes(const ToolChain &TC, const ArgList &Args,
+ const SanitizerArgs &SanArgs,
SmallVectorImpl &SharedRuntimes,
SmallVectorImpl &StaticRuntimes,
SmallVectorImpl &NonWholeStaticRuntimes,
SmallVectorImpl &HelperStaticRuntimes,
SmallVectorImpl &RequiredSymbols) {
- const SanitizerArgs &SanArgs = TC.getSanitizerArgs(Args);
// Collect shared runtimes.
if (SanArgs.needsSharedRt()) {
if (SanArgs.needsAsanRt()) {
@@ -1555,12 +1556,12 @@ collectSanitizerRuntimes(const ToolChain &TC, const
ArgList &Args,
// Should be called before we add system libraries (C++ ABI, libstdc++/libc++,
// C runtime, etc). Returns true if sanitizer system deps need to be linked in.
bool tools::addSanitizerRuntimes(const ToolChain &TC, const ArgList &Args,
+ const SanitizerArgs &SanArgs,
ArgStringList &CmdArgs) {
- const SanitizerArgs &SanArgs = TC.getSanitizerArgs(Args);
SmallVector SharedRuntimes, StaticRuntimes,
NonWholeStaticRuntimes, HelperStaticRuntimes, RequiredSymbols;
if (SanArgs.linkRuntimes()) {
-collectSanitizerRuntimes(TC, Args, SharedRuntimes, StaticRuntimes,
+collectSanitizerRuntimes(TC, Args, SanArgs, SharedRuntimes, StaticRuntimes,
NonWholeStaticRuntimes, HelperStaticRuntimes,
RequiredSymbols);
}
diff --git a/clang/lib/Driver/ToolChains/CommonArgs.h
b/clang/lib/Driver/ToolChains/CommonArgs.h
index b6ddd99b872798..de2d143b904790 100644
--- a/clang/lib/Driver/ToolChains/CommonArgs.h
+++ b/clang/lib/Driver/ToolChains/CommonArgs.h
@@ -38,10 +38,12 @@ void addLinkerCompressDebugSectionsOption(const ToolChain
&TC,
void claimNoWarnArgs(const llvm::opt::ArgList &Args);
bool addSanitizerRuntimes(const ToolChain &TC, const llvm::opt::ArgList &Args,
+ const SanitizerArgs &SanArgs,
llvm::opt::ArgStringList &CmdArgs);
void linkSanitizerRuntimeDeps(const ToolChain &TC,
const llvm::opt::ArgList &Args,
+ const SanitizerArgs &SanArgs,
llvm::opt::ArgStringList &CmdArgs);
bool addXRayRuntime(const ToolChain &TC, const llvm::opt::ArgList &Args,
diff --git a/clang/lib/Driver/ToolChains/FreeBSD.cpp
b/clang/lib/Driver/ToolChains/FreeBSD.cpp
index 3d744bc087f467..c78f4e26a2f107 100644
--- a/clang/lib/Driver/ToolChains/FreeBSD.cpp
+++ b/clang/lib/Driver/ToolChains/FreeBSD.cpp
@@ -283,7 +283,9 @@ void freebsd::Linker::ConstructJob(Compilation &C, const
JobAction &JA,
D.getLTOMode() == LTOK_Thin);
}
- bool NeedsSanitizerDeps = addSanitizerRuntimes(ToolChain, Args, CmdArgs);
+ const SanitizerArgs &SanArgs = ToolChai
[clang] [llvm] [NVPTX] Add NVVMUpgradeAnnotations pass to cleanup legacy annotations (PR #119261)
@@ -5022,6 +5022,69 @@ bool llvm::UpgradeDebugInfo(Module &M) {
return Modified;
}
+bool static upgradeSingleNVVMAnnotation(GlobalValue *GV, StringRef K,
+const Metadata *V) {
+ if (K == "kernel") {
+assert(mdconst::extract(V)->getZExtValue() == 1);
+cast(GV)->addFnAttr("nvvm.kernel");
AlexMaclean wrote:
Some annotations (such as "texture") are applied to global variables, not
functions. I cannot unconditionally cast to a Function until confirming the
annotation kind.
https://github.com/llvm/llvm-project/pull/119261
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] Cache SanitizerArgs in Clang driver (NFC) (PR #119442)
llvmbot wrote:
@llvm/pr-subscribers-clang-driver
Author: Aaron Puchert (aaronpuchert)
Changes
The name getSanitizerArgs seems to mislead callers that this is a cheap
function, but it extracts the SanitizerArgs each time it is called.
So we try to reuse it a bit more.
---
Full diff: https://github.com/llvm/llvm-project/pull/119442.diff
9 Files Affected:
- (modified) clang/lib/Driver/ToolChains/CommonArgs.cpp (+5-4)
- (modified) clang/lib/Driver/ToolChains/CommonArgs.h (+2)
- (modified) clang/lib/Driver/ToolChains/FreeBSD.cpp (+4-2)
- (modified) clang/lib/Driver/ToolChains/Fuchsia.cpp (+5-5)
- (modified) clang/lib/Driver/ToolChains/Gnu.cpp (+5-2)
- (modified) clang/lib/Driver/ToolChains/Hexagon.cpp (+6-3)
- (modified) clang/lib/Driver/ToolChains/NetBSD.cpp (+4-3)
- (modified) clang/lib/Driver/ToolChains/OpenBSD.cpp (+4-2)
- (modified) clang/lib/Driver/ToolChains/Solaris.cpp (+3-3)
``diff
diff --git a/clang/lib/Driver/ToolChains/CommonArgs.cpp
b/clang/lib/Driver/ToolChains/CommonArgs.cpp
index 03dbdc27975b42..6702032172ddd3 100644
--- a/clang/lib/Driver/ToolChains/CommonArgs.cpp
+++ b/clang/lib/Driver/ToolChains/CommonArgs.cpp
@@ -1384,6 +1384,7 @@ void tools::addAsNeededOption(const ToolChain &TC,
void tools::linkSanitizerRuntimeDeps(const ToolChain &TC,
const llvm::opt::ArgList &Args,
+ const SanitizerArgs &SanArgs,
ArgStringList &CmdArgs) {
// Force linking against the system libraries sanitizers depends on
// (see PR15823 why this is necessary).
@@ -1410,18 +1411,18 @@ void tools::linkSanitizerRuntimeDeps(const ToolChain
&TC,
// libresolv.a, even if exists, is an empty archive to satisfy POSIX -lresolv
// requirement.
if (TC.getTriple().isOSLinux() && !TC.getTriple().isAndroid() &&
- !TC.getTriple().isMusl() && TC.getSanitizerArgs(Args).needsMsanRt())
+ !TC.getTriple().isMusl() && SanArgs.needsMsanRt())
CmdArgs.push_back("-lresolv");
}
static void
collectSanitizerRuntimes(const ToolChain &TC, const ArgList &Args,
+ const SanitizerArgs &SanArgs,
SmallVectorImpl &SharedRuntimes,
SmallVectorImpl &StaticRuntimes,
SmallVectorImpl &NonWholeStaticRuntimes,
SmallVectorImpl &HelperStaticRuntimes,
SmallVectorImpl &RequiredSymbols) {
- const SanitizerArgs &SanArgs = TC.getSanitizerArgs(Args);
// Collect shared runtimes.
if (SanArgs.needsSharedRt()) {
if (SanArgs.needsAsanRt()) {
@@ -1555,12 +1556,12 @@ collectSanitizerRuntimes(const ToolChain &TC, const
ArgList &Args,
// Should be called before we add system libraries (C++ ABI, libstdc++/libc++,
// C runtime, etc). Returns true if sanitizer system deps need to be linked in.
bool tools::addSanitizerRuntimes(const ToolChain &TC, const ArgList &Args,
+ const SanitizerArgs &SanArgs,
ArgStringList &CmdArgs) {
- const SanitizerArgs &SanArgs = TC.getSanitizerArgs(Args);
SmallVector SharedRuntimes, StaticRuntimes,
NonWholeStaticRuntimes, HelperStaticRuntimes, RequiredSymbols;
if (SanArgs.linkRuntimes()) {
-collectSanitizerRuntimes(TC, Args, SharedRuntimes, StaticRuntimes,
+collectSanitizerRuntimes(TC, Args, SanArgs, SharedRuntimes, StaticRuntimes,
NonWholeStaticRuntimes, HelperStaticRuntimes,
RequiredSymbols);
}
diff --git a/clang/lib/Driver/ToolChains/CommonArgs.h
b/clang/lib/Driver/ToolChains/CommonArgs.h
index b6ddd99b872798..de2d143b904790 100644
--- a/clang/lib/Driver/ToolChains/CommonArgs.h
+++ b/clang/lib/Driver/ToolChains/CommonArgs.h
@@ -38,10 +38,12 @@ void addLinkerCompressDebugSectionsOption(const ToolChain
&TC,
void claimNoWarnArgs(const llvm::opt::ArgList &Args);
bool addSanitizerRuntimes(const ToolChain &TC, const llvm::opt::ArgList &Args,
+ const SanitizerArgs &SanArgs,
llvm::opt::ArgStringList &CmdArgs);
void linkSanitizerRuntimeDeps(const ToolChain &TC,
const llvm::opt::ArgList &Args,
+ const SanitizerArgs &SanArgs,
llvm::opt::ArgStringList &CmdArgs);
bool addXRayRuntime(const ToolChain &TC, const llvm::opt::ArgList &Args,
diff --git a/clang/lib/Driver/ToolChains/FreeBSD.cpp
b/clang/lib/Driver/ToolChains/FreeBSD.cpp
index 3d744bc087f467..c78f4e26a2f107 100644
--- a/clang/lib/Driver/ToolChains/FreeBSD.cpp
+++ b/clang/lib/Driver/ToolChains/FreeBSD.cpp
@@ -283,7 +283,9 @@ void freebsd::Linker::ConstructJob(Compilation &C, const
JobAction &JA,
D.getLTOMode() == LTOK_Thin);
}
- bool NeedsSanitizerDeps = addSanitizerRuntimes(ToolChain, Args, CmdArgs);
+ const SanitizerArgs &SanArgs = T
[clang] Cache SanitizerArgs in Clang driver (NFC) (PR #119442)
llvmbot wrote:
@llvm/pr-subscribers-backend-hexagon
Author: Aaron Puchert (aaronpuchert)
Changes
The name getSanitizerArgs seems to mislead callers that this is a cheap
function, but it extracts the SanitizerArgs each time it is called.
So we try to reuse it a bit more.
---
Full diff: https://github.com/llvm/llvm-project/pull/119442.diff
9 Files Affected:
- (modified) clang/lib/Driver/ToolChains/CommonArgs.cpp (+5-4)
- (modified) clang/lib/Driver/ToolChains/CommonArgs.h (+2)
- (modified) clang/lib/Driver/ToolChains/FreeBSD.cpp (+4-2)
- (modified) clang/lib/Driver/ToolChains/Fuchsia.cpp (+5-5)
- (modified) clang/lib/Driver/ToolChains/Gnu.cpp (+5-2)
- (modified) clang/lib/Driver/ToolChains/Hexagon.cpp (+6-3)
- (modified) clang/lib/Driver/ToolChains/NetBSD.cpp (+4-3)
- (modified) clang/lib/Driver/ToolChains/OpenBSD.cpp (+4-2)
- (modified) clang/lib/Driver/ToolChains/Solaris.cpp (+3-3)
``diff
diff --git a/clang/lib/Driver/ToolChains/CommonArgs.cpp
b/clang/lib/Driver/ToolChains/CommonArgs.cpp
index 03dbdc27975b42..6702032172ddd3 100644
--- a/clang/lib/Driver/ToolChains/CommonArgs.cpp
+++ b/clang/lib/Driver/ToolChains/CommonArgs.cpp
@@ -1384,6 +1384,7 @@ void tools::addAsNeededOption(const ToolChain &TC,
void tools::linkSanitizerRuntimeDeps(const ToolChain &TC,
const llvm::opt::ArgList &Args,
+ const SanitizerArgs &SanArgs,
ArgStringList &CmdArgs) {
// Force linking against the system libraries sanitizers depends on
// (see PR15823 why this is necessary).
@@ -1410,18 +1411,18 @@ void tools::linkSanitizerRuntimeDeps(const ToolChain
&TC,
// libresolv.a, even if exists, is an empty archive to satisfy POSIX -lresolv
// requirement.
if (TC.getTriple().isOSLinux() && !TC.getTriple().isAndroid() &&
- !TC.getTriple().isMusl() && TC.getSanitizerArgs(Args).needsMsanRt())
+ !TC.getTriple().isMusl() && SanArgs.needsMsanRt())
CmdArgs.push_back("-lresolv");
}
static void
collectSanitizerRuntimes(const ToolChain &TC, const ArgList &Args,
+ const SanitizerArgs &SanArgs,
SmallVectorImpl &SharedRuntimes,
SmallVectorImpl &StaticRuntimes,
SmallVectorImpl &NonWholeStaticRuntimes,
SmallVectorImpl &HelperStaticRuntimes,
SmallVectorImpl &RequiredSymbols) {
- const SanitizerArgs &SanArgs = TC.getSanitizerArgs(Args);
// Collect shared runtimes.
if (SanArgs.needsSharedRt()) {
if (SanArgs.needsAsanRt()) {
@@ -1555,12 +1556,12 @@ collectSanitizerRuntimes(const ToolChain &TC, const
ArgList &Args,
// Should be called before we add system libraries (C++ ABI, libstdc++/libc++,
// C runtime, etc). Returns true if sanitizer system deps need to be linked in.
bool tools::addSanitizerRuntimes(const ToolChain &TC, const ArgList &Args,
+ const SanitizerArgs &SanArgs,
ArgStringList &CmdArgs) {
- const SanitizerArgs &SanArgs = TC.getSanitizerArgs(Args);
SmallVector SharedRuntimes, StaticRuntimes,
NonWholeStaticRuntimes, HelperStaticRuntimes, RequiredSymbols;
if (SanArgs.linkRuntimes()) {
-collectSanitizerRuntimes(TC, Args, SharedRuntimes, StaticRuntimes,
+collectSanitizerRuntimes(TC, Args, SanArgs, SharedRuntimes, StaticRuntimes,
NonWholeStaticRuntimes, HelperStaticRuntimes,
RequiredSymbols);
}
diff --git a/clang/lib/Driver/ToolChains/CommonArgs.h
b/clang/lib/Driver/ToolChains/CommonArgs.h
index b6ddd99b872798..de2d143b904790 100644
--- a/clang/lib/Driver/ToolChains/CommonArgs.h
+++ b/clang/lib/Driver/ToolChains/CommonArgs.h
@@ -38,10 +38,12 @@ void addLinkerCompressDebugSectionsOption(const ToolChain
&TC,
void claimNoWarnArgs(const llvm::opt::ArgList &Args);
bool addSanitizerRuntimes(const ToolChain &TC, const llvm::opt::ArgList &Args,
+ const SanitizerArgs &SanArgs,
llvm::opt::ArgStringList &CmdArgs);
void linkSanitizerRuntimeDeps(const ToolChain &TC,
const llvm::opt::ArgList &Args,
+ const SanitizerArgs &SanArgs,
llvm::opt::ArgStringList &CmdArgs);
bool addXRayRuntime(const ToolChain &TC, const llvm::opt::ArgList &Args,
diff --git a/clang/lib/Driver/ToolChains/FreeBSD.cpp
b/clang/lib/Driver/ToolChains/FreeBSD.cpp
index 3d744bc087f467..c78f4e26a2f107 100644
--- a/clang/lib/Driver/ToolChains/FreeBSD.cpp
+++ b/clang/lib/Driver/ToolChains/FreeBSD.cpp
@@ -283,7 +283,9 @@ void freebsd::Linker::ConstructJob(Compilation &C, const
JobAction &JA,
D.getLTOMode() == LTOK_Thin);
}
- bool NeedsSanitizerDeps = addSanitizerRuntimes(ToolChain, Args, CmdArgs);
+ const SanitizerArgs &SanArgs
[clang] [CIR] Infrastructure: class CIRGenBuilderTy; cache CIR types (PR #119037)
https://github.com/dkolsen-pgi closed https://github.com/llvm/llvm-project/pull/119037 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [NVPTX] Add NVVMUpgradeAnnotations pass to cleanup legacy annotations (PR #119261)
https://github.com/AlexMaclean edited https://github.com/llvm/llvm-project/pull/119261 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [flang] [clang][driver] When -fveclib=ArmPL flag is in use, always link against libamath (PR #116432)
https://github.com/pawosm-arm updated
https://github.com/llvm/llvm-project/pull/116432
>From d19c8ea5e02ad92c1f89b129cd8fe2fd3cf1d4e3 Mon Sep 17 00:00:00 2001
From: Paul Osmialowski
Date: Tue, 3 Dec 2024 12:41:15 +
Subject: [PATCH] [clang][driver] When -fveclib=ArmPL flag is in use, always
link against libamath
Using `-fveclib=ArmPL` without libamath likely effects in the
link-time errors.
---
clang/lib/Driver/ToolChains/CommonArgs.cpp | 33 ++
clang/lib/Driver/ToolChains/MSVC.cpp | 6
clang/test/Driver/fveclib.c| 17 +++
flang/test/Driver/fveclib.f90 | 17 +++
4 files changed, 73 insertions(+)
diff --git a/clang/lib/Driver/ToolChains/CommonArgs.cpp
b/clang/lib/Driver/ToolChains/CommonArgs.cpp
index 8d977149e62485..a0d77f254bd8c3 100644
--- a/clang/lib/Driver/ToolChains/CommonArgs.cpp
+++ b/clang/lib/Driver/ToolChains/CommonArgs.cpp
@@ -490,6 +490,39 @@ void tools::AddLinkerInputs(const ToolChain &TC, const
InputInfoList &Inputs,
else
A.renderAsInput(Args, CmdArgs);
}
+ if (const Arg *A = Args.getLastArg(options::OPT_fveclib)) {
+const llvm::Triple &Triple = TC.getTriple();
+StringRef V = A->getValue();
+if (V == "ArmPL" && (Triple.isOSLinux() || Triple.isOSDarwin())) {
+ // To support -fveclib=ArmPL we need to link against libamath. Some of
the
+ // libamath functions depend on libm, at the same time, libamath exports
+ // its own implementation of some of the libm functions. These are faster
+ // and potentially less accurate implementations, hence we need to be
+ // careful what is being linked in. Since here we are interested only in
+ // the subset of libamath functions that is covered by the veclib
+ // mappings, we need to prioritize libm functions by putting -lm before
+ // -lamath (and then -lm again, to fulfill libamath requirements).
+ //
+ // Therefore we need to do the following:
+ //
+ // 1. On Linux, link only when actually needed.
+ //
+ // 2. Prefer libm functions over libamath.
+ //
+ // 3. Link against libm to resolve libamath dependencies.
+ //
+ if (Triple.isOSLinux()) {
+CmdArgs.push_back(Args.MakeArgString("--push-state"));
+CmdArgs.push_back(Args.MakeArgString("--as-needed"));
+ }
+ CmdArgs.push_back(Args.MakeArgString("-lm"));
+ CmdArgs.push_back(Args.MakeArgString("-lamath"));
+ CmdArgs.push_back(Args.MakeArgString("-lm"));
+ if (Triple.isOSLinux())
+CmdArgs.push_back(Args.MakeArgString("--pop-state"));
+ addArchSpecificRPath(TC, Args, CmdArgs);
+}
+ }
}
void tools::addLinkerCompressDebugSectionsOption(
diff --git a/clang/lib/Driver/ToolChains/MSVC.cpp
b/clang/lib/Driver/ToolChains/MSVC.cpp
index 80799d1e715f07..752c2e2751ab67 100644
--- a/clang/lib/Driver/ToolChains/MSVC.cpp
+++ b/clang/lib/Driver/ToolChains/MSVC.cpp
@@ -84,6 +84,12 @@ void visualstudio::Linker::ConstructJob(Compilation &C,
const JobAction &JA,
else if (TC.getTriple().isWindowsArm64EC())
CmdArgs.push_back("-machine:arm64ec");
+ if (const Arg *A = Args.getLastArg(options::OPT_fveclib)) {
+StringRef V = A->getValue();
+if (V == "ArmPL")
+ CmdArgs.push_back(Args.MakeArgString("--dependent-lib=amath"));
+ }
+
if (!Args.hasArg(options::OPT_nostdlib, options::OPT_nostartfiles) &&
!C.getDriver().IsCLMode() && !C.getDriver().IsFlangMode()) {
CmdArgs.push_back("-defaultlib:libcmt");
diff --git a/clang/test/Driver/fveclib.c b/clang/test/Driver/fveclib.c
index 09a12c2327137c..7d0985c4dd4f48 100644
--- a/clang/test/Driver/fveclib.c
+++ b/clang/test/Driver/fveclib.c
@@ -112,3 +112,20 @@
/* Verify no warning when math-errno is re-enabled for a different veclib
(that does not imply -fno-math-errno). */
// RUN: %clang -### --target=aarch64-linux-gnu -fveclib=ArmPL -fmath-errno
-fveclib=LIBMVEC %s 2>&1 | FileCheck --check-prefix=CHECK-REPEAT-VECLIB %s
// CHECK-REPEAT-VECLIB-NOT: math errno enabled
+
+/// Verify that vectorized routines library is being linked in.
+// RUN: %clang -### --target=aarch64-pc-windows-msvc -fveclib=ArmPL %s 2>&1 |
FileCheck --check-prefix=CHECK-LINKING-ARMPL-MSVC %s
+// RUN: %clang -### --target=aarch64-linux-gnu -fveclib=ArmPL %s 2>&1 |
FileCheck --check-prefix=CHECK-LINKING-ARMPL-LINUX %s
+// RUN: %clang -### --target=aarch64-linux-gnu -fveclib=ArmPL %s -lamath 2>&1
| FileCheck --check-prefix=CHECK-LINKING-AMATH-BEFORE-ARMPL-LINUX %s
+// RUN: %clang -### --target=arm64-apple-darwin -fveclib=ArmPL %s 2>&1 |
FileCheck --check-prefix=CHECK-LINKING-ARMPL-DARWIN %s
+// RUN: %clang -### --target=arm64-apple-darwin -fveclib=ArmPL %s -lamath 2>&1
| FileCheck --check-prefix=CHECK-LINKING-AMATH-BEFORE-ARMPL-DARWIN %s
+// CHECK-LINKING-ARMPL-LINUX: "--push-state" "--as-needed" "-lm" "-lamath"
"-lm" "--pop-state"
+// CHECK-LINKING-ARMPL-DARWIN: "-lm" "-lamath" "-lm"
+// CHECK-LINKING-ARMPL-MSV
[clang] Cache SanitizerArgs in Clang driver (NFC) (PR #119442)
https://github.com/Alexizx0078 approved this pull request. https://github.com/llvm/llvm-project/pull/119442 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [llvm] [HLSL] Add implicit resource element type concepts to AST (PR #116413)
@@ -5720,8 +5720,7 @@ static bool EvaluateUnaryTypeTrait(Sema &Self, TypeTrait UTT, case UTT_IsTypedResourceElementCompatible: assert(Self.getLangOpts().HLSL && "typed resource element compatible types are an HLSL-only feature"); -if (Self.RequireCompleteType(TInfo->getTypeLoc().getBeginLoc(), T, - diag::err_incomplete_type)) +if (T->isIncompleteType()) bob80905 wrote: Consider `RWBuffer > r8;` If we left the old condition as it was, then in evaluating the outer RWBuffer, we'd need to require that the inner element type `TemplatedBuffer` is a complete type. It is not yet, however. But we don't need to require that the element type is complete, because the type trait already rejects incomplete types. Requiring that condition caused assertion errors IIRC when recursing into deeper element types that aren't yet complete. Types would attempt to be completed earlier than necessary. This is unnecessary since the type trait itself will return false on incomplete types. "The only allowed types are builtin types or vectors (which are builtin), any incomplete type is guaranteed to not conform to the concept." https://github.com/llvm/llvm-project/pull/116413 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [NFC] Cleanup comments in HLSLExternalSemaSource.cpp (PR #119444)
llvmbot wrote:
@llvm/pr-subscribers-hlsl
@llvm/pr-subscribers-clang
Author: Joshua Batista (bob80905)
Changes
Cleaning up some comments that @bogner pointed out were unpolished.
---
Full diff: https://github.com/llvm/llvm-project/pull/119444.diff
1 Files Affected:
- (modified) clang/lib/Sema/HLSLExternalSemaSource.cpp (+1-1)
``diff
diff --git a/clang/lib/Sema/HLSLExternalSemaSource.cpp
b/clang/lib/Sema/HLSLExternalSemaSource.cpp
index 29672658525403..9eda5e59053e7a 100644
--- a/clang/lib/Sema/HLSLExternalSemaSource.cpp
+++ b/clang/lib/Sema/HLSLExternalSemaSource.cpp
@@ -851,7 +851,7 @@ static Expr *constructTypedBufferConstraintExpr(Sema &S,
SourceLocation NameLoc,
TemplateTypeParmDecl *T) {
ASTContext &Context = S.getASTContext();
- // Obtain the QualType for 'unsigned long'
+ // Obtain the QualType for 'bool'
QualType BoolTy = Context.BoolTy;
// Create a QualType that points to this TemplateTypeParmDecl
``
https://github.com/llvm/llvm-project/pull/119444
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang-tools-extra] [clang-tidy][NFC] reorder cmake source file (PR #119374)
https://github.com/carlosgalvezp approved this pull request. https://github.com/llvm/llvm-project/pull/119374 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
