From 63ef2eca887f8bce604fcf59cecdca986eb7f132 Mon Sep 17 00:00:00 2001 From: Craig Topper Date: Sun, 9 Jun 2024 18:59:03 -0700 Subject: [PATCH 01/36] [RISCV] Remove unused tablegen classes. NFC --- .../Target/RISCV/RISCVInstrInfoVPseudos.td | 110 +----------------- 1 file changed, 3 insertions(+), 107 deletions(-) diff --git a/llvm/lib/Target/RISCV/RISCVInstrInfoVPseudos.td b/llvm/lib/Target/RISCV/RISCVInstrInfoVPseudos.td index 7a95ebad364c247..08543c6bd72d890 100644 --- a/llvm/lib/Target/RISCV/RISCVInstrInfoVPseudos.td +++ b/llvm/lib/Target/RISCV/RISCVInstrInfoVPseudos.td @@ -1373,23 +1373,6 @@ class VPseudoIStoreMask LMUL, let HasSEWOp = 1; } -class VPseudoBinaryMask : - Pseudo<(outs GetVRegNoV0.R:$rd), - (ins GetVRegNoV0.R:$merge, - Op1Class:$rs2, Op2Class:$rs1, - VMaskOp:$vm, AVL:$vl, ixlenimm:$sew), []>, - RISCVVPseudo { - let mayLoad = 0; - let mayStore = 0; - let hasSideEffects = 0; - let Constraints = !interleave([Constraint, "$rd = $merge"], ","); - let HasVLOp = 1; - let HasSEWOp = 1; -} - class VPseudoBinaryMaskPolicy; -class VPatTernaryNoMask : - Pat<(result_type (!cast(intrinsic) - (result_type result_reg_class:$rs3), - (op1_type op1_reg_class:$rs1), - (op2_type op2_kind:$rs2), - VLOpFrag)), - (!cast(inst#"_"#kind#"_"#vlmul.MX) - result_reg_class:$rs3, - (op1_type op1_reg_class:$rs1), - op2_kind:$rs2, - GPR:$vl, sew)>; - class VPatTernaryNoMaskTU; -class VPatTernaryMask : - Pat<(result_type (!cast(intrinsic#"_mask") - (result_type result_reg_class:$rs3), - (op1_type op1_reg_class:$rs1), - (op2_type op2_kind:$rs2), - (mask_type V0), - VLOpFrag)), - (!cast(inst#"_"#kind#"_"#vlmul.MX # "_MASK") - result_reg_class:$rs3, - (op1_type op1_reg_class:$rs1), - op2_kind:$rs2, - (mask_type V0), - GPR:$vl, sew)>; - class VPatTernaryMaskPolicy : VPatBinaryV_V, VPatBinaryV_X; -multiclass VPatTernary { - def : VPatTernaryNoMask; - def : VPatTernaryMask; -} - -multiclass VPatTernaryNoMaskNoPolicy { - def : VPatTernaryNoMask; - def : VPatTernaryMaskPolicy; -} - multiclass VPatTernaryWithPolicy Date: Sun, 9 Jun 2024 19:07:52 -0700 Subject: [PATCH 02/36] [RISCV] Replace TU with Policy in tablegen class name. NFC --- .../Target/RISCV/RISCVInstrInfoVPseudos.td | 34 +++++++++---------- 1 file changed, 17 insertions(+), 17 deletions(-) diff --git a/llvm/lib/Target/RISCV/RISCVInstrInfoVPseudos.td b/llvm/lib/Target/RISCV/RISCVInstrInfoVPseudos.td index 08543c6bd72d890..77d3644687fddd5 100644 --- a/llvm/lib/Target/RISCV/RISCVInstrInfoVPseudos.td +++ b/llvm/lib/Target/RISCV/RISCVInstrInfoVPseudos.td @@ -4197,16 +4197,16 @@ class VPatBinaryMaskPolicy; -class VPatBinaryMaskTARoundingMode : +class VPatBinaryMaskPolicyRoundingMode : Pat<(result_type (!cast(intrinsic_name#"_mask") (result_type result_reg_class:$merge), (op1_type op1_reg_class:$rs1), @@ -4720,9 +4720,9 @@ multiclass VPatBinaryRoundingMode { def : VPatBinaryNoMaskTURoundingMode; - def : VPatBinaryMaskTARoundingMode; + def : VPatBinaryMaskPolicyRoundingMode; } multiclass VPatBinaryMSwapped; } - def : VPatBinaryMaskTARoundingMode; + def : VPatBinaryMaskPolicyRoundingMode; } } } From 367d50278aa1808cd2a2213ff1ae3c5f861667b5 Mon Sep 17 00:00:00 2001 From: Egor Ospadov Date: Sun, 9 Jun 2024 23:06:46 -0400 Subject: [PATCH 03/36] [mlir][python] Fix attribute registration in ir.py (#94615) This PR fixes attribute registration for `SI8Attr` and `UI8Attr` in `ir.py`. --- mlir/python/mlir/ir.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/mlir/python/mlir/ir.py b/mlir/python/mlir/ir.py index 80c965b2d0eb2cc..a9ac765fe1c178a 100644 --- a/mlir/python/mlir/ir.py +++ b/mlir/python/mlir/ir.py @@ -68,7 +68,7 @@ def _si1Attr(x, context): @register_attribute_builder("SI8Attr") -def _i8Attr(x, context): +def _si8Attr(x, context): return IntegerAttr.get(IntegerType.get_signed(8, context=context), x) @@ -93,7 +93,7 @@ def _ui1Attr(x, context): @register_attribute_builder("UI8Attr") -def _i8Attr(x, context): +def _ui8Attr(x, context): return IntegerAttr.get(IntegerType.get_unsigned(8, context=context), x) From 4403cdbaf01379de96f8d0d6ea4f51a085e37766 Mon Sep 17 00:00:00 2001 From: Kazu Hirata Date: Sun, 9 Jun 2024 20:08:45 -0700 Subject: [PATCH 04/36] [ProfileData] Refactor BinaryIdsStart and BinaryIdsSize (NFC) (#94922) BinaryIdsStart and BinaryIdsSize in IndexedInstrProfReader are always used together, so this patch packages them into an ArrayRef. For now, readBinaryIdsInternal immediately unpacks ArrayRef into its constituents to avoid touching the rest of readBinaryIdsInternal. --- .../llvm/ProfileData/InstrProfReader.h | 6 ++--- llvm/lib/ProfileData/InstrProfReader.cpp | 24 ++++++++++--------- 2 files changed, 15 insertions(+), 15 deletions(-) diff --git a/llvm/include/llvm/ProfileData/InstrProfReader.h b/llvm/include/llvm/ProfileData/InstrProfReader.h index ae0fd257bb05f41..3b307d083599806 100644 --- a/llvm/include/llvm/ProfileData/InstrProfReader.h +++ b/llvm/include/llvm/ProfileData/InstrProfReader.h @@ -697,10 +697,8 @@ class IndexedInstrProfReader : public InstrProfReader { /// A compiler that reads indexed profiles could construct symtab from module /// IR so it doesn't need the decompressed names. StringRef VTableName; - /// Total size of binary ids. - uint64_t BinaryIdsSize{0}; - /// Start address of binary id length and data pairs. - const uint8_t *BinaryIdsStart = nullptr; + /// A memory buffer holding binary ids. + ArrayRef BinaryIdsBuffer; // Index to the current record in the record array. unsigned RecordIndex = 0; diff --git a/llvm/lib/ProfileData/InstrProfReader.cpp b/llvm/lib/ProfileData/InstrProfReader.cpp index ca2fe4af94528ed..e18ce5d373d1cb7 100644 --- a/llvm/lib/ProfileData/InstrProfReader.cpp +++ b/llvm/lib/ProfileData/InstrProfReader.cpp @@ -91,12 +91,14 @@ static Error initializeReader(InstrProfReader &Reader) { /// associated endian format to read the binary ids correctly. static Error readBinaryIdsInternal(const MemoryBuffer &DataBuffer, - const uint64_t BinaryIdsSize, - const uint8_t *BinaryIdsStart, + ArrayRef BinaryIdsBuffer, std::vector &BinaryIds, const llvm::endianness Endian) { using namespace support; + const uint64_t BinaryIdsSize = BinaryIdsBuffer.size(); + const uint8_t *BinaryIdsStart = BinaryIdsBuffer.data(); + if (BinaryIdsSize == 0) return Error::success(); @@ -584,10 +586,10 @@ Error RawInstrProfReader::readHeader( const uint8_t *BufferEnd = (const uint8_t *)DataBuffer->getBufferEnd(); if (BinaryIdSize % sizeof(uint64_t) || BinaryIdEnd > BufferEnd) return error(instrprof_error::bad_header); - if (BinaryIdSize != 0) { - if (Error Err = - readBinaryIdsInternal(*DataBuffer, BinaryIdSize, BinaryIdStart, - BinaryIds, getDataEndianness())) + ArrayRef BinaryIdsBuffer(BinaryIdStart, BinaryIdSize); + if (!BinaryIdsBuffer.empty()) { + if (Error Err = readBinaryIdsInternal(*DataBuffer, BinaryIdsBuffer, + BinaryIds, getDataEndianness())) return Err; } @@ -1383,13 +1385,13 @@ Error IndexedInstrProfReader::readHeader() { if (Header->getIndexedProfileVersion() >= 9) { const unsigned char *Ptr = Start + Header->BinaryIdOffset; // Read binary ids size. - BinaryIdsSize = + uint64_t BinaryIdsSize = support::endian::readNext(Ptr); if (BinaryIdsSize % sizeof(uint64_t)) return error(instrprof_error::bad_header); // Set the binary ids start. - BinaryIdsStart = Ptr; - if (BinaryIdsStart > (const unsigned char *)DataBuffer->getBufferEnd()) + BinaryIdsBuffer = ArrayRef(Ptr, BinaryIdsSize); + if (Ptr > (const unsigned char *)DataBuffer->getBufferEnd()) return make_error(instrprof_error::malformed, "corrupted binary ids"); } @@ -1691,8 +1693,8 @@ Error IndexedInstrProfReader::readNextRecord(NamedInstrProfRecord &Record) { Error IndexedInstrProfReader::readBinaryIds( std::vector &BinaryIds) { - return readBinaryIdsInternal(*DataBuffer, BinaryIdsSize, BinaryIdsStart, - BinaryIds, llvm::endianness::little); + return readBinaryIdsInternal(*DataBuffer, BinaryIdsBuffer, BinaryIds, + llvm::endianness::little); } Error IndexedInstrProfReader::printBinaryIds(raw_ostream &OS) { From bf0d76d185a7c0fbedb75633afb467c74e8c0e81 Mon Sep 17 00:00:00 2001 From: Fangrui Song Date: Sun, 9 Jun 2024 21:14:10 -0700 Subject: [PATCH 05/36] [MC,test] Reorganize relax-recompute-align.s & layout-interdependency.s relax-recompute-align.s might change when we change the fragment relaxation approach. --- .../layout-interdependency.s | 0 llvm/test/MC/ELF/relax-recompute-align.s | 23 ++++++++++ llvm/test/MC/MachO/relax-recompute-align.s | 42 ------------------- 3 files changed, 23 insertions(+), 42 deletions(-) rename llvm/test/MC/{AsmParser => ELF}/layout-interdependency.s (100%) create mode 100644 llvm/test/MC/ELF/relax-recompute-align.s delete mode 100644 llvm/test/MC/MachO/relax-recompute-align.s diff --git a/llvm/test/MC/AsmParser/layout-interdependency.s b/llvm/test/MC/ELF/layout-interdependency.s similarity index 100% rename from llvm/test/MC/AsmParser/layout-interdependency.s rename to llvm/test/MC/ELF/layout-interdependency.s diff --git a/llvm/test/MC/ELF/relax-recompute-align.s b/llvm/test/MC/ELF/relax-recompute-align.s new file mode 100644 index 000000000000000..508fb5969b3b504 --- /dev/null +++ b/llvm/test/MC/ELF/relax-recompute-align.s @@ -0,0 +1,23 @@ +// RUN: llvm-mc -filetype=obj -triple i386 %s -o - | llvm-objdump -d --no-show-raw-insn - | FileCheck %s + +// This is a case where llvm-mc computes a better layout than Darwin 'as'. This +// issue is that after the first jmp slides, the .align size must be +// recomputed -- otherwise the second jump will appear to be out-of-range for a +// 1-byte jump. + +// CHECK: int3 +// CHECK-NEXT: ce: int3 +// CHECK: d0: pushal +// CHECK: 130: jl 0xd0 + +L0: + .space 0x8a, 0x90 + jmp L0 + .space (0xb3 - 0x8f), 0x90 + jle L2 + .space (0xcd - 0xb5), 0x90 + .p2align 4, 0xcc +L1: + .space (0x130 - 0xd0),0x60 + jl L1 +L2: diff --git a/llvm/test/MC/MachO/relax-recompute-align.s b/llvm/test/MC/MachO/relax-recompute-align.s deleted file mode 100644 index bfb17b71274e9c8..000000000000000 --- a/llvm/test/MC/MachO/relax-recompute-align.s +++ /dev/null @@ -1,42 +0,0 @@ -// RUN: llvm-mc -triple i386-apple-darwin9 %s -filetype=obj -o - | llvm-readobj -S - | FileCheck %s - -// FIXME: This is a horrible way of checking the output, we need an llvm-mc -// based 'otool'. - -// This is a case where llvm-mc computes a better layout than Darwin 'as'. This -// issue is that after the first jmp slides, the .align size must be -// recomputed -- otherwise the second jump will appear to be out-of-range for a -// 1-byte jump. - -L0: - .space 0x8a, 0x90 - jmp L0 - .space (0xb3 - 0x8f), 0x90 - jle L2 - .space (0xcd - 0xb5), 0x90 - .align 4, 0x90 -L1: - .space (0x130 - 0xd0),0x90 - jl L1 -L2: - -.zerofill __DATA,__bss,_sym,4,2 - -// CHECK: Section { -// CHECK-NEXT: Index: 0 -// CHECK-NEXT: Name: __text (5F 5F 74 65 78 74 00 00 00 00 00 00 00 00 00 00) -// CHECK-NEXT: Segment: __TEXT (5F 5F 54 45 58 54 00 00 00 00 00 00 00 00 00 00) -// CHECK-NEXT: Address: 0x0 -// CHECK-NEXT: Size: 0x132 -// CHECK-NEXT: Offset: 340 -// CHECK-NEXT: Alignment: 4 -// CHECK-NEXT: RelocationOffset: 0x0 -// CHECK-NEXT: RelocationCount: 0 -// CHECK-NEXT: Type: Regular (0x0) -// CHECK-NEXT: Attributes [ (0x800004) -// CHECK-NEXT: PureInstructions (0x800000) -// CHECK-NEXT: SomeInstructions (0x4) -// CHECK-NEXT: ] -// CHECK-NEXT: Reserved1: 0x0 -// CHECK-NEXT: Reserved2: 0x0 -// CHECK-NEXT: } From cb1a727dea9f063085db10190971dbd7aa7fb8e2 Mon Sep 17 00:00:00 2001 From: OverMighty Date: Mon, 10 Jun 2024 06:19:22 +0200 Subject: [PATCH 06/36] [libc][math][c23] Add nanf16 C23 math function (#94767) Part of #93566. --- libc/config/linux/aarch64/entrypoints.txt | 1 + libc/config/linux/x86_64/entrypoints.txt | 1 + libc/docs/math/index.rst | 2 +- libc/include/llvm-libc-types/CMakeLists.txt | 2 +- libc/spec/stdc.td | 1 + libc/src/math/CMakeLists.txt | 1 + libc/src/math/generic/CMakeLists.txt | 13 ++++++ libc/src/math/generic/nanf16.cpp | 23 ++++++++++ libc/src/math/nanf16.h | 20 ++++++++ libc/test/src/math/smoke/CMakeLists.txt | 16 +++++++ libc/test/src/math/smoke/nanf16_test.cpp | 51 +++++++++++++++++++++ 11 files changed, 129 insertions(+), 2 deletions(-) create mode 100644 libc/src/math/generic/nanf16.cpp create mode 100644 libc/src/math/nanf16.h create mode 100644 libc/test/src/math/smoke/nanf16_test.cpp diff --git a/libc/config/linux/aarch64/entrypoints.txt b/libc/config/linux/aarch64/entrypoints.txt index adfc1d23a1e38e1..ee86822d96d9126 100644 --- a/libc/config/linux/aarch64/entrypoints.txt +++ b/libc/config/linux/aarch64/entrypoints.txt @@ -522,6 +522,7 @@ if(LIBC_TYPES_HAS_FLOAT16) libc.src.math.llroundf16 libc.src.math.lrintf16 libc.src.math.lroundf16 + libc.src.math.nanf16 libc.src.math.nearbyintf16 libc.src.math.nextafterf16 libc.src.math.nextdownf16 diff --git a/libc/config/linux/x86_64/entrypoints.txt b/libc/config/linux/x86_64/entrypoints.txt index e86274ac34b2faa..c4b3b1aece639f0 100644 --- a/libc/config/linux/x86_64/entrypoints.txt +++ b/libc/config/linux/x86_64/entrypoints.txt @@ -555,6 +555,7 @@ if(LIBC_TYPES_HAS_FLOAT16) libc.src.math.llroundf16 libc.src.math.lrintf16 libc.src.math.lroundf16 + libc.src.math.nanf16 libc.src.math.nearbyintf16 libc.src.math.nextafterf16 libc.src.math.nextdownf16 diff --git a/libc/docs/math/index.rst b/libc/docs/math/index.rst index ee2e7643c34119e..43f63e910750f1c 100644 --- a/libc/docs/math/index.rst +++ b/libc/docs/math/index.rst @@ -186,7 +186,7 @@ Basic Operations +------------------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+----------------------------+ | modf | |check| | |check| | |check| | | |check| | 7.12.6.18 | F.10.3.18 | +------------------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+----------------------------+ -| nan | |check| | |check| | |check| | | |check| | 7.12.11.2 | F.10.8.2 | +| nan | |check| | |check| | |check| | |check| | |check| | 7.12.11.2 | F.10.8.2 | +------------------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+----------------------------+ | nearbyint | |check| | |check| | |check| | |check| | |check| | 7.12.9.3 | F.10.6.3 | +------------------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+----------------------------+ diff --git a/libc/include/llvm-libc-types/CMakeLists.txt b/libc/include/llvm-libc-types/CMakeLists.txt index ee2c910b85b000c..356ac037770d25e 100644 --- a/libc/include/llvm-libc-types/CMakeLists.txt +++ b/libc/include/llvm-libc-types/CMakeLists.txt @@ -58,7 +58,7 @@ add_header(pthread_rwlockattr_t HDR pthread_rwlockattr_t.h) add_header(pthread_t HDR pthread_t.h DEPENDS .__thread_type) add_header(rlim_t HDR rlim_t.h) add_header(time_t HDR time_t.h) -add_header(stack_t HDR stack_t.h) +add_header(stack_t HDR stack_t.h DEPENDS .size_t) add_header(suseconds_t HDR suseconds_t.h) add_header(struct_flock HDR struct_flock.h DEPENDS .off_t .pid_t) add_header(struct_flock64 HDR struct_flock64.h DEPENDS .off64_t .pid_t) diff --git a/libc/spec/stdc.td b/libc/spec/stdc.td index 38521f5d364b976..ac16041e44c99dc 100644 --- a/libc/spec/stdc.td +++ b/libc/spec/stdc.td @@ -695,6 +695,7 @@ def StdC : StandardSpec<"stdc"> { FunctionSpec<"nanf", RetValSpec, [ArgSpec]>, FunctionSpec<"nan", RetValSpec, [ArgSpec]>, FunctionSpec<"nanl", RetValSpec, [ArgSpec]>, + GuardedFunctionSpec<"nanf16", RetValSpec, [ArgSpec], "LIBC_TYPES_HAS_FLOAT16">, GuardedFunctionSpec<"nanf128", RetValSpec, [ArgSpec], "LIBC_TYPES_HAS_FLOAT128">, FunctionSpec<"canonicalize", RetValSpec, [ArgSpec, ArgSpec]>, diff --git a/libc/src/math/CMakeLists.txt b/libc/src/math/CMakeLists.txt index e6cb8fd61597146..68b8924460bf492 100644 --- a/libc/src/math/CMakeLists.txt +++ b/libc/src/math/CMakeLists.txt @@ -272,6 +272,7 @@ add_math_entrypoint_object(modff128) add_math_entrypoint_object(nan) add_math_entrypoint_object(nanf) add_math_entrypoint_object(nanl) +add_math_entrypoint_object(nanf16) add_math_entrypoint_object(nanf128) add_math_entrypoint_object(nearbyint) diff --git a/libc/src/math/generic/CMakeLists.txt b/libc/src/math/generic/CMakeLists.txt index 97740834fb0293f..4f94468939ed862 100644 --- a/libc/src/math/generic/CMakeLists.txt +++ b/libc/src/math/generic/CMakeLists.txt @@ -2631,6 +2631,19 @@ add_entrypoint_object( -O3 ) +add_entrypoint_object( + nanf16 + SRCS + nanf16.cpp + HDRS + ../nanf16.h + DEPENDS + libc.src.__support.str_to_float + libc.src.errno.errno + COMPILE_OPTIONS + -O3 +) + add_entrypoint_object( nanf128 SRCS diff --git a/libc/src/math/generic/nanf16.cpp b/libc/src/math/generic/nanf16.cpp new file mode 100644 index 000000000000000..c42cd25fe3ef0c7 --- /dev/null +++ b/libc/src/math/generic/nanf16.cpp @@ -0,0 +1,23 @@ +//===-- Implementation of nanf16 function ---------------------------------===// +// +// 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 +// +//===----------------------------------------------------------------------===// + +#include "src/math/nanf16.h" +#include "src/__support/common.h" +#include "src/__support/str_to_float.h" +#include "src/errno/libc_errno.h" + +namespace LIBC_NAMESPACE { + +LLVM_LIBC_FUNCTION(float16, nanf16, (const char *arg)) { + auto result = internal::strtonan(arg); + if (result.has_error()) + libc_errno = result.error; + return result.value; +} + +} // namespace LIBC_NAMESPACE diff --git a/libc/src/math/nanf16.h b/libc/src/math/nanf16.h new file mode 100644 index 000000000000000..c2db4bac0a15515 --- /dev/null +++ b/libc/src/math/nanf16.h @@ -0,0 +1,20 @@ +//===-- Implementation header for nanf16 ------------------------*- C++ -*-===// +// +// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. +// See https://llvm.org/LICENSE.txt for license information. +// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception +// +//===----------------------------------------------------------------------===// + +#ifndef LLVM_LIBC_SRC_MATH_NANF16_H +#define LLVM_LIBC_SRC_MATH_NANF16_H + +#include "src/__support/macros/properties/types.h" + +namespace LIBC_NAMESPACE { + +float16 nanf16(const char *arg); + +} // namespace LIBC_NAMESPACE + +#endif // LLVM_LIBC_SRC_MATH_NANF16_H diff --git a/libc/test/src/math/smoke/CMakeLists.txt b/libc/test/src/math/smoke/CMakeLists.txt index dfda1bc2457e088..cc0e00aeb1dd256 100644 --- a/libc/test/src/math/smoke/CMakeLists.txt +++ b/libc/test/src/math/smoke/CMakeLists.txt @@ -2636,6 +2636,22 @@ add_fp_unittest( UNIT_TEST_ONLY ) +add_fp_unittest( + nanf16_test + SUITE + libc-math-smoke-tests + SRCS + nanf16_test.cpp + DEPENDS + libc.include.signal + libc.src.math.nanf16 + libc.src.__support.FPUtil.fp_bits + libc.src.__support.macros.sanitizer + # FIXME: The nan tests currently have death tests, which aren't supported for + # hermetic tests. + UNIT_TEST_ONLY +) + add_fp_unittest( nanf128_test SUITE diff --git a/libc/test/src/math/smoke/nanf16_test.cpp b/libc/test/src/math/smoke/nanf16_test.cpp new file mode 100644 index 000000000000000..ec17a73d881aa62 --- /dev/null +++ b/libc/test/src/math/smoke/nanf16_test.cpp @@ -0,0 +1,51 @@ +//===-- Unittests for nanf16 ----------------------------------------------===// +// +// 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 +// +//===----------------------------------------------------------------------===// + +#include "src/__support/FPUtil/FPBits.h" +#include "src/__support/macros/sanitizer.h" +#include "src/math/nanf16.h" +#include "test/UnitTest/FEnvSafeTest.h" +#include "test/UnitTest/FPMatcher.h" +#include "test/UnitTest/Test.h" + +#include + +class LlvmLibcNanf16Test : public LIBC_NAMESPACE::testing::FEnvSafeTest { +public: + using StorageType = LIBC_NAMESPACE::fputil::FPBits::StorageType; + + void run_test(const char *input_str, StorageType bits) { + float16 result = LIBC_NAMESPACE::nanf16(input_str); + auto actual_fp = LIBC_NAMESPACE::fputil::FPBits(result); + auto expected_fp = LIBC_NAMESPACE::fputil::FPBits(bits); + EXPECT_EQ(actual_fp.uintval(), expected_fp.uintval()); + }; +}; + +TEST_F(LlvmLibcNanf16Test, NCharSeq) { + run_test("", 0x7e00); + run_test("123", 0x7e7b); + run_test("0x123", 0x7f23); + run_test("1a", 0x7e00); + run_test("1234567890qwertyuiopasdfghjklzxcvbnmQWERTYUIOPASDFGHJKLZXCVBNM_", + 0x7e00); + run_test("10000000000000000000000000000000000000000000000000", 0x7e00); +} + +TEST_F(LlvmLibcNanf16Test, RandomString) { + run_test(" 1234", 0x7e00); + run_test("-1234", 0x7e00); + run_test("asd&f", 0x7e00); + run_test("123 ", 0x7e00); +} + +#ifndef LIBC_HAVE_ADDRESS_SANITIZER +TEST_F(LlvmLibcNanf16Test, InvalidInput) { + EXPECT_DEATH([] { LIBC_NAMESPACE::nanf16(nullptr); }, WITH_SIGNAL(SIGSEGV)); +} +#endif // LIBC_HAVE_ADDRESS_SANITIZER From 9d0754ada5dbbc0c009bcc2f7824488419cc5530 Mon Sep 17 00:00:00 2001 From: Fangrui Song Date: Sun, 9 Jun 2024 23:05:05 -0700 Subject: [PATCH 07/36] [MC] Relax fragments eagerly Lazy relaxation caused hash table lookups (`getFragmentOffset`) and complex use/compute interdependencies. Some expressions involding forward declared symbols (e.g. `subsection-if.s`) cannot be computed. Recursion detection requires complex `IsBeingLaidOut` (https://reviews.llvm.org/D79570). D76114's `invalidateFragmentsFrom` makes lazy relaxation even less useful. Switch to eager relaxation to greatly simplify code and resolve these issues. This change also removes a `getPrevNode` use, which makes it more feasible to replace the fragment representation, which might yield a large peak RSS win. Minor downsides: The number of section relaxations may increase (offset by avoiding the hash table lookup). For relax-recompute-align.s, the computed layout is not optimal. --- llvm/include/llvm/MC/MCAsmLayout.h | 20 +--- llvm/include/llvm/MC/MCAssembler.h | 4 - llvm/include/llvm/MC/MCFragment.h | 3 - llvm/include/llvm/MC/MCSection.h | 5 + llvm/lib/MC/MCAssembler.cpp | 117 ++++++++-------------- llvm/lib/MC/MCExpr.cpp | 4 - llvm/lib/MC/MCFragment.cpp | 60 +---------- llvm/lib/MC/MCSection.cpp | 4 +- llvm/test/MC/ELF/layout-interdependency.s | 8 +- llvm/test/MC/ELF/relax-recompute-align.s | 8 +- llvm/test/MC/ELF/subsection-if.s | 4 +- 11 files changed, 63 insertions(+), 174 deletions(-) diff --git a/llvm/include/llvm/MC/MCAsmLayout.h b/llvm/include/llvm/MC/MCAsmLayout.h index 94cfb76bdce4db7..6fbfce78759e1b8 100644 --- a/llvm/include/llvm/MC/MCAsmLayout.h +++ b/llvm/include/llvm/MC/MCAsmLayout.h @@ -31,37 +31,21 @@ class MCAsmLayout { /// List of sections in layout order. llvm::SmallVector SectionOrder; - /// The last fragment which was laid out, or 0 if nothing has been laid - /// out. Fragments are always laid out in order, so all fragments with a - /// lower ordinal will be valid. - mutable DenseMap LastValidFragment; - - /// Make sure that the layout for the given fragment is valid, lazily - /// computing it if necessary. + /// Compute the layout for the section if necessary. void ensureValid(const MCFragment *F) const; - /// Is the layout for this fragment valid? - bool isFragmentValid(const MCFragment *F) const; - public: MCAsmLayout(MCAssembler &Assembler); /// Get the assembler object this is a layout for. MCAssembler &getAssembler() const { return Assembler; } - /// \returns whether the offset of fragment \p F can be obtained via - /// getFragmentOffset. - bool canGetFragmentOffset(const MCFragment *F) const; - /// Invalidate the fragments starting with F because it has been /// resized. The fragment's size should have already been updated, but /// its bundle padding will be recomputed. void invalidateFragmentsFrom(MCFragment *F); - /// Perform layout for a single fragment, assuming that the previous - /// fragment has already been laid out correctly, and the parent section has - /// been initialized. - void layoutFragment(MCFragment *Fragment); + void layoutBundle(MCFragment *F); /// \name Section Access (in layout order) /// @{ diff --git a/llvm/include/llvm/MC/MCAssembler.h b/llvm/include/llvm/MC/MCAssembler.h index 3fa5f2fe655e86e..914c7506e754b39 100644 --- a/llvm/include/llvm/MC/MCAssembler.h +++ b/llvm/include/llvm/MC/MCAssembler.h @@ -203,10 +203,6 @@ class MCAssembler { /// were adjusted. bool layoutOnce(MCAsmLayout &Layout); - /// Perform one layout iteration of the given section and return true - /// if any offsets were adjusted. - bool layoutSectionOnce(MCAsmLayout &Layout, MCSection &Sec); - /// Perform relaxation on a single fragment - returns true if the fragment /// changes as a result of relaxation. bool relaxFragment(MCAsmLayout &Layout, MCFragment &F); diff --git a/llvm/include/llvm/MC/MCFragment.h b/llvm/include/llvm/MC/MCFragment.h index a9b19dc56f16a5e..67e10c35894954a 100644 --- a/llvm/include/llvm/MC/MCFragment.h +++ b/llvm/include/llvm/MC/MCFragment.h @@ -70,9 +70,6 @@ class MCFragment : public ilist_node_with_parent { FragmentType Kind; - /// Whether fragment is being laid out. - bool IsBeingLaidOut; - protected: bool HasInstructions; bool LinkerRelaxable = false; diff --git a/llvm/include/llvm/MC/MCSection.h b/llvm/include/llvm/MC/MCSection.h index ad41d4475047a82..90effde5bb670a5 100644 --- a/llvm/include/llvm/MC/MCSection.h +++ b/llvm/include/llvm/MC/MCSection.h @@ -89,6 +89,8 @@ class MCSection { /// Whether this section has had instructions emitted into it. bool HasInstructions : 1; + bool HasLayout : 1; + bool IsRegistered : 1; MCDummyFragment DummyFragment; @@ -166,6 +168,9 @@ class MCSection { bool hasInstructions() const { return HasInstructions; } void setHasInstructions(bool Value) { HasInstructions = Value; } + bool hasLayout() const { return HasLayout; } + void setHasLayout(bool Value) { HasLayout = Value; } + bool isRegistered() const { return IsRegistered; } void setIsRegistered(bool Value) { IsRegistered = Value; } diff --git a/llvm/lib/MC/MCAssembler.cpp b/llvm/lib/MC/MCAssembler.cpp index ad30b5ce9e63144..c56d28bb70e7305 100644 --- a/llvm/lib/MC/MCAssembler.cpp +++ b/llvm/lib/MC/MCAssembler.cpp @@ -66,7 +66,6 @@ STATISTIC(EmittedFillFragments, STATISTIC(EmittedNopsFragments, "Number of emitted assembler fragments - nops"); STATISTIC(EmittedOrgFragments, "Number of emitted assembler fragments - org"); STATISTIC(evaluateFixup, "Number of evaluated fixups"); -STATISTIC(FragmentLayouts, "Number of fragment layouts"); STATISTIC(ObjectBytes, "Number of emitted object file bytes"); STATISTIC(RelaxationSteps, "Number of assembler layout and relaxation steps"); STATISTIC(RelaxedInstructions, "Number of relaxed instructions"); @@ -404,29 +403,7 @@ uint64_t MCAssembler::computeFragmentSize(const MCAsmLayout &Layout, llvm_unreachable("invalid fragment kind"); } -void MCAsmLayout::layoutFragment(MCFragment *F) { - MCFragment *Prev = F->getPrevNode(); - - // We should never try to recompute something which is valid. - assert(!isFragmentValid(F) && "Attempt to recompute a valid fragment!"); - // We should never try to compute the fragment layout if its predecessor - // isn't valid. - assert((!Prev || isFragmentValid(Prev)) && - "Attempt to compute fragment before its predecessor!"); - - assert(!F->IsBeingLaidOut && "Already being laid out!"); - F->IsBeingLaidOut = true; - - ++stats::FragmentLayouts; - - // Compute fragment offset and size. - if (Prev) - F->Offset = Prev->Offset + getAssembler().computeFragmentSize(*this, *Prev); - else - F->Offset = 0; - F->IsBeingLaidOut = false; - LastValidFragment[F->getParent()] = F; - +void MCAsmLayout::layoutBundle(MCFragment *F) { // If bundling is enabled and this fragment has instructions in it, it has to // obey the bundling restrictions. With padding, we'll have: // @@ -454,21 +431,40 @@ void MCAsmLayout::layoutFragment(MCFragment *F) { // within-fragment padding (which would produce less padding when N is less // than the bundle size), but for now we don't. // - if (Assembler.isBundlingEnabled() && F->hasInstructions()) { - assert(isa(F) && - "Only MCEncodedFragment implementations have instructions"); - MCEncodedFragment *EF = cast(F); - uint64_t FSize = Assembler.computeFragmentSize(*this, *EF); - - if (!Assembler.getRelaxAll() && FSize > Assembler.getBundleAlignSize()) - report_fatal_error("Fragment can't be larger than a bundle size"); - - uint64_t RequiredBundlePadding = - computeBundlePadding(Assembler, EF, EF->Offset, FSize); - if (RequiredBundlePadding > UINT8_MAX) - report_fatal_error("Padding cannot exceed 255 bytes"); - EF->setBundlePadding(static_cast(RequiredBundlePadding)); - EF->Offset += RequiredBundlePadding; + assert(isa(F) && + "Only MCEncodedFragment implementations have instructions"); + MCEncodedFragment *EF = cast(F); + uint64_t FSize = Assembler.computeFragmentSize(*this, *EF); + + if (!Assembler.getRelaxAll() && FSize > Assembler.getBundleAlignSize()) + report_fatal_error("Fragment can't be larger than a bundle size"); + + uint64_t RequiredBundlePadding = + computeBundlePadding(Assembler, EF, EF->Offset, FSize); + if (RequiredBundlePadding > UINT8_MAX) + report_fatal_error("Padding cannot exceed 255 bytes"); + EF->setBundlePadding(static_cast(RequiredBundlePadding)); + EF->Offset += RequiredBundlePadding; +} + +uint64_t MCAsmLayout::getFragmentOffset(const MCFragment *F) const { + ensureValid(F); + return F->Offset; +} + +void MCAsmLayout::ensureValid(const MCFragment *Frag) const { + MCSection &Sec = *Frag->getParent(); + if (Sec.hasLayout()) + return; + Sec.setHasLayout(true); + uint64_t Offset = 0; + for (MCFragment &F : Sec) { + F.Offset = Offset; + if (Assembler.isBundlingEnabled() && F.hasInstructions()) { + const_cast(this)->layoutBundle(&F); + Offset = F.Offset; + } + Offset += getAssembler().computeFragmentSize(*this, F); } } @@ -848,7 +844,7 @@ void MCAssembler::layout(MCAsmLayout &Layout) { // another. If any fragment has changed size, we have to re-layout (and // as a result possibly further relax) all. for (MCSection &Sec : *this) - Layout.invalidateFragmentsFrom(&*Sec.begin()); + Sec.setHasLayout(false); } DEBUG_WITH_TYPE("mc-dump", { @@ -1109,7 +1105,6 @@ bool MCAssembler::relaxBoundaryAlign(MCAsmLayout &Layout, if (NewSize == BF.getSize()) return false; BF.setSize(NewSize); - Layout.invalidateFragmentsFrom(&BF); return true; } @@ -1219,47 +1214,19 @@ bool MCAssembler::relaxFragment(MCAsmLayout &Layout, MCFragment &F) { } } -bool MCAssembler::layoutSectionOnce(MCAsmLayout &Layout, MCSection &Sec) { - // Holds the first fragment which needed relaxing during this layout. It will - // remain NULL if none were relaxed. - // When a fragment is relaxed, all the fragments following it should get - // invalidated because their offset is going to change. - MCFragment *FirstRelaxedFragment = nullptr; - - // Attempt to relax all the fragments in the section. - for (MCFragment &Frag : Sec) { - // Check if this is a fragment that needs relaxation. - bool RelaxedFrag = relaxFragment(Layout, Frag); - if (RelaxedFrag && !FirstRelaxedFragment) - FirstRelaxedFragment = &Frag; - } - if (FirstRelaxedFragment) { - Layout.invalidateFragmentsFrom(FirstRelaxedFragment); - return true; - } - return false; -} - bool MCAssembler::layoutOnce(MCAsmLayout &Layout) { ++stats::RelaxationSteps; - bool WasRelaxed = false; - for (MCSection &Sec : *this) { - while (layoutSectionOnce(Layout, Sec)) - WasRelaxed = true; - } - - return WasRelaxed; + bool Changed = false; + for (MCSection &Sec : *this) + for (MCFragment &Frag : Sec) + if (relaxFragment(Layout, Frag)) + Changed = true; + return Changed; } void MCAssembler::finishLayout(MCAsmLayout &Layout) { assert(getBackendPtr() && "Expected assembler backend"); - // The layout is done. Mark every fragment as valid. - for (unsigned int i = 0, n = Layout.getSectionOrder().size(); i != n; ++i) { - MCSection &Section = *Layout.getSectionOrder()[i]; - Layout.getFragmentOffset(&*Section.getFragmentList().rbegin()); - computeFragmentSize(Layout, *Section.getFragmentList().rbegin()); - } getBackend().finishLayout(*this, Layout); } diff --git a/llvm/lib/MC/MCExpr.cpp b/llvm/lib/MC/MCExpr.cpp index b065d03651c4527..b70ac86c18ccf16 100644 --- a/llvm/lib/MC/MCExpr.cpp +++ b/llvm/lib/MC/MCExpr.cpp @@ -645,10 +645,6 @@ static void AttemptToFoldSymbolOffsetDifference( Addend += SA.getOffset() - SB.getOffset(); return FinalizeFolding(); } - // One of the symbol involved is part of a fragment being laid out. Quit now - // to avoid a self loop. - if (!Layout->canGetFragmentOffset(FA) || !Layout->canGetFragmentOffset(FB)) - return; // Eagerly evaluate when layout is finalized. Addend += Layout->getSymbolOffset(A->getSymbol()) - diff --git a/llvm/lib/MC/MCFragment.cpp b/llvm/lib/MC/MCFragment.cpp index 7d0826802d0afd4..84a587164c788d2 100644 --- a/llvm/lib/MC/MCFragment.cpp +++ b/llvm/lib/MC/MCFragment.cpp @@ -39,64 +39,8 @@ MCAsmLayout::MCAsmLayout(MCAssembler &Asm) : Assembler(Asm) { SectionOrder.push_back(&Sec); } -bool MCAsmLayout::isFragmentValid(const MCFragment *F) const { - const MCSection *Sec = F->getParent(); - const MCFragment *LastValid = LastValidFragment.lookup(Sec); - if (!LastValid) - return false; - assert(LastValid->getParent() == Sec); - return F->getLayoutOrder() <= LastValid->getLayoutOrder(); -} - -bool MCAsmLayout::canGetFragmentOffset(const MCFragment *F) const { - MCSection *Sec = F->getParent(); - MCSection::iterator I; - if (MCFragment *LastValid = LastValidFragment[Sec]) { - // Fragment already valid, offset is available. - if (F->getLayoutOrder() <= LastValid->getLayoutOrder()) - return true; - I = ++MCSection::iterator(LastValid); - } else - I = Sec->begin(); - - // A fragment ordered before F is currently being laid out. - const MCFragment *FirstInvalidFragment = &*I; - if (FirstInvalidFragment->IsBeingLaidOut) - return false; - - return true; -} - void MCAsmLayout::invalidateFragmentsFrom(MCFragment *F) { - // If this fragment wasn't already valid, we don't need to do anything. - if (!isFragmentValid(F)) - return; - - // Otherwise, reset the last valid fragment to the previous fragment - // (if this is the first fragment, it will be NULL). - LastValidFragment[F->getParent()] = F->getPrevNode(); -} - -void MCAsmLayout::ensureValid(const MCFragment *F) const { - MCSection *Sec = F->getParent(); - MCSection::iterator I; - if (MCFragment *Cur = LastValidFragment[Sec]) - I = ++MCSection::iterator(Cur); - else - I = Sec->begin(); - - // Advance the layout position until the fragment is valid. - while (!isFragmentValid(F)) { - assert(I != Sec->end() && "Layout bookkeeping error"); - const_cast(this)->layoutFragment(&*I); - ++I; - } -} - -uint64_t MCAsmLayout::getFragmentOffset(const MCFragment *F) const { - ensureValid(F); - assert(F->Offset != ~UINT64_C(0) && "Address not set!"); - return F->Offset; + F->getParent()->setHasLayout(false); } // Simple getSymbolOffset helper for the non-variable case. @@ -258,7 +202,7 @@ void ilist_alloc_traits::deleteNode(MCFragment *V) { V->destroy(); } MCFragment::MCFragment(FragmentType Kind, bool HasInstructions, MCSection *Parent) : Parent(Parent), Atom(nullptr), Offset(~UINT64_C(0)), LayoutOrder(0), - Kind(Kind), IsBeingLaidOut(false), HasInstructions(HasInstructions) { + Kind(Kind), HasInstructions(HasInstructions) { if (Parent && !isa(*this)) Parent->addFragment(*this); } diff --git a/llvm/lib/MC/MCSection.cpp b/llvm/lib/MC/MCSection.cpp index ea3baf96b38ed18..12e69f70537b78c 100644 --- a/llvm/lib/MC/MCSection.cpp +++ b/llvm/lib/MC/MCSection.cpp @@ -23,8 +23,8 @@ using namespace llvm; MCSection::MCSection(SectionVariant V, StringRef Name, SectionKind K, MCSymbol *Begin) : Begin(Begin), BundleGroupBeforeFirstInst(false), HasInstructions(false), - IsRegistered(false), DummyFragment(this), Name(Name), Variant(V), - Kind(K) {} + HasLayout(false), IsRegistered(false), DummyFragment(this), Name(Name), + Variant(V), Kind(K) {} MCSymbol *MCSection::getEndSymbol(MCContext &Ctx) { if (!End) diff --git a/llvm/test/MC/ELF/layout-interdependency.s b/llvm/test/MC/ELF/layout-interdependency.s index d275614e87e74d3..5ebcdfacdbad764 100644 --- a/llvm/test/MC/ELF/layout-interdependency.s +++ b/llvm/test/MC/ELF/layout-interdependency.s @@ -1,12 +1,10 @@ -# RUN: not llvm-mc --filetype=obj %s -o /dev/null 2>&1 | FileCheck %s -# REQUIRES: object-emission -# UNSUPPORTED: target={{.*}}-zos{{.*}} +# RUN: not llvm-mc -filetype=obj -triple=x86_64 %s -o /dev/null 2>&1 | FileCheck %s fct_end: -# CHECK: layout-interdependency.s:[[#@LINE+1]]:7: error: expected assembly-time absolute expression +# CHECK: layout-interdependency.s:[[#@LINE+1]]:7: error: invalid number of bytes .fill (data_start - fct_end), 1, 42 -# CHECK: layout-interdependency.s:[[#@LINE+1]]:7: error: expected assembly-time absolute expression +# CHECK: layout-interdependency.s:[[#@LINE+1]]:7: error: invalid number of bytes .fill (fct_end - data_start), 1, 42 data_start: diff --git a/llvm/test/MC/ELF/relax-recompute-align.s b/llvm/test/MC/ELF/relax-recompute-align.s index 508fb5969b3b504..44e1f1f9719970f 100644 --- a/llvm/test/MC/ELF/relax-recompute-align.s +++ b/llvm/test/MC/ELF/relax-recompute-align.s @@ -1,14 +1,14 @@ // RUN: llvm-mc -filetype=obj -triple i386 %s -o - | llvm-objdump -d --no-show-raw-insn - | FileCheck %s -// This is a case where llvm-mc computes a better layout than Darwin 'as'. This +/// This is a case where the computed layout is not optimal. The // issue is that after the first jmp slides, the .align size must be // recomputed -- otherwise the second jump will appear to be out-of-range for a // 1-byte jump. // CHECK: int3 -// CHECK-NEXT: ce: int3 -// CHECK: d0: pushal -// CHECK: 130: jl 0xd0 +// CHECK-NEXT: d2: int3 +// CHECK: e0: pushal +// CHECK: 140: jl 0xe0 L0: .space 0x8a, 0x90 diff --git a/llvm/test/MC/ELF/subsection-if.s b/llvm/test/MC/ELF/subsection-if.s index 7f2cba6f52100ec..905cb5afff2ea1d 100644 --- a/llvm/test/MC/ELF/subsection-if.s +++ b/llvm/test/MC/ELF/subsection-if.s @@ -1,8 +1,10 @@ # RUN: llvm-mc -filetype=obj -triple=x86_64 %s -o %t # RUN: llvm-readelf -x .text %t | FileCheck %s -# RUN: not llvm-mc -filetype=obj -triple=x86_64 --defsym ERR=1 %s -o /dev/null 2>&1 | FileCheck %s --check-prefix=ERR +# RUN: llvm-mc -filetype=obj -triple=x86_64 --defsym ERR=1 %s -o %t1 +# RUN: llvm-readelf -x .text %t1 | FileCheck %s --check-prefix=CHECK1 # CHECK: 0x00000000 9090 +# CHECK1: 0x00000000 90909090 90 .subsection 1 661: From 13896b6ce9b791db6d1ac40c6a05eb893405fcd7 Mon Sep 17 00:00:00 2001 From: Matthias Springer Date: Mon, 10 Jun 2024 08:07:24 +0200 Subject: [PATCH 08/36] [mlir][bufferization] Fix handling of indirect function calls (#94896) This commit fixes a crash in the ownership-based buffer deallocation pass when indirectly calling a function via SSA value. Such functions must be conservatively assumed to be public. Fixes #94780. --- .../OwnershipBasedBufferDeallocation.cpp | 7 ++++--- .../dealloc-callop-interface.mlir | 19 +++++++++++++++++++ 2 files changed, 23 insertions(+), 3 deletions(-) diff --git a/mlir/lib/Dialect/Bufferization/Transforms/OwnershipBasedBufferDeallocation.cpp b/mlir/lib/Dialect/Bufferization/Transforms/OwnershipBasedBufferDeallocation.cpp index a8ec111f8c304ba..bd5c4d476921696 100644 --- a/mlir/lib/Dialect/Bufferization/Transforms/OwnershipBasedBufferDeallocation.cpp +++ b/mlir/lib/Dialect/Bufferization/Transforms/OwnershipBasedBufferDeallocation.cpp @@ -822,10 +822,11 @@ FailureOr BufferDeallocation::handleInterface(CallOpInterface op) { // Lookup the function operation and check if it has private visibility. If // the function is referenced by SSA value instead of a Symbol, it's assumed - // to be always private. + // to be public. (And we cannot easily change the type of the SSA value + // anyway.) Operation *funcOp = op.resolveCallable(state.getSymbolTable()); - bool isPrivate = true; - if (auto symbol = dyn_cast(funcOp)) + bool isPrivate = false; + if (auto symbol = dyn_cast_or_null(funcOp)) isPrivate = symbol.isPrivate() && !symbol.isDeclaration(); // If the private-function-dynamic-ownership option is enabled and we are diff --git a/mlir/test/Dialect/Bufferization/Transforms/OwnershipBasedBufferDeallocation/dealloc-callop-interface.mlir b/mlir/test/Dialect/Bufferization/Transforms/OwnershipBasedBufferDeallocation/dealloc-callop-interface.mlir index 63947150c23b30e..a77442dfe40e5b9 100644 --- a/mlir/test/Dialect/Bufferization/Transforms/OwnershipBasedBufferDeallocation/dealloc-callop-interface.mlir +++ b/mlir/test/Dialect/Bufferization/Transforms/OwnershipBasedBufferDeallocation/dealloc-callop-interface.mlir @@ -131,3 +131,22 @@ func.func @g(%arg0: memref) -> memref { // CHECK-DYNAMIC-LABEL: func private @f // CHECK-DYNAMIC-SAME: (memref) -> memref // CHECK-DYNAMIC: call @f({{.*}}) : (memref) -> memref + +// ----- + +func.func @func_call_indirect(%m: memref, %f: (memref) -> (memref)) { + %0 = func.call_indirect %f(%m) : (memref) -> (memref) + return +} + +// CHECK-LABEL: func @func_call_indirect( +// CHECK: %[[true:.*]] = arith.constant true +// CHECK: %[[call:.*]] = call_indirect {{.*}} : (memref) -> memref +// CHECK: %[[base_call:.*]], %{{.*}}, %{{.*}}, %{{.*}} = memref.extract_strided_metadata %[[call]] +// CHECK: bufferization.dealloc (%[[base_call]] : {{.*}}) if (%[[true]]) + +// CHECK-DYNAMIC-LABEL: func @func_call_indirect( +// CHECK-DYNAMIC: %[[true:.*]] = arith.constant true +// CHECK-DYNAMIC: %[[call:.*]] = call_indirect {{.*}} : (memref) -> memref +// CHECK-DYNAMIC: %[[base_call:.*]], %{{.*}}, %{{.*}}, %{{.*}} = memref.extract_strided_metadata %[[call]] +// CHECK-DYNAMIC: bufferization.dealloc (%[[base_call]] : {{.*}}) if (%[[true]]) From bb4ee27a31695bb06bfe56d442b30b25ac88d7cd Mon Sep 17 00:00:00 2001 From: Fangrui Song Date: Sun, 9 Jun 2024 23:14:15 -0700 Subject: [PATCH 09/36] [MC] Remove the last MCFragment::getPrevNode use. NFC --- llvm/lib/MC/MCAssembler.cpp | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/llvm/lib/MC/MCAssembler.cpp b/llvm/lib/MC/MCAssembler.cpp index c56d28bb70e7305..a7157e7a37b4399 100644 --- a/llvm/lib/MC/MCAssembler.cpp +++ b/llvm/lib/MC/MCAssembler.cpp @@ -1094,9 +1094,11 @@ bool MCAssembler::relaxBoundaryAlign(MCAsmLayout &Layout, uint64_t AlignedOffset = Layout.getFragmentOffset(&BF); uint64_t AlignedSize = 0; - for (const MCFragment *F = BF.getLastFragment(); F != &BF; - F = F->getPrevNode()) + for (const MCFragment *F = BF.getNextNode();; F = F->getNextNode()) { AlignedSize += computeFragmentSize(Layout, *F); + if (F == BF.getLastFragment()) + break; + } Align BoundaryAlignment = BF.getAlignment(); uint64_t NewSize = needPadding(AlignedOffset, AlignedSize, BoundaryAlignment) From 87cedbec1cce1cbfc6ad15ae9b88e3481ba9eead Mon Sep 17 00:00:00 2001 From: Mark de Wever Date: Mon, 10 Jun 2024 08:40:04 +0200 Subject: [PATCH 10/36] [libc++][TZDB] Implements time_zone::to_sys. (#90901) This implements the overload with the choose argument and adds this enum. Implements parts of: - P0355 Extending chrono to Calendars and Time Zones --- libcxx/include/__chrono/time_zone.h | 32 ++++ libcxx/include/chrono | 4 + libcxx/modules/std/chrono.inc | 3 - .../diagnostics/chrono.nodiscard.verify.cpp | 3 + .../time.zone.timezone/choose.pass.cpp | 37 +++++ .../assert.to_sys_choose.pass.cpp | 41 +++++ .../time.zone.members/to_sys_choose.pass.cpp | 147 ++++++++++++++++++ 7 files changed, 264 insertions(+), 3 deletions(-) create mode 100644 libcxx/test/libcxx/time/time.zone/time.zone.timezone/choose.pass.cpp create mode 100644 libcxx/test/libcxx/time/time.zone/time.zone.timezone/time.zone.members/assert.to_sys_choose.pass.cpp create mode 100644 libcxx/test/std/time/time.zone/time.zone.timezone/time.zone.members/to_sys_choose.pass.cpp diff --git a/libcxx/include/__chrono/time_zone.h b/libcxx/include/__chrono/time_zone.h index b4069823dbe74fd..89ab6d1f2110271 100644 --- a/libcxx/include/__chrono/time_zone.h +++ b/libcxx/include/__chrono/time_zone.h @@ -42,6 +42,8 @@ _LIBCPP_BEGIN_NAMESPACE_STD namespace chrono { +enum class choose { earliest, latest }; + class _LIBCPP_AVAILABILITY_TZDB time_zone { _LIBCPP_HIDE_FROM_ABI time_zone() = default; @@ -97,6 +99,36 @@ class _LIBCPP_AVAILABILITY_TZDB time_zone { return {}; } + template + [[nodiscard]] _LIBCPP_HIDE_FROM_ABI sys_time> + to_sys(const local_time<_Duration>& __time, choose __z) const { + local_info __info = get_info(__time); + switch (__info.result) { + case local_info::unique: + case local_info::nonexistent: // first and second are the same + return sys_time>{__time.time_since_epoch() - __info.first.offset}; + + case local_info::ambiguous: + switch (__z) { + case choose::earliest: + return sys_time>{__time.time_since_epoch() - __info.first.offset}; + + case choose::latest: + return sys_time>{__time.time_since_epoch() - __info.second.offset}; + + // Note a value out of bounds is not specified. + } + } + + // TODO TZDB The standard does not specify anything in these cases. + _LIBCPP_ASSERT_ARGUMENT_WITHIN_DOMAIN( + __info.result != -1, "cannot convert the local time; it would be before the minimum system clock value"); + _LIBCPP_ASSERT_ARGUMENT_WITHIN_DOMAIN( + __info.result != -2, "cannot convert the local time; it would be after the maximum system clock value"); + + return {}; + } + [[nodiscard]] _LIBCPP_HIDE_FROM_ABI const __impl& __implementation() const noexcept { return *__impl_; } private: diff --git a/libcxx/include/chrono b/libcxx/include/chrono index 46bca597ee1fd94..78e885f6ca373d9 100644 --- a/libcxx/include/chrono +++ b/libcxx/include/chrono @@ -774,6 +774,10 @@ class time_zone { template sys_time> to_sys(const local_time& tp) const; + + template + sys_time> + to_sys(const local_time& tp, choose z) const; }; bool operator==(const time_zone& x, const time_zone& y) noexcept; // C++20 strong_ordering operator<=>(const time_zone& x, const time_zone& y) noexcept; // C++20 diff --git a/libcxx/modules/std/chrono.inc b/libcxx/modules/std/chrono.inc index bd3b6c3c12f29c6..f12c2425592e0bf 100644 --- a/libcxx/modules/std/chrono.inc +++ b/libcxx/modules/std/chrono.inc @@ -223,11 +223,8 @@ export namespace std { # if !defined(_LIBCPP_HAS_NO_TIME_ZONE_DATABASE) && !defined(_LIBCPP_HAS_NO_FILESYSTEM) && \ !defined(_LIBCPP_HAS_NO_LOCALIZATION) -# if 0 // [time.zone.timezone], class time_zone using std::chrono::choose; -# endif // if 0 - using std::chrono::time_zone; # if 0 diff --git a/libcxx/test/libcxx/diagnostics/chrono.nodiscard.verify.cpp b/libcxx/test/libcxx/diagnostics/chrono.nodiscard.verify.cpp index fea1e4417cc12c0..cba7916ff2c646f 100644 --- a/libcxx/test/libcxx/diagnostics/chrono.nodiscard.verify.cpp +++ b/libcxx/test/libcxx/diagnostics/chrono.nodiscard.verify.cpp @@ -49,9 +49,12 @@ void test() { { std::chrono::sys_seconds s{}; std::chrono::local_seconds l{}; + std::chrono::choose z = std::chrono::choose::earliest; tz.name(); // expected-warning {{ignoring return value of function declared with 'nodiscard' attribute}} tz.get_info(s); // expected-warning {{ignoring return value of function declared with 'nodiscard' attribute}} tz.get_info(l); // expected-warning {{ignoring return value of function declared with 'nodiscard' attribute}} + tz.to_sys(l); // not nodiscard + tz.to_sys(l, z); // expected-warning {{ignoring return value of function declared with 'nodiscard' attribute}} operator==(tz, tz); // expected-warning {{ignoring return value of function declared with 'nodiscard' attribute}} operator<=>(tz, tz); // expected-warning {{ignoring return value of function declared with 'nodiscard' attribute}} } diff --git a/libcxx/test/libcxx/time/time.zone/time.zone.timezone/choose.pass.cpp b/libcxx/test/libcxx/time/time.zone/time.zone.timezone/choose.pass.cpp new file mode 100644 index 000000000000000..23ef9c848096083 --- /dev/null +++ b/libcxx/test/libcxx/time/time.zone/time.zone.timezone/choose.pass.cpp @@ -0,0 +1,37 @@ +//===----------------------------------------------------------------------===// +// +// 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 +// +//===----------------------------------------------------------------------===// + +// UNSUPPORTED: c++03, c++11, c++14, c++17 +// UNSUPPORTED: no-filesystem, no-localization, no-tzdb + +// XFAIL: libcpp-has-no-experimental-tzdb +// XFAIL: availability-tzdb-missing + +// + +// enum class choose; + +#include +#include +#include + +#include "test_macros.h" + +int main(int, char**) { + using E = std::chrono::choose; + static_assert(std::is_enum_v); + + // Check that E is a scoped enum by checking for conversions. + using UT = std::underlying_type_t; + static_assert(!std::is_convertible_v); + + [[maybe_unused]] const E& early = E::earliest; + [[maybe_unused]] const E& late = E::latest; + + return 0; +} diff --git a/libcxx/test/libcxx/time/time.zone/time.zone.timezone/time.zone.members/assert.to_sys_choose.pass.cpp b/libcxx/test/libcxx/time/time.zone/time.zone.timezone/time.zone.members/assert.to_sys_choose.pass.cpp new file mode 100644 index 000000000000000..65429345ae794f7 --- /dev/null +++ b/libcxx/test/libcxx/time/time.zone/time.zone.timezone/time.zone.members/assert.to_sys_choose.pass.cpp @@ -0,0 +1,41 @@ +//===----------------------------------------------------------------------===// +// +// 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 +// +//===----------------------------------------------------------------------===// + +// UNSUPPORTED: c++03, c++11, c++14, c++17 + +// REQUIRES: has-unix-headers +// REQUIRES: libcpp-hardening-mode={{extensive|debug}} +// XFAIL: libcpp-hardening-mode=debug && availability-verbose_abort-missing + +// XFAIL: libcpp-has-no-experimental-tzdb + +// + +// template +// sys_time> +// to_sys(const local_time& tp, choose z) const; + +#include + +#include "check_assertion.h" + +// Tests values that cannot be converted. To make sure the test is does not depend on changes +// in the database it uses a time zone with a fixed offset. +int main(int, char**) { + TEST_LIBCPP_ASSERT_FAILURE( + std::chrono::locate_zone("Etc/GMT-1")->to_sys(std::chrono::local_seconds::min(), std::chrono::choose::earliest), + "cannot convert the local time; it would be before the minimum system clock value"); + + // TODO TZDB look why std::chrono::local_seconds::max() fails + TEST_LIBCPP_ASSERT_FAILURE( + std::chrono::locate_zone("Etc/GMT+1") + ->to_sys(std::chrono::local_seconds::max() - std::chrono::seconds(1), std::chrono::choose::latest), + "cannot convert the local time; it would be after the maximum system clock value"); + + return 0; +} diff --git a/libcxx/test/std/time/time.zone/time.zone.timezone/time.zone.members/to_sys_choose.pass.cpp b/libcxx/test/std/time/time.zone/time.zone.timezone/time.zone.members/to_sys_choose.pass.cpp new file mode 100644 index 000000000000000..bad4ef352e9b917 --- /dev/null +++ b/libcxx/test/std/time/time.zone/time.zone.timezone/time.zone.members/to_sys_choose.pass.cpp @@ -0,0 +1,147 @@ +//===----------------------------------------------------------------------===// +// +// 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 +// +//===----------------------------------------------------------------------===// + +// UNSUPPORTED: c++03, c++11, c++14, c++17 +// UNSUPPORTED: no-filesystem, no-localization, no-tzdb + +// XFAIL: libcpp-has-no-experimental-tzdb +// XFAIL: availability-tzdb-missing + +// + +// class time_zone; + +// template +// sys_time> +// to_sys(const local_time& tp, choose z) const; + +#include +#include +#include +#include + +#include "test_macros.h" + +// Tests unique conversions. To make sure the test is does not depend on changes +// in the database it uses a time zone with a fixed offset. +static void test_unique() { + using namespace std::literals::chrono_literals; + + const std::chrono::time_zone* tz = std::chrono::locate_zone("Etc/GMT+1"); + + assert(tz->to_sys(std::chrono::local_time{-1ns}, std::chrono::choose::earliest) == + std::chrono::sys_time{-1ns + 1h}); + + assert(tz->to_sys(std::chrono::local_time{0us}, std::chrono::choose::latest) == + std::chrono::sys_time{1h}); + + assert(tz->to_sys( + std::chrono::local_time{ + (std::chrono::sys_days{std::chrono::January / 1 / -21970}).time_since_epoch()}, + std::chrono::choose::earliest) == + std::chrono::sys_time{ + (std::chrono::sys_days{std::chrono::January / 1 / -21970}).time_since_epoch() + 1h}); + + // sys_time> is seconds for the larger types + assert(tz->to_sys( + std::chrono::local_time{ + (std::chrono::sys_days{std::chrono::January / 1 / 21970}).time_since_epoch()}, + std::chrono::choose::latest) == + std::chrono::sys_time{ + (std::chrono::sys_days{std::chrono::January / 1 / 21970}).time_since_epoch() + 1h}); + + assert(tz->to_sys(std::chrono::local_time{}, std::chrono::choose::earliest) == + std::chrono::sys_time{ + (std::chrono::sys_days{std::chrono::January / 1 / 1970}).time_since_epoch() + 1h}); + + // Note months and years cannot be streamed; however these functions don't + // throw an exception and thus can be used. + assert(tz->to_sys(std::chrono::local_time{}, std::chrono::choose::latest) == + std::chrono::sys_time{ + (std::chrono::sys_days{std::chrono::January / 1 / 1970}).time_since_epoch() + 1h}); + + assert(tz->to_sys(std::chrono::local_time{}, std::chrono::choose::earliest) == + std::chrono::sys_time{ + (std::chrono::sys_days{std::chrono::January / 1 / 1970}).time_since_epoch() + 1h}); +} + +// Tests non-existant conversions. +static void test_nonexistent() { + using namespace std::literals::chrono_literals; + + const std::chrono::time_zone* tz = std::chrono::locate_zone("Europe/Berlin"); + + // Z Europe/Berlin 0:53:28 - LMT 1893 Ap + // ... + // 1 DE CE%sT 1980 + // 1 E CE%sT + // + // ... + // R E 1981 ma - Mar lastSu 1u 1 S + // R E 1996 ma - O lastSu 1u 0 - + + // Pick an historic date where it's well known what the time zone rules were. + // This makes it unlikely updates to the database change these rules. + std::chrono::local_time time{ + (std::chrono::sys_days{std::chrono::March / 30 / 1986} + 2h + 30min).time_since_epoch()}; + + std::chrono::sys_seconds expected{time.time_since_epoch() - 1h}; + + // Validates whether the database did not change. + std::chrono::local_info info = tz->get_info(time); + assert(info.result == std::chrono::local_info::nonexistent); + + assert(tz->to_sys(time + 0ns, std::chrono::choose::earliest) == expected); + assert(tz->to_sys(time + 0us, std::chrono::choose::latest) == expected); + assert(tz->to_sys(time + 0ms, std::chrono::choose::earliest) == expected); + assert(tz->to_sys(time + 0s, std::chrono::choose::latest) == expected); +} + +// Tests ambiguous conversions. +static void test_ambiguous() { + using namespace std::literals::chrono_literals; + + const std::chrono::time_zone* tz = std::chrono::locate_zone("Europe/Berlin"); + + // Z Europe/Berlin 0:53:28 - LMT 1893 Ap + // ... + // 1 DE CE%sT 1980 + // 1 E CE%sT + // + // ... + // R E 1981 ma - Mar lastSu 1u 1 S + // R E 1996 ma - O lastSu 1u 0 - + + // Pick an historic date where it's well known what the time zone rules were. + // This makes it unlikely updates to the database change these rules. + std::chrono::local_time time{ + (std::chrono::sys_days{std::chrono::September / 28 / 1986} + 2h + 30min).time_since_epoch()}; + + std::chrono::sys_seconds earlier{time.time_since_epoch() - 2h}; + std::chrono::sys_seconds later{time.time_since_epoch() - 1h}; + + // Validates whether the database did not change. + std::chrono::local_info info = tz->get_info(time); + assert(info.result == std::chrono::local_info::ambiguous); + + assert(tz->to_sys(time + 0ns, std::chrono::choose::earliest) == earlier); + assert(tz->to_sys(time + 0us, std::chrono::choose::latest) == later); + assert(tz->to_sys(time + 0ms, std::chrono::choose::earliest) == earlier); + assert(tz->to_sys(time + 0s, std::chrono::choose::latest) == later); +} + +// This test does the basic validations of this function. The library function +// uses `local_info get_info(const local_time& tp)` as implementation +// detail. The get_info function does extensive testing of the data. +int main(int, char**) { + test_unique(); + test_nonexistent(); + test_ambiguous(); + + return 0; +} From a47e40bce086c506b672cbd8fbb2abdc6619c0a6 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Timm=20B=C3=A4der?= Date: Sun, 9 Jun 2024 21:13:55 +0200 Subject: [PATCH 11/36] [clang][Interp] Disallow ptr-to-int casts on dummy pointers --- clang/lib/AST/Interp/Interp.h | 9 +++++++++ clang/test/AST/Interp/builtin-align-cxx.cpp | 3 +-- clang/test/AST/Interp/const-eval.c | 9 +-------- 3 files changed, 11 insertions(+), 10 deletions(-) diff --git a/clang/lib/AST/Interp/Interp.h b/clang/lib/AST/Interp/Interp.h index f63711da90c7e0a..4bb9e7f0d22ea63 100644 --- a/clang/lib/AST/Interp/Interp.h +++ b/clang/lib/AST/Interp/Interp.h @@ -1937,6 +1937,9 @@ template ::T> bool CastPointerIntegral(InterpState &S, CodePtr OpPC) { const Pointer &Ptr = S.Stk.pop(); + if (Ptr.isDummy()) + return false; + const SourceInfo &E = S.Current->getSource(OpPC); S.CCEDiag(E, diag::note_constexpr_invalid_cast) << 2 << S.getLangOpts().CPlusPlus << S.Current->getRange(OpPC); @@ -1949,6 +1952,9 @@ static inline bool CastPointerIntegralAP(InterpState &S, CodePtr OpPC, uint32_t BitWidth) { const Pointer &Ptr = S.Stk.pop(); + if (Ptr.isDummy()) + return false; + const SourceInfo &E = S.Current->getSource(OpPC); S.CCEDiag(E, diag::note_constexpr_invalid_cast) << 2 << S.getLangOpts().CPlusPlus << S.Current->getRange(OpPC); @@ -1962,6 +1968,9 @@ static inline bool CastPointerIntegralAPS(InterpState &S, CodePtr OpPC, uint32_t BitWidth) { const Pointer &Ptr = S.Stk.pop(); + if (Ptr.isDummy()) + return false; + const SourceInfo &E = S.Current->getSource(OpPC); S.CCEDiag(E, diag::note_constexpr_invalid_cast) << 2 << S.getLangOpts().CPlusPlus << S.Current->getRange(OpPC); diff --git a/clang/test/AST/Interp/builtin-align-cxx.cpp b/clang/test/AST/Interp/builtin-align-cxx.cpp index c4103953df026a1..a1edf307d6c479c 100644 --- a/clang/test/AST/Interp/builtin-align-cxx.cpp +++ b/clang/test/AST/Interp/builtin-align-cxx.cpp @@ -202,8 +202,7 @@ static_assert(__builtin_align_down(&align32array[7], 4) == &align32array[4], "") static_assert(__builtin_align_down(&align32array[8], 4) == &align32array[8], ""); // Achieving the same thing using casts to uintptr_t is not allowed: -static_assert((char *)((__UINTPTR_TYPE__)&align32array[7] & ~3) == &align32array[4], ""); // both-error{{not an integral constant expression}} \ - // expected-note {{cast that performs the conversions of a reinterpret_cast is not allowed in a constant expression}} +static_assert((char *)((__UINTPTR_TYPE__)&align32array[7] & ~3) == &align32array[4], ""); // both-error{{not an integral constant expression}} static_assert(__builtin_align_down(&align32array[1], 4) == &align32array[0], ""); static_assert(__builtin_align_down(&align32array[1], 64) == &align32array[0], ""); // both-error{{not an integral constant expression}} diff --git a/clang/test/AST/Interp/const-eval.c b/clang/test/AST/Interp/const-eval.c index 72c0833a0f6309b..eab14c08ec809ac 100644 --- a/clang/test/AST/Interp/const-eval.c +++ b/clang/test/AST/Interp/const-eval.c @@ -140,15 +140,8 @@ EVAL_EXPR(47, &x < &x + 1 ? 1 : -1) EVAL_EXPR(48, &x != &x - 1 ? 1 : -1) EVAL_EXPR(49, &x < &x - 100 ? 1 : -1) // ref-error {{not an integer constant expression}} -/// FIXME: Rejecting this is correct, BUT when converting the innermost pointer -/// to an integer, we do not preserve the information where it came from. So when we later -/// create a pointer from it, it also doesn't have that information, which means -/// hasSameBase() for those two pointers will return false. And in those cases, we emit -/// the diagnostic: -/// comparison between '&Test50' and '&(631578)' has unspecified value extern struct Test50S Test50; -EVAL_EXPR(50, &Test50 < (struct Test50S*)((unsigned long)&Test50 + 10)) // both-error {{not an integer constant expression}} \ - // expected-note {{comparison between}} +EVAL_EXPR(50, &Test50 < (struct Test50S*)((unsigned long)&Test50 + 10)) // both-error {{not an integer constant expression}} EVAL_EXPR(51, 0 != (float)1e99) From a6929dbea8b4525ee41c1269e8e4a2b7182d9ce2 Mon Sep 17 00:00:00 2001 From: Nikita Popov Date: Mon, 10 Jun 2024 09:01:37 +0200 Subject: [PATCH 12/36] [lld] Remove const qualifier on symbolKind (NFC) (#94753) The symbol including this member is being overwritten by memcpy here: https://github.com/llvm/llvm-project/blob/2117677e304d334326f6591f3c75fb2f34dc4bcb/lld/COFF/SymbolTable.cpp#L496-L509 --- lld/COFF/Symbols.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/lld/COFF/Symbols.h b/lld/COFF/Symbols.h index ca69fb2d052706d..5ef46f5af6a6c8d 100644 --- a/lld/COFF/Symbols.h +++ b/lld/COFF/Symbols.h @@ -106,7 +106,7 @@ class Symbol { "If the name is empty, the Symbol must be a DefinedCOFF."); } - const unsigned symbolKind : 8; + unsigned symbolKind : 8; unsigned isExternal : 1; public: From 12d24e0c56a154c88247e55c7c352030e4d9073d Mon Sep 17 00:00:00 2001 From: Nikita Popov Date: Mon, 10 Jun 2024 09:19:55 +0200 Subject: [PATCH 13/36] [CodeGen] Simplify codegen for array initialization (#93956) This makes codegen for array initialization simpler in two ways: 1. Drop the zero-index GEP at the start, which is no longer needed with opaque pointers. 2. Emit GEPs directly to the correct element, instead of having a long chain of +1 GEPs. This is more canonical, and also avoids regressions in unoptimized builds from #93823. --- clang/lib/CodeGen/CGExprAgg.cpp | 29 +- clang/test/CodeGen/paren-list-agg-init.cpp | 44 +- .../CodeGenCXX/control-flow-in-stmt-expr.cpp | 32 +- .../cxx0x-initializer-references.cpp | 3 +- ...nitializer-stdinitializerlist-startend.cpp | 3 +- .../cxx0x-initializer-stdinitializerlist.cpp | 3 +- .../cxx11-initializer-array-new.cpp | 42 +- clang/test/CodeGenCXX/partial-destruction.cpp | 38 +- clang/test/CodeGenCXX/temporaries.cpp | 5 +- clang/test/CodeGenCXX/value-init.cpp | 23 +- .../coro-suspend-cleanups.cpp | 17 +- clang/test/CodeGenObjC/arc-ternary-op.m | 10 +- clang/test/CodeGenObjC/arc.m | 5 +- clang/test/CodeGenObjCXX/arc-exceptions.mm | 19 +- .../distribute_firstprivate_codegen.cpp | 20 +- .../OpenMP/distribute_lastprivate_codegen.cpp | 20 +- ...bute_parallel_for_firstprivate_codegen.cpp | 20 +- ...ibute_parallel_for_lastprivate_codegen.cpp | 20 +- ...istribute_parallel_for_private_codegen.cpp | 20 +- ...parallel_for_simd_firstprivate_codegen.cpp | 40 +- ..._parallel_for_simd_lastprivate_codegen.cpp | 40 +- ...bute_parallel_for_simd_private_codegen.cpp | 40 +- .../OpenMP/distribute_private_codegen.cpp | 20 +- .../distribute_simd_firstprivate_codegen.cpp | 40 +- .../distribute_simd_lastprivate_codegen.cpp | 40 +- .../distribute_simd_private_codegen.cpp | 40 +- .../test/OpenMP/for_firstprivate_codegen.cpp | 5 +- clang/test/OpenMP/for_lastprivate_codegen.cpp | 20 +- clang/test/OpenMP/for_private_codegen.cpp | 10 +- clang/test/OpenMP/for_reduction_codegen.cpp | 14 +- .../test/OpenMP/for_reduction_codegen_UDR.cpp | 28 +- clang/test/OpenMP/parallel_copyin_codegen.cpp | 10 +- .../OpenMP/parallel_firstprivate_codegen.cpp | 20 +- ...l_master_taskloop_firstprivate_codegen.cpp | 30 +- ...el_master_taskloop_lastprivate_codegen.cpp | 10 +- ...ter_taskloop_simd_firstprivate_codegen.cpp | 30 +- ...ster_taskloop_simd_lastprivate_codegen.cpp | 20 +- .../test/OpenMP/parallel_private_codegen.cpp | 10 +- .../OpenMP/parallel_reduction_codegen.cpp | 10 +- .../OpenMP/sections_firstprivate_codegen.cpp | 5 +- .../OpenMP/sections_lastprivate_codegen.cpp | 20 +- .../test/OpenMP/sections_private_codegen.cpp | 10 +- .../OpenMP/sections_reduction_codegen.cpp | 10 +- .../OpenMP/simd_private_taskloop_codegen.cpp | 40 +- .../OpenMP/single_firstprivate_codegen.cpp | 5 +- clang/test/OpenMP/single_private_codegen.cpp | 10 +- ..._teams_distribute_firstprivate_codegen.cpp | 10 +- ...t_teams_distribute_lastprivate_codegen.cpp | 20 +- ...bute_parallel_for_firstprivate_codegen.cpp | 10 +- ...ibute_parallel_for_lastprivate_codegen.cpp | 20 +- ...istribute_parallel_for_private_codegen.cpp | 10 +- ...parallel_for_simd_firstprivate_codegen.cpp | 20 +- ..._parallel_for_simd_lastprivate_codegen.cpp | 40 +- ...bute_parallel_for_simd_private_codegen.cpp | 20 +- ...arget_teams_distribute_private_codegen.cpp | 10 +- ...s_distribute_simd_firstprivate_codegen.cpp | 20 +- ...ms_distribute_simd_lastprivate_codegen.cpp | 40 +- ..._teams_distribute_simd_private_codegen.cpp | 20 +- ...get_teams_generic_loop_private_codegen.cpp | 10 +- .../teams_distribute_firstprivate_codegen.cpp | 10 +- .../teams_distribute_lastprivate_codegen.cpp | 20 +- ...bute_parallel_for_firstprivate_codegen.cpp | 10 +- ...ibute_parallel_for_lastprivate_codegen.cpp | 20 +- ...istribute_parallel_for_private_codegen.cpp | 10 +- ...parallel_for_simd_firstprivate_codegen.cpp | 20 +- ..._parallel_for_simd_lastprivate_codegen.cpp | 40 +- ...bute_parallel_for_simd_private_codegen.cpp | 20 +- .../teams_distribute_private_codegen.cpp | 10 +- ...s_distribute_simd_firstprivate_codegen.cpp | 20 +- ...ms_distribute_simd_lastprivate_codegen.cpp | 40 +- .../teams_distribute_simd_private_codegen.cpp | 20 +- .../OpenMP/teams_firstprivate_codegen.cpp | 20 +- .../teams_generic_loop_private_codegen.cpp | 10 +- clang/test/OpenMP/teams_private_codegen.cpp | 20 +- clang/test/OpenMP/threadprivate_codegen.cpp | 468 +++++++++--------- clang/test/PCH/cxx_paren_init.cpp | 9 +- 76 files changed, 835 insertions(+), 1132 deletions(-) diff --git a/clang/lib/CodeGen/CGExprAgg.cpp b/clang/lib/CodeGen/CGExprAgg.cpp index 5b2039af6128bf0..b2a5ceeeae08ba1 100644 --- a/clang/lib/CodeGen/CGExprAgg.cpp +++ b/clang/lib/CodeGen/CGExprAgg.cpp @@ -513,15 +513,6 @@ void AggExprEmitter::EmitArrayInit(Address DestPtr, llvm::ArrayType *AType, QualType elementType = CGF.getContext().getAsArrayType(ArrayQTy)->getElementType(); - - // DestPtr is an array*. Construct an elementType* by drilling - // down a level. - llvm::Value *zero = llvm::ConstantInt::get(CGF.SizeTy, 0); - llvm::Value *indices[] = { zero, zero }; - llvm::Value *begin = Builder.CreateInBoundsGEP(DestPtr.getElementType(), - DestPtr.emitRawPointer(CGF), - indices, "arrayinit.begin"); - CharUnits elementSize = CGF.getContext().getTypeSizeInChars(elementType); CharUnits elementAlign = DestPtr.getAlignment().alignmentOfArrayElement(elementSize); @@ -562,6 +553,7 @@ void AggExprEmitter::EmitArrayInit(Address DestPtr, llvm::ArrayType *AType, Address endOfInit = Address::invalid(); CodeGenFunction::CleanupDeactivationScope deactivation(CGF); + llvm::Value *begin = DestPtr.emitRawPointer(CGF); if (dtorKind) { CodeGenFunction::AllocaTrackerRAII allocaTracker(CGF); // In principle we could tell the cleanup where we are more @@ -585,19 +577,13 @@ void AggExprEmitter::EmitArrayInit(Address DestPtr, llvm::ArrayType *AType, llvm::Value *one = llvm::ConstantInt::get(CGF.SizeTy, 1); - // The 'current element to initialize'. The invariants on this - // variable are complicated. Essentially, after each iteration of - // the loop, it points to the last initialized element, except - // that it points to the beginning of the array before any - // elements have been initialized. - llvm::Value *element = begin; - // Emit the explicit initializers. for (uint64_t i = 0; i != NumInitElements; ++i) { - // Advance to the next element. + llvm::Value *element = begin; if (i > 0) { - element = Builder.CreateInBoundsGEP( - llvmElementType, element, one, "arrayinit.element"); + element = Builder.CreateInBoundsGEP(llvmElementType, begin, + llvm::ConstantInt::get(CGF.SizeTy, i), + "arrayinit.element"); // Tell the cleanup that it needs to destroy up to this // element. TODO: some of these stores can be trivially @@ -624,9 +610,12 @@ void AggExprEmitter::EmitArrayInit(Address DestPtr, llvm::ArrayType *AType, // do { *array++ = filler; } while (array != end); // Advance to the start of the rest of the array. + llvm::Value *element = begin; if (NumInitElements) { element = Builder.CreateInBoundsGEP( - llvmElementType, element, one, "arrayinit.start"); + llvmElementType, element, + llvm::ConstantInt::get(CGF.SizeTy, NumInitElements), + "arrayinit.start"); if (endOfInit.isValid()) Builder.CreateStore(element, endOfInit); } diff --git a/clang/test/CodeGen/paren-list-agg-init.cpp b/clang/test/CodeGen/paren-list-agg-init.cpp index 94d42431d125d45..88b1834d42d8791 100644 --- a/clang/test/CodeGen/paren-list-agg-init.cpp +++ b/clang/test/CodeGen/paren-list-agg-init.cpp @@ -271,14 +271,13 @@ const int* foo10() { // CHECK-NEXT: [[ARR_2:%.*]] = alloca [4 x i32], align 16 // CHECK-NEXT: store i32 [[A:%.*]], ptr [[A_ADDR]], align 4 // CHECK-NEXT: store i32 [[B:%.*]], ptr [[B_ADDR]], align 4 -// CHECK-NEXT: [[ARRINIT_BEGIN:%.*]] = getelementptr inbounds [4 x i32], ptr [[ARR_2]], i64 0, i64 0 // CHECK-NEXT: [[TMP_0:%.*]] = load i32, ptr [[A_ADDR]], align 4 -// CHECK-NEXT: store i32 [[TMP_0]], ptr [[ARRINIT_BEGIN]], align 4 -// CHECK-NEXT: [[ARRINIT_ELEM:%.*]] = getelementptr inbounds i32, ptr [[ARRINIT_BEGIN]], i64 1 +// CHECK-NEXT: store i32 [[TMP_0]], ptr [[ARR_2]], align 4 +// CHECK-NEXT: [[ARRINIT_ELEM:%.*]] = getelementptr inbounds i32, ptr [[ARR_2]], i64 1 // CHECK-NEXT: [[TMP_1:%.*]] = load i32, ptr [[B_ADDR]], align 4 // CHECK-NEXT: store i32 [[TMP_1]], ptr [[ARRINIT_ELEM]], align 4 -// CHECK-NEXT: [[ARRINIT_START:%.*]] = getelementptr inbounds i32, ptr [[ARRINIT_ELEM]], i64 1 -// CHECK-NEXT: [[ARRINIT_END:%.*]] = getelementptr inbounds i32, ptr [[ARRINIT_BEGIN]], i64 4 +// CHECK-NEXT: [[ARRINIT_START:%.*]] = getelementptr inbounds i32, ptr [[ARR_2]], i64 2 +// CHECK-NEXT: [[ARRINIT_END:%.*]] = getelementptr inbounds i32, ptr [[ARR_2]], i64 4 // CHECK-NEXT: br label [[ARRINIT_BODY:%.*]] // CHECK: [[ARRINIT_CUR:%.*]] = phi ptr [ [[ARRINIT_START]], %entry ], [ [[ARRINIT_NEXT:%.*]], [[ARRINIT_BODY]] ] // CHECK-NEXT: store i32 0, ptr [[ARRINIT_CUR]], align 4 @@ -297,10 +296,9 @@ void foo11(int a, int b) { // CHECK-NEXT: [[ARR_3:%.*]] = alloca [2 x i32], align 4 // CHECK-NEXT: store i32 [[A:%.*]], ptr [[A_ADDR]], align 4 // CHECK-NEXT: store i32 [[B:%.*]], ptr [[B_ADDR]], align 4 -// CHECK-NEXT: [[ARRINIT_BEGIN:%.*]] = getelementptr inbounds [2 x i32], ptr [[ARR_3]], i64 0, i64 0 // CHECK-NEXT: [[TMP_0:%.*]] = load i32, ptr [[A_ADDR]], align 4 -// CHECK-NEXT: store i32 [[TMP_0]], ptr [[ARRINIT_BEGIN]], align 4 -// CHECK-NEXT: [[ARRINIT_ELEMENT:%.*]] = getelementptr inbounds i32, ptr [[ARRINIT_BEGIN]], i64 1 +// CHECK-NEXT: store i32 [[TMP_0]], ptr [[ARR_3]], align 4 +// CHECK-NEXT: [[ARRINIT_ELEMENT:%.*]] = getelementptr inbounds i32, ptr [[ARR_3]], i64 1 // CHECK-NEXT: [[TMP_1:%.*]] = load i32, ptr [[B_ADDR]], align 4 // CHECK-NEXT: store i32 [[TMP_1]], ptr [[ARRINIT_ELEMENT]], align 4 // CHECK-NEXT: ret void @@ -336,8 +334,7 @@ const int* foo15() { // CHECK-NEXT: entry: // CHECK-NEXT: [[ARR_6:%.*arr6.*]] = alloca ptr, align 8 // CHECK-NEXT: [[REF_TMP:%.*]] = alloca [1 x i32], align 4 -// CHECK-NEXT: [[ARRINIT_BEGIN:%.*]] = getelementptr inbounds [1 x i32], ptr [[REF_TMP]], i64 0, i64 0 -// CHECK-NEXT: store i32 3, ptr [[ARRINIT_BEGIN]], align 4 +// CHECK-NEXT: store i32 3, ptr [[REF_TMP]], align 4 // CHECK-NEXT: store ptr [[REF_TMP]], ptr [[ARR_6]], align 8 // CHECK-NEXT: ret void void foo16() { @@ -348,10 +345,9 @@ void foo16() { // CHECK-NEXT: entry: // CHECK-NEXT: [[ARR_7:%.*arr7.*]] = alloca ptr, align 8 // CHECK-NEXT: [[REF_TMP:%.*]] = alloca [2 x i32], align 4 -// CHECK-NEXT: [[ARRINIT_BEGIN:%.*]] = getelementptr inbounds [2 x i32], ptr [[REF_TMP]], i64 0, i64 0 -// CHECK-NEXT: store i32 4, ptr [[ARRINIT_BEGIN]], align 4 -// CHECK-NEXT: [[ARRINIT_START:%.*]] = getelementptr inbounds i32, ptr [[ARRINIT_BEGIN]], i64 1 -// CHECK-NEXT: [[ARRINIT_END:%.*]] = getelementptr inbounds i32, ptr [[ARRINIT_BEGIN]], i64 2 +// CHECK-NEXT: store i32 4, ptr [[REF_TMP]], align 4 +// CHECK-NEXT: [[ARRINIT_START:%.*]] = getelementptr inbounds i32, ptr [[REF_TMP]], i64 1 +// CHECK-NEXT: [[ARRINIT_END:%.*]] = getelementptr inbounds i32, ptr [[REF_TMP]], i64 2 // CHECK-NEXT: br label [[ARRINIT_BODY]] // CHECK: [[ARRINIT_CUR:%.*]] = phi ptr [ [[ARRINIT_START]], %entry ], [ [[ARRINIT_NEXT:%.*]], [[ARRINIT_BODY]] ] // CHECK-NEXT: store i32 0, ptr [[ARRINIT_CUR]], align 4 @@ -533,14 +529,12 @@ namespace gh68198 { // CHECK-NEXT: entry // CHECK-NEXT: [[ARR_10:%.*arr9.*]] = alloca ptr, align 8 // CHECK-NEXT: [[CALL_PTR]] = call noalias noundef nonnull ptr @_Znam(i64 noundef 16) - // CHECK-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x i32], ptr [[CALL]], i64 0, i64 0 - // CHECK-NEXT: store i32 1, ptr [[ARRAYINIT_BEGIN]], align 4 - // CHECK-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds i32, ptr [[ARRAYINIT_BEGIN]], i64 1 + // CHECK-NEXT: store i32 1, ptr [[CALL]], align 4 + // CHECK-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds i32, ptr [[CALL]], i64 1 // CHECK-NEXT: store i32 2, ptr [[ARRAYINIT_ELEMENT]], align 4 // CHECK-NEXT: [[ARRAY_EXP_NEXT:%.*]] = getelementptr inbounds [2 x i32], ptr %call, i64 1 - // CHECK-NEXT: [[ARRAYINIT_BEGIN1:%.*]] = getelementptr inbounds [2 x i32], ptr [[ARRAY_EXP_NEXT]], i64 0, i64 0 - // CHECK-NEXT: store i32 3, ptr [[ARRAYINIT_BEGIN1]], align 4 - // CHECK-NEXT: [[ARRAYINIT_ELEMENT2:%.*]] = getelementptr inbounds i32, ptr [[ARRAYINIT_BEGIN1]], i64 1 + // CHECK-NEXT: store i32 3, ptr [[ARRAY_EXP_NEXT]], align 4 + // CHECK-NEXT: [[ARRAYINIT_ELEMENT2:%.*]] = getelementptr inbounds i32, ptr [[ARRAY_EXP_NEXT]], i64 1 // CHECK-NEXT: store i32 4, ptr [[ARRAYINIT_ELEMENT2]], align 4 // CHECK-NEXT: [[ARRAY_EXP_NEXT3:%.*]] = getelementptr inbounds [2 x i32], ptr [[ARRAY_EXP_NEXT]], i64 1 // CHECK-NEXT: store ptr [[CALL_PTR]], ptr [[ARR_10]], align 8 @@ -553,14 +547,12 @@ namespace gh68198 { // CHECK-NEXT: entry // CHECK-NEXT: [[ARR_10:%.*arr10.*]] = alloca ptr, align 8 // CHECK-NEXT: [[CALL_PTR]] = call noalias noundef nonnull ptr @_Znam(i64 noundef 32) - // CHECK-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x i32], ptr [[CALL]], i64 0, i64 0 - // CHECK-NEXT: store i32 5, ptr [[ARRAYINIT_BEGIN]], align 4 - // CHECK-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds i32, ptr [[ARRAYINIT_BEGIN]], i64 1 + // CHECK-NEXT: store i32 5, ptr [[CALL]], align 4 + // CHECK-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds i32, ptr [[CALL]], i64 1 // CHECK-NEXT: store i32 6, ptr [[ARRAYINIT_ELEMENT]], align 4 // CHECK-NEXT: [[ARRAY_EXP_NEXT:%.*]] = getelementptr inbounds [2 x i32], ptr %call, i64 1 - // CHECK-NEXT: [[ARRAYINIT_BEGIN1:%.*]] = getelementptr inbounds [2 x i32], ptr [[ARRAY_EXP_NEXT]], i64 0, i64 0 - // CHECK-NEXT: store i32 7, ptr [[ARRAYINIT_BEGIN1]], align 4 - // CHECK-NEXT: [[ARRAYINIT_ELEMENT2:%.*]] = getelementptr inbounds i32, ptr [[ARRAYINIT_BEGIN1]], i64 1 + // CHECK-NEXT: store i32 7, ptr [[ARRAY_EXP_NEXT]], align 4 + // CHECK-NEXT: [[ARRAYINIT_ELEMENT2:%.*]] = getelementptr inbounds i32, ptr [[ARRAY_EXP_NEXT]], i64 1 // CHECK-NEXT: store i32 8, ptr [[ARRAYINIT_ELEMENT2]], align 4 // CHECK-NEXT: [[ARRAY_EXP_NEXT3:%.*]] = getelementptr inbounds [2 x i32], ptr [[ARRAY_EXP_NEXT]], i64 1 // CHECK-NEXT: call void @llvm.memset.p0.i64(ptr align 4 [[ARRAY_EXP_NEXT3]], i8 0, i64 16, i1 false) diff --git a/clang/test/CodeGenCXX/control-flow-in-stmt-expr.cpp b/clang/test/CodeGenCXX/control-flow-in-stmt-expr.cpp index ac466ee5bba40b0..4eafa720e0cb41b 100644 --- a/clang/test/CodeGenCXX/control-flow-in-stmt-expr.cpp +++ b/clang/test/CodeGenCXX/control-flow-in-stmt-expr.cpp @@ -158,16 +158,15 @@ void ArrayInit() { // CHECK-LABEL: define dso_local void @_Z9ArrayInitv() // CHECK: %arrayinit.endOfInit = alloca ptr, align 8 // CHECK: %cleanup.dest.slot = alloca i32, align 4 - // CHECK: %arrayinit.begin = getelementptr inbounds [4 x %struct.Printy], ptr %arr, i64 0, i64 0 - // CHECK: store ptr %arrayinit.begin, ptr %arrayinit.endOfInit, align 8 + // CHECK: store ptr %arr, ptr %arrayinit.endOfInit, align 8 Printy arr[4] = { Printy("a"), - // CHECK: call void @_ZN6PrintyC1EPKc(ptr noundef nonnull align 8 dereferenceable(8) %arrayinit.begin, ptr noundef @.str) - // CHECK: [[ARRAYINIT_ELEMENT1:%.+]] = getelementptr inbounds %struct.Printy, ptr %arrayinit.begin, i64 1 + // CHECK: call void @_ZN6PrintyC1EPKc(ptr noundef nonnull align 8 dereferenceable(8) %arr, ptr noundef @.str) + // CHECK: [[ARRAYINIT_ELEMENT1:%.+]] = getelementptr inbounds %struct.Printy, ptr %arr, i64 1 // CHECK: store ptr [[ARRAYINIT_ELEMENT1]], ptr %arrayinit.endOfInit, align 8 Printy("b"), // CHECK: call void @_ZN6PrintyC1EPKc(ptr noundef nonnull align 8 dereferenceable(8) [[ARRAYINIT_ELEMENT1]], ptr noundef @.str.1) - // CHECK: [[ARRAYINIT_ELEMENT2:%.+]] = getelementptr inbounds %struct.Printy, ptr [[ARRAYINIT_ELEMENT1]], i64 1 + // CHECK: [[ARRAYINIT_ELEMENT2:%.+]] = getelementptr inbounds %struct.Printy, ptr %arr, i64 2 // CHECK: store ptr [[ARRAYINIT_ELEMENT2]], ptr %arrayinit.endOfInit, align 8 ({ // CHECK: br i1 {{.*}}, label %if.then, label %if.end @@ -180,7 +179,7 @@ void ArrayInit() { // CHECK: if.end: Printy("c"); // CHECK-NEXT: call void @_ZN6PrintyC1EPKc - // CHECK-NEXT: %arrayinit.element2 = getelementptr inbounds %struct.Printy, ptr %arrayinit.element1, i64 1 + // CHECK-NEXT: %arrayinit.element2 = getelementptr inbounds %struct.Printy, ptr %arr, i64 3 // CHECK-NEXT: store ptr %arrayinit.element2, ptr %arrayinit.endOfInit, align 8 }), ({ @@ -212,14 +211,14 @@ void ArrayInit() { // CHECK: cleanup: // CHECK-NEXT: %1 = load ptr, ptr %arrayinit.endOfInit, align 8 - // CHECK-NEXT: %arraydestroy.isempty = icmp eq ptr %arrayinit.begin, %1 + // CHECK-NEXT: %arraydestroy.isempty = icmp eq ptr %arr, %1 // CHECK-NEXT: br i1 %arraydestroy.isempty, label %[[ARRAY_DESTROY_DONE2:.+]], label %[[ARRAY_DESTROY_BODY2:.+]] // CHECK: [[ARRAY_DESTROY_BODY2]]: // CHECK-NEXT: %arraydestroy.elementPast = phi ptr [ %1, %cleanup ], [ %arraydestroy.element, %[[ARRAY_DESTROY_BODY2]] ] // CHECK-NEXT: %arraydestroy.element = getelementptr inbounds %struct.Printy, ptr %arraydestroy.elementPast, i64 -1 // CHECK-NEXT: call void @_ZN6PrintyD1Ev(ptr noundef nonnull align 8 dereferenceable(8) %arraydestroy.element) - // CHECK-NEXT: %arraydestroy.done = icmp eq ptr %arraydestroy.element, %arrayinit.begin + // CHECK-NEXT: %arraydestroy.done = icmp eq ptr %arraydestroy.element, %arr // CHECK-NEXT: br i1 %arraydestroy.done, label %[[ARRAY_DESTROY_DONE2]], label %[[ARRAY_DESTROY_BODY2]] // CHECK: [[ARRAY_DESTROY_DONE2]]: @@ -238,8 +237,7 @@ void ArraySubobjects() { // CHECK: call void @_ZN6PrintyC1EPKc // CHECK: call void @_ZN6PrintyC1EPKc {Printy("a"), - // CHECK: [[ARRAYINIT_BEGIN:%.+]] = getelementptr inbounds [2 x %struct.Printy] - // CHECK: store ptr [[ARRAYINIT_BEGIN]], ptr %arrayinit.endOfInit, align 8 + // CHECK: store ptr %arr2, ptr %arrayinit.endOfInit, align 8 // CHECK: call void @_ZN6PrintyC1EPKc // CHECK: [[ARRAYINIT_ELEMENT:%.+]] = getelementptr inbounds %struct.Printy // CHECK: store ptr [[ARRAYINIT_ELEMENT]], ptr %arrayinit.endOfInit, align 8 @@ -248,7 +246,7 @@ void ArraySubobjects() { return; // CHECK: if.then: // CHECK-NEXT: [[V0:%.+]] = load ptr, ptr %arrayinit.endOfInit, align 8 - // CHECK-NEXT: %arraydestroy.isempty = icmp eq ptr [[ARRAYINIT_BEGIN]], [[V0]] + // CHECK-NEXT: %arraydestroy.isempty = icmp eq ptr %arr2, [[V0]] // CHECK-NEXT: br i1 %arraydestroy.isempty, label %[[ARRAY_DESTROY_DONE:.+]], label %[[ARRAY_DESTROY_BODY:.+]] } Printy("b"); @@ -268,7 +266,7 @@ void ArraySubobjects() { // CHECK-NEXT: %arraydestroy.elementPast = phi ptr [ %0, %if.then ], [ %arraydestroy.element, %[[ARRAY_DESTROY_BODY]] ] // CHECK-NEXT: %arraydestroy.element = getelementptr inbounds %struct.Printy, ptr %arraydestroy.elementPast, i64 -1 // CHECK-NEXT: call void @_ZN6PrintyD1Ev(ptr noundef nonnull align 8 dereferenceable(8) %arraydestroy.element) - // CHECK-NEXT: %arraydestroy.done = icmp eq ptr %arraydestroy.element, [[ARRAYINIT_BEGIN]] + // CHECK-NEXT: %arraydestroy.done = icmp eq ptr %arraydestroy.element, %arr2 // CHECK-NEXT: br i1 %arraydestroy.done, label %[[ARRAY_DESTROY_DONE]], label %[[ARRAY_DESTROY_BODY]] // CHECK: [[ARRAY_DESTROY_DONE]] @@ -277,11 +275,11 @@ void ArraySubobjects() { // CHECK-NEXT: br label %[[ARRAY_DESTROY_BODY2:.+]] // CHECK: [[ARRAY_DESTROY_BODY2]]: - // CHECK-NEXT: %arraydestroy.elementPast5 = phi ptr [ %1, %[[ARRAY_DESTROY_DONE]] ], [ %arraydestroy.element6, %[[ARRAY_DESTROY_BODY2]] ] - // CHECK-NEXT: %arraydestroy.element6 = getelementptr inbounds %struct.Printy, ptr %arraydestroy.elementPast5, i64 -1 - // CHECK-NEXT: call void @_ZN6PrintyD1Ev(ptr noundef nonnull align 8 dereferenceable(8) %arraydestroy.element6) - // CHECK-NEXT: %arraydestroy.done7 = icmp eq ptr %arraydestroy.element6, [[ARRAY_BEGIN]] - // CHECK-NEXT: br i1 %arraydestroy.done7, label %[[ARRAY_DESTROY_DONE2:.+]], label %[[ARRAY_DESTROY_BODY2]] + // CHECK-NEXT: %arraydestroy.elementPast4 = phi ptr [ %1, %[[ARRAY_DESTROY_DONE]] ], [ %arraydestroy.element5, %[[ARRAY_DESTROY_BODY2]] ] + // CHECK-NEXT: %arraydestroy.element5 = getelementptr inbounds %struct.Printy, ptr %arraydestroy.elementPast4, i64 -1 + // CHECK-NEXT: call void @_ZN6PrintyD1Ev(ptr noundef nonnull align 8 dereferenceable(8) %arraydestroy.element5) + // CHECK-NEXT: %arraydestroy.done6 = icmp eq ptr %arraydestroy.element5, [[ARRAY_BEGIN]] + // CHECK-NEXT: br i1 %arraydestroy.done6, label %[[ARRAY_DESTROY_DONE2:.+]], label %[[ARRAY_DESTROY_BODY2]] // CHECK: [[ARRAY_DESTROY_DONE2]]: diff --git a/clang/test/CodeGenCXX/cxx0x-initializer-references.cpp b/clang/test/CodeGenCXX/cxx0x-initializer-references.cpp index b19ca1d861b11c7..419525d0544d52e 100644 --- a/clang/test/CodeGenCXX/cxx0x-initializer-references.cpp +++ b/clang/test/CodeGenCXX/cxx0x-initializer-references.cpp @@ -52,11 +52,10 @@ namespace reference { // CHECK-NEXT: store ptr %{{.*}}, ptr %{{.*}}, align const A &ra1{1, i}; - // CHECK-NEXT: getelementptr inbounds [3 x i32], ptr %{{.*}}, i{{32|64}} 0, i{{32|64}} 0 // CHECK-NEXT: store i32 1 // CHECK-NEXT: getelementptr inbounds i32, ptr %{{.*}}, i{{32|64}} 1 // CHECK-NEXT: store i32 2 - // CHECK-NEXT: getelementptr inbounds i32, ptr %{{.*}}, i{{32|64}} 1 + // CHECK-NEXT: getelementptr inbounds i32, ptr %{{.*}}, i{{32|64}} 2 // CHECK-NEXT: %[[I2:.*]] = load i32, ptr // CHECK-NEXT: store i32 %[[I2]] // CHECK-NEXT: store ptr %{{.*}}, ptr %{{.*}}, align diff --git a/clang/test/CodeGenCXX/cxx0x-initializer-stdinitializerlist-startend.cpp b/clang/test/CodeGenCXX/cxx0x-initializer-stdinitializerlist-startend.cpp index d9e4c5d5403c251..e2d566121331fdc 100644 --- a/clang/test/CodeGenCXX/cxx0x-initializer-stdinitializerlist-startend.cpp +++ b/clang/test/CodeGenCXX/cxx0x-initializer-stdinitializerlist-startend.cpp @@ -40,8 +40,7 @@ void fn1(int i) { // CHECK-LABEL: define{{.*}} void @_Z3fn1i // temporary array // CHECK: [[array:%[^ ]+]] = alloca [3 x i32] - // CHECK: getelementptr inbounds [3 x i32], ptr [[array]], i{{32|64}} 0 - // CHECK-NEXT: store i32 1, ptr + // CHECK: store i32 1, ptr // CHECK-NEXT: getelementptr // CHECK-NEXT: store // CHECK-NEXT: getelementptr diff --git a/clang/test/CodeGenCXX/cxx0x-initializer-stdinitializerlist.cpp b/clang/test/CodeGenCXX/cxx0x-initializer-stdinitializerlist.cpp index aa2f078a5fb0ca6..3d0cf968bfd54cb 100644 --- a/clang/test/CodeGenCXX/cxx0x-initializer-stdinitializerlist.cpp +++ b/clang/test/CodeGenCXX/cxx0x-initializer-stdinitializerlist.cpp @@ -122,8 +122,7 @@ void fn1(int i) { // X86: [[array:%[^ ]+]] = alloca [3 x i32] // AMDGCN: [[alloca:%[^ ]+]] = alloca [3 x i32], align 4, addrspace(5) // AMDGCN: [[array:%[^ ]+]] ={{.*}} addrspacecast ptr addrspace(5) [[alloca]] to ptr - // CHECK: getelementptr inbounds [3 x i32], ptr [[array]], i{{32|64}} 0 - // CHECK-NEXT: store i32 1, ptr + // CHECK: store i32 1, ptr // CHECK-NEXT: getelementptr // CHECK-NEXT: store // CHECK-NEXT: getelementptr diff --git a/clang/test/CodeGenCXX/cxx11-initializer-array-new.cpp b/clang/test/CodeGenCXX/cxx11-initializer-array-new.cpp index 48ee019a55f555a..d86c712372cd89c 100644 --- a/clang/test/CodeGenCXX/cxx11-initializer-array-new.cpp +++ b/clang/test/CodeGenCXX/cxx11-initializer-array-new.cpp @@ -16,22 +16,20 @@ void *p = new S[2][3]{ { 1, 2, 3 }, { 4, 5, 6 } }; // { 1, 2, 3 } // // -// CHECK: %[[S_0_0:.*]] = getelementptr inbounds [3 x %[[S:.*]]], ptr %[[START_AS_i8]], i64 0, i64 0 -// CHECK: call void @_ZN1SC1Ei(ptr {{[^,]*}} %[[S_0_0]], i32 noundef 1) -// CHECK: %[[S_0_1:.*]] = getelementptr inbounds %[[S]], ptr %[[S_0_0]], i64 1 +// CHECK: call void @_ZN1SC1Ei(ptr {{[^,]*}} %[[START_AS_i8]], i32 noundef 1) +// CHECK: %[[S_0_1:.*]] = getelementptr inbounds %[[S:.+]], ptr %[[START_AS_i8]], i64 1 // CHECK: call void @_ZN1SC1Ei(ptr {{[^,]*}} %[[S_0_1]], i32 noundef 2) -// CHECK: %[[S_0_2:.*]] = getelementptr inbounds %[[S]], ptr %[[S_0_1]], i64 1 +// CHECK: %[[S_0_2:.*]] = getelementptr inbounds %[[S]], ptr %[[START_AS_i8]], i64 2 // CHECK: call void @_ZN1SC1Ei(ptr {{[^,]*}} %[[S_0_2]], i32 noundef 3) // // { 4, 5, 6 } // // CHECK: %[[S_1:.*]] = getelementptr inbounds [3 x %[[S]]], ptr %[[START_AS_i8]], i64 1 // -// CHECK: %[[S_1_0:.*]] = getelementptr inbounds [3 x %[[S]]], ptr %[[S_1]], i64 0, i64 0 -// CHECK: call void @_ZN1SC1Ei(ptr {{[^,]*}} %[[S_1_0]], i32 noundef 4) -// CHECK: %[[S_1_1:.*]] = getelementptr inbounds %[[S]], ptr %[[S_1_0]], i64 1 +// CHECK: call void @_ZN1SC1Ei(ptr {{[^,]*}} %[[S_1]], i32 noundef 4) +// CHECK: %[[S_1_1:.*]] = getelementptr inbounds %[[S]], ptr %[[S_1]], i64 1 // CHECK: call void @_ZN1SC1Ei(ptr {{[^,]*}} %[[S_1_1]], i32 noundef 5) -// CHECK: %[[S_1_2:.*]] = getelementptr inbounds %[[S]], ptr %[[S_1_1]], i64 1 +// CHECK: %[[S_1_2:.*]] = getelementptr inbounds %[[S]], ptr %[[S_1]], i64 2 // CHECK: call void @_ZN1SC1Ei(ptr {{[^,]*}} %[[S_1_2]], i32 noundef 6) // // CHECK-NOT: br i1 @@ -57,22 +55,20 @@ void *q = new S[n][3]{ { 1, 2, 3 }, { 4, 5, 6 } }; // { 1, 2, 3 } // // -// CHECK: %[[S_0_0:.*]] = getelementptr inbounds [3 x %[[S]]], ptr %[[START_AS_i8]], i64 0, i64 0 -// CHECK: call void @_ZN1SC1Ei(ptr {{[^,]*}} %[[S_0_0]], i32 noundef 1) -// CHECK: %[[S_0_1:.*]] = getelementptr inbounds %[[S]], ptr %[[S_0_0]], i64 1 +// CHECK: call void @_ZN1SC1Ei(ptr {{[^,]*}} %[[START_AS_i8]], i32 noundef 1) +// CHECK: %[[S_0_1:.*]] = getelementptr inbounds %[[S]], ptr %[[START_AS_i8]], i64 1 // CHECK: call void @_ZN1SC1Ei(ptr {{[^,]*}} %[[S_0_1]], i32 noundef 2) -// CHECK: %[[S_0_2:.*]] = getelementptr inbounds %[[S]], ptr %[[S_0_1]], i64 1 +// CHECK: %[[S_0_2:.*]] = getelementptr inbounds %[[S]], ptr %[[START_AS_i8]], i64 2 // CHECK: call void @_ZN1SC1Ei(ptr {{[^,]*}} %[[S_0_2]], i32 noundef 3) // // { 4, 5, 6 } // // CHECK: %[[S_1:.*]] = getelementptr inbounds [3 x %[[S]]], ptr %[[START_AS_i8]], i64 1 // -// CHECK: %[[S_1_0:.*]] = getelementptr inbounds [3 x %[[S]]], ptr %[[S_1]], i64 0, i64 0 -// CHECK: call void @_ZN1SC1Ei(ptr {{[^,]*}} %[[S_1_0]], i32 noundef 4) -// CHECK: %[[S_1_1:.*]] = getelementptr inbounds %[[S]], ptr %[[S_1_0]], i64 1 +// CHECK: call void @_ZN1SC1Ei(ptr {{[^,]*}} %[[S_1]], i32 noundef 4) +// CHECK: %[[S_1_1:.*]] = getelementptr inbounds %[[S]], ptr %[[S_1]], i64 1 // CHECK: call void @_ZN1SC1Ei(ptr {{[^,]*}} %[[S_1_1]], i32 noundef 5) -// CHECK: %[[S_1_2:.*]] = getelementptr inbounds %[[S]], ptr %[[S_1_1]], i64 1 +// CHECK: %[[S_1_2:.*]] = getelementptr inbounds %[[S]], ptr %[[S_1]], i64 2 // CHECK: call void @_ZN1SC1Ei(ptr {{[^,]*}} %[[S_1_2]], i32 noundef 6) // // And the rest. @@ -114,13 +110,12 @@ void *r = new T[n][3]{ { 1, 2, 3 }, { 4, 5, 6 } }; // { 1, 2, 3 } // // -// CHECK: %[[T_0_0:.*]] = getelementptr inbounds [3 x %[[T:.*]]], ptr %[[ALLOC]], i64 0, i64 0 -// CHECK: %[[T_0_0_0:.*]] = getelementptr inbounds %[[T]], ptr %[[T_0_0]], i32 0, i32 0 +// CHECK: %[[T_0_0_0:.*]] = getelementptr inbounds %[[T:.+]], ptr %[[ALLOC]], i32 0, i32 0 // CHECK: store i32 1, ptr %[[T_0_0_0]] -// CHECK: %[[T_0_1:.*]] = getelementptr inbounds %[[T]], ptr %[[T_0_0]], i64 1 +// CHECK: %[[T_0_1:.*]] = getelementptr inbounds %[[T]], ptr %[[ALLOC]], i64 1 // CHECK: %[[T_0_1_0:.*]] = getelementptr inbounds %[[T]], ptr %[[T_0_1]], i32 0, i32 0 // CHECK: store i32 2, ptr %[[T_0_1_0]] -// CHECK: %[[T_0_2:.*]] = getelementptr inbounds %[[T]], ptr %[[T_0_1]], i64 1 +// CHECK: %[[T_0_2:.*]] = getelementptr inbounds %[[T]], ptr %[[ALLOC]], i64 2 // CHECK: %[[T_0_2_0:.*]] = getelementptr inbounds %[[T]], ptr %[[T_0_2]], i32 0, i32 0 // CHECK: store i32 3, ptr %[[T_0_2_0]] // @@ -128,13 +123,12 @@ void *r = new T[n][3]{ { 1, 2, 3 }, { 4, 5, 6 } }; // // CHECK: %[[T_1:.*]] = getelementptr inbounds [3 x %[[T]]], ptr %[[ALLOC]], i64 1 // -// CHECK: %[[T_1_0:.*]] = getelementptr inbounds [3 x %[[T]]], ptr %[[T_1]], i64 0, i64 0 -// CHECK: %[[T_1_0_0:.*]] = getelementptr inbounds %[[T]], ptr %[[T_1_0]], i32 0, i32 0 +// CHECK: %[[T_1_0_0:.*]] = getelementptr inbounds %[[T]], ptr %[[T_1]], i32 0, i32 0 // CHECK: store i32 4, ptr %[[T_1_0_0]] -// CHECK: %[[T_1_1:.*]] = getelementptr inbounds %[[T]], ptr %[[T_1_0]], i64 1 +// CHECK: %[[T_1_1:.*]] = getelementptr inbounds %[[T]], ptr %[[T_1]], i64 1 // CHECK: %[[T_1_1_0:.*]] = getelementptr inbounds %[[T]], ptr %[[T_1_1]], i32 0, i32 0 // CHECK: store i32 5, ptr %[[T_1_1_0]] -// CHECK: %[[T_1_2:.*]] = getelementptr inbounds %[[T]], ptr %[[T_1_1]], i64 1 +// CHECK: %[[T_1_2:.*]] = getelementptr inbounds %[[T]], ptr %[[T_1]], i64 2 // CHECK: %[[T_1_2_0:.*]] = getelementptr inbounds %[[T]], ptr %[[T_1_2]], i32 0, i32 0 // CHECK: store i32 6, ptr %[[T_1_2_0]] // diff --git a/clang/test/CodeGenCXX/partial-destruction.cpp b/clang/test/CodeGenCXX/partial-destruction.cpp index 840e43880293b9f..8ceb4b9bbedeaae 100644 --- a/clang/test/CodeGenCXX/partial-destruction.cpp +++ b/clang/test/CodeGenCXX/partial-destruction.cpp @@ -20,15 +20,14 @@ namespace test0 { // CHECK-NEXT: [[SEL:%.*]] = alloca i32 // Initialize. - // CHECK-NEXT: [[E_BEGIN:%.*]] = getelementptr inbounds [10 x [[A]]], ptr [[AS]], i64 0, i64 0 - // CHECK-NEXT: store ptr [[E_BEGIN]], ptr [[ENDVAR]] - // CHECK-NEXT: invoke void @_ZN5test01AC1Ei(ptr {{[^,]*}} [[E_BEGIN]], i32 noundef 5) - // CHECK: [[E1:%.*]] = getelementptr inbounds [[A]], ptr [[E_BEGIN]], i64 1 + // CHECK-NEXT: store ptr [[AS]], ptr [[ENDVAR]] + // CHECK-NEXT: invoke void @_ZN5test01AC1Ei(ptr {{[^,]*}} [[AS]], i32 noundef 5) + // CHECK: [[E1:%.*]] = getelementptr inbounds [[A]], ptr [[AS]], i64 1 // CHECK-NEXT: store ptr [[E1]], ptr [[ENDVAR]] // CHECK-NEXT: invoke void @_ZN5test01AC1Ei(ptr {{[^,]*}} [[E1]], i32 noundef 7) - // CHECK: [[E2:%.*]] = getelementptr inbounds [[A]], ptr [[E1]], i64 1 + // CHECK: [[E2:%.*]] = getelementptr inbounds [[A]], ptr [[AS]], i64 2 // CHECK-NEXT: store ptr [[E2]], ptr [[ENDVAR]] - // CHECK-NEXT: [[E_END:%.*]] = getelementptr inbounds [[A]], ptr [[E_BEGIN]], i64 10 + // CHECK-NEXT: [[E_END:%.*]] = getelementptr inbounds [[A]], ptr [[AS]], i64 10 // CHECK-NEXT: br label // CHECK: [[E_CUR:%.*]] = phi ptr [ [[E2]], {{%.*}} ], [ [[E_NEXT:%.*]], {{%.*}} ] // CHECK-NEXT: invoke void @_ZN5test01AC1Ev(ptr {{[^,]*}} [[E_CUR]]) @@ -56,13 +55,13 @@ namespace test0 { // CHECK: landingpad { ptr, i32 } // CHECK-NEXT: cleanup // CHECK: [[PARTIAL_END:%.*]] = load ptr, ptr [[ENDVAR]] - // CHECK-NEXT: [[T0:%.*]] = icmp eq ptr [[E_BEGIN]], [[PARTIAL_END]] + // CHECK-NEXT: [[T0:%.*]] = icmp eq ptr [[AS]], [[PARTIAL_END]] // CHECK-NEXT: br i1 [[T0]], // CHECK: [[E_AFTER:%.*]] = phi ptr [ [[PARTIAL_END]], {{%.*}} ], [ [[E_CUR:%.*]], {{%.*}} ] // CHECK-NEXT: [[E_CUR]] = getelementptr inbounds [[A]], ptr [[E_AFTER]], i64 -1 // CHECKv03-NEXT: invoke void @_ZN5test01AD1Ev(ptr {{[^,]*}} [[E_CUR]]) // CHECKv11-NEXT: call void @_ZN5test01AD1Ev(ptr {{[^,]*}} [[E_CUR]]) - // CHECK: [[T0:%.*]] = icmp eq ptr [[E_CUR]], [[E_BEGIN]] + // CHECK: [[T0:%.*]] = icmp eq ptr [[E_CUR]], [[AS]] // CHECK-NEXT: br i1 [[T0]], // Primary EH destructor. @@ -189,25 +188,22 @@ namespace test4 { } // CHECK-LABEL: define{{.*}} void @_ZN5test44testEv() // CHECK: [[ARRAY:%.*]] = alloca [2 x [3 x [[A:%.*]]]], align -// CHECK: [[A0:%.*]] = getelementptr inbounds [2 x [3 x [[A]]]], ptr [[ARRAY]], i64 0, i64 0 -// CHECK-NEXT: store ptr [[A0]], -// CHECK-NEXT: [[A00:%.*]] = getelementptr inbounds [3 x [[A]]], ptr [[A0]], i64 0, i64 0 -// CHECK-NEXT: store ptr [[A00]], -// CHECK-NEXT: invoke void @_ZN5test41AC1Ej(ptr {{[^,]*}} [[A00]], i32 noundef 0) -// CHECK: [[A01:%.*]] = getelementptr inbounds [[A]], ptr [[A00]], i64 1 +// CHECK: store ptr [[ARRAY]], +// CHECK-NEXT: store ptr [[ARRAY]], +// CHECK-NEXT: invoke void @_ZN5test41AC1Ej(ptr {{[^,]*}} [[ARRAY]], i32 noundef 0) +// CHECK: [[A01:%.*]] = getelementptr inbounds [[A]], ptr [[ARRAY]], i64 1 // CHECK-NEXT: store ptr [[A01]], // CHECK-NEXT: invoke void @_ZN5test41AC1Ej(ptr {{[^,]*}} [[A01]], i32 noundef 1) -// CHECK: [[A02:%.*]] = getelementptr inbounds [[A]], ptr [[A01]], i64 1 +// CHECK: [[A02:%.*]] = getelementptr inbounds [[A]], ptr [[ARRAY]], i64 2 // CHECK-NEXT: store ptr [[A02]], // CHECK-NEXT: invoke void @_ZN5test41AC1Ej(ptr {{[^,]*}} [[A02]], i32 noundef 2) -// CHECK: [[A1:%.*]] = getelementptr inbounds [3 x [[A]]], ptr [[A0]], i64 1 +// CHECK: [[A1:%.*]] = getelementptr inbounds [3 x [[A]]], ptr [[ARRAY]], i64 1 // CHECK-NEXT: store ptr [[A1]], -// CHECK-NEXT: [[A10:%.*]] = getelementptr inbounds [3 x [[A]]], ptr [[A1]], i64 0, i64 0 -// CHECK-NEXT: store ptr [[A10]], -// CHECK-NEXT: invoke void @_ZN5test41AC1Ej(ptr {{[^,]*}} [[A10]], i32 noundef 3) -// CHECK: [[A11:%.*]] = getelementptr inbounds [[A]], ptr [[A10]], i64 1 +// CHECK-NEXT: store ptr [[A1]], +// CHECK-NEXT: invoke void @_ZN5test41AC1Ej(ptr {{[^,]*}} [[A1]], i32 noundef 3) +// CHECK: [[A11:%.*]] = getelementptr inbounds [[A]], ptr [[A1]], i64 1 // CHECK-NEXT: store ptr [[A11]], // CHECK-NEXT: invoke void @_ZN5test41AC1Ej(ptr {{[^,]*}} [[A11]], i32 noundef 4) -// CHECK: [[A12:%.*]] = getelementptr inbounds [[A]], ptr [[A11]], i64 1 +// CHECK: [[A12:%.*]] = getelementptr inbounds [[A]], ptr [[A1]], i64 2 // CHECK-NEXT: store ptr [[A12]], // CHECK-NEXT: invoke void @_ZN5test41AC1Ej(ptr {{[^,]*}} [[A12]], i32 noundef 5) diff --git a/clang/test/CodeGenCXX/temporaries.cpp b/clang/test/CodeGenCXX/temporaries.cpp index 9f29d3b73278395..186f4934873fa94 100644 --- a/clang/test/CodeGenCXX/temporaries.cpp +++ b/clang/test/CodeGenCXX/temporaries.cpp @@ -477,9 +477,8 @@ namespace Elision { // CHECK-NEXT: call void @_ZN7Elision1AC1Ev(ptr {{[^,]*}} [[X]]) A x; - // CHECK-NEXT: [[XS0:%.*]] = getelementptr inbounds [2 x [[A]]], ptr [[XS]], i64 0, i64 0 - // CHECK-NEXT: call void @_ZN7Elision1AC1Ev(ptr {{[^,]*}} [[XS0]]) - // CHECK-NEXT: [[XS1:%.*]] = getelementptr inbounds [[A]], ptr [[XS0]], i64 1 + // CHECK-NEXT: call void @_ZN7Elision1AC1Ev(ptr {{[^,]*}} [[XS]]) + // CHECK-NEXT: [[XS1:%.*]] = getelementptr inbounds [[A]], ptr [[XS]], i64 1 // CHECK-NEXT: call void @_ZN7Elision1AC1ERKS0_(ptr {{[^,]*}} [[XS1]], ptr noundef {{(nonnull )?}}align {{[0-9]+}} dereferenceable({{[0-9]+}}) [[X]]) A xs[] = { A(), x }; diff --git a/clang/test/CodeGenCXX/value-init.cpp b/clang/test/CodeGenCXX/value-init.cpp index 42181be92ace9e1..9e72769d87f696a 100644 --- a/clang/test/CodeGenCXX/value-init.cpp +++ b/clang/test/CodeGenCXX/value-init.cpp @@ -205,11 +205,9 @@ namespace test6 { // CHECK-LABEL: define{{.*}} void @_ZN5test64testEv() // CHECK: [[ARR:%.*]] = alloca [10 x [20 x [[A:%.*]]]], - // CHECK-NEXT: [[INNER:%.*]] = getelementptr inbounds [10 x [20 x [[A]]]], ptr [[ARR]], i64 0, i64 0 - // CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [20 x [[A]]], ptr [[INNER]], i64 0, i64 0 - // CHECK-NEXT: call void @_ZN5test61AC1Ei(ptr {{[^,]*}} [[T0]], i32 noundef 5) - // CHECK-NEXT: [[BEGIN:%.*]] = getelementptr inbounds [[A]], ptr [[T0]], i64 1 - // CHECK-NEXT: [[END:%.*]] = getelementptr inbounds [[A]], ptr [[T0]], i64 20 + // CHECK-NEXT: call void @_ZN5test61AC1Ei(ptr {{[^,]*}} [[ARR]], i32 noundef 5) + // CHECK-NEXT: [[BEGIN:%.*]] = getelementptr inbounds [[A]], ptr [[ARR]], i64 1 + // CHECK-NEXT: [[END:%.*]] = getelementptr inbounds [[A]], ptr [[ARR]], i64 20 // CHECK-NEXT: br label // CHECK: [[CUR:%.*]] = phi ptr [ [[BEGIN]], {{%.*}} ], [ [[NEXT:%.*]], {{%.*}} ] // CHECK-NEXT: call void @_ZN5test61AC1Ev(ptr {{[^,]*}} [[CUR]]) @@ -217,13 +215,15 @@ namespace test6 { // CHECK-NEXT: [[T0:%.*]] = icmp eq ptr [[NEXT]], [[END]] // CHECK-NEXT: br i1 - // CHECK: [[BEGIN:%.*]] = getelementptr inbounds [20 x [[A]]], ptr [[INNER]], i64 1 - // CHECK-NEXT: [[END:%.*]] = getelementptr inbounds [20 x [[A]]], ptr [[INNER]], i64 10 + // CHECK: [[BEGIN:%.*]] = getelementptr inbounds [20 x [[A]]], ptr [[ARR]], i64 1 + // CHECK-NEXT: [[END:%.*]] = getelementptr inbounds [20 x [[A]]], ptr [[ARR]], i64 10 // CHECK-NEXT: br label - // CHECK: [[CUR:%.*]] = phi ptr [ [[BEGIN]], {{%.*}} ], [ [[NEXT:%.*]], {{%.*}} ] - // Inner loop. - // CHECK-NEXT: [[IBEGIN:%.*]] = getelementptr inbounds [20 x [[A]]], ptr [[CUR]], i{{32|64}} 0, i{{32|64}} 0 + // CHECK-CXX98: [[CUR:%.*]] = phi ptr [ [[BEGIN]], {{%.*}} ], [ [[NEXT:%.*]], {{%.*}} ] + + // CHECK-CXX98: [[IBEGIN:%.*]] = getelementptr inbounds [20 x [[A]]], ptr [[CUR]], i{{32|64}} 0, i{{32|64}} 0 + // CHECK-CXX17: [[IBEGIN:%.*]] = phi ptr [ [[BEGIN]], {{%.*}} ], [ [[NEXT:%.*]], {{%.*}} ] + // CHECK-NEXT: [[IEND:%.*]] = getelementptr inbounds [[A]], ptr [[IBEGIN]], i64 20 // CHECK-NEXT: br label // CHECK: [[ICUR:%.*]] = phi ptr [ [[IBEGIN]], {{%.*}} ], [ [[INEXT:%.*]], {{%.*}} ] @@ -232,7 +232,8 @@ namespace test6 { // CHECK-NEXT: [[T0:%.*]] = icmp eq ptr [[INEXT]], [[IEND]] // CHECK-NEXT: br i1 [[T0]], - // CHECK: [[NEXT]] = getelementptr inbounds [20 x [[A]]], ptr [[CUR]], i64 1 + // CHECK-CXX98: [[NEXT]] = getelementptr inbounds [20 x [[A]]], ptr [[CUR]], i64 1 + // CHECK-CXX17: [[NEXT]] = getelementptr inbounds [20 x [[A]]], ptr [[IBEGIN]], i64 1 // CHECK-NEXT: [[T0:%.*]] = icmp eq ptr [[NEXT]], [[END]] // CHECK-NEXT: br i1 [[T0]] // CHECK: ret void diff --git a/clang/test/CodeGenCoroutines/coro-suspend-cleanups.cpp b/clang/test/CodeGenCoroutines/coro-suspend-cleanups.cpp index 06cc2069dbe9aed..d71c2c558996a7a 100644 --- a/clang/test/CodeGenCoroutines/coro-suspend-cleanups.cpp +++ b/clang/test/CodeGenCoroutines/coro-suspend-cleanups.cpp @@ -43,14 +43,11 @@ coroutine ArrayInitCoro() { // CHECK: %cleanup.isactive = alloca i1, align 1 Printy arr[2] = { Printy("a"), - // CHECK: %arrayinit.begin = getelementptr inbounds [2 x %struct.Printy], ptr %arr.reload.addr, i64 0, i64 0 - // CHECK-NEXT: %arrayinit.begin.spill.addr = getelementptr inbounds %_Z13ArrayInitCorov.Frame, ptr %0, i32 0, i32 10 - // CHECK-NEXT: store ptr %arrayinit.begin, ptr %arrayinit.begin.spill.addr, align 8 - // CHECK-NEXT: store i1 true, ptr %cleanup.isactive.reload.addr, align 1 - // CHECK-NEXT: store ptr %arrayinit.begin, ptr %arrayinit.endOfInit.reload.addr, align 8 - // CHECK-NEXT: call void @_ZN6PrintyC1EPKc(ptr noundef nonnull align 8 dereferenceable(8) %arrayinit.begin, ptr noundef @.str) - // CHECK-NEXT: %arrayinit.element = getelementptr inbounds %struct.Printy, ptr %arrayinit.begin, i64 1 - // CHECK-NEXT: %arrayinit.element.spill.addr = getelementptr inbounds %_Z13ArrayInitCorov.Frame, ptr %0, i32 0, i32 11 + // CHECK: store i1 true, ptr %cleanup.isactive.reload.addr, align 1 + // CHECK-NEXT: store ptr %arr.reload.addr, ptr %arrayinit.endOfInit.reload.addr, align 8 + // CHECK-NEXT: call void @_ZN6PrintyC1EPKc(ptr noundef nonnull align 8 dereferenceable(8) %arr.reload.addr, ptr noundef @.str) + // CHECK-NEXT: %arrayinit.element = getelementptr inbounds %struct.Printy, ptr %arr.reload.addr, i64 1 + // CHECK-NEXT: %arrayinit.element.spill.addr = getelementptr inbounds %_Z13ArrayInitCorov.Frame, ptr %0, i32 0, i32 10 // CHECK-NEXT: store ptr %arrayinit.element, ptr %arrayinit.element.spill.addr, align 8 // CHECK-NEXT: store ptr %arrayinit.element, ptr %arrayinit.endOfInit.reload.addr, align 8 co_await Awaiter{} @@ -64,7 +61,7 @@ coroutine ArrayInitCoro() { // CHECK: br label %cleanup{{.*}} // CHECK: await.ready: - // CHECK-NEXT: %arrayinit.element.reload.addr = getelementptr inbounds %_Z13ArrayInitCorov.Frame, ptr %0, i32 0, i32 11 + // CHECK-NEXT: %arrayinit.element.reload.addr = getelementptr inbounds %_Z13ArrayInitCorov.Frame, ptr %0, i32 0, i32 10 // CHECK-NEXT: %arrayinit.element.reload = load ptr, ptr %arrayinit.element.reload.addr, align 8 // CHECK-NEXT: call void @_ZN7Awaiter12await_resumeEv // CHECK-NEXT: store i1 false, ptr %cleanup.isactive.reload.addr, align 1 @@ -75,7 +72,7 @@ coroutine ArrayInitCoro() { // CHECK-NEXT: br i1 %cleanup.is_active, label %cleanup.action, label %cleanup.done // CHECK: cleanup.action: - // CHECK: %arraydestroy.isempty = icmp eq ptr %arrayinit.begin.reload{{.*}}, %{{.*}} + // CHECK: %arraydestroy.isempty = icmp eq ptr %arr.reload.addr, %{{.*}} // CHECK-NEXT: br i1 %arraydestroy.isempty, label %arraydestroy.done{{.*}}, label %arraydestroy.body.from.cleanup.action // Ignore rest of the array cleanup. } diff --git a/clang/test/CodeGenObjC/arc-ternary-op.m b/clang/test/CodeGenObjC/arc-ternary-op.m index 87167d911ea8ef1..d633851ac76174e 100644 --- a/clang/test/CodeGenObjC/arc-ternary-op.m +++ b/clang/test/CodeGenObjC/arc-ternary-op.m @@ -153,22 +153,20 @@ void test3(int cond) { // CHECK: %[[_COMPOUNDLITERAL1:.*]] = alloca [2 x ptr], align 8 // CHECK: %[[CLEANUP_COND4:.*]] = alloca i1, align 1 - // CHECK: %[[ARRAYINIT_BEGIN:.*]] = getelementptr inbounds [2 x ptr], ptr %[[_COMPOUNDLITERAL]], i64 0, i64 0 // CHECK: %[[V2:.*]] = load ptr, ptr @g0, align 8 // CHECK: %[[V3:.*]] = call ptr @llvm.objc.retain(ptr %[[V2]]) - // CHECK: store ptr %[[V3]], ptr %[[ARRAYINIT_BEGIN]], align 8 - // CHECK: %[[ARRAYINIT_ELEMENT:.*]] = getelementptr inbounds ptr, ptr %[[ARRAYINIT_BEGIN]], i64 1 + // CHECK: store ptr %[[V3]], ptr %[[_COMPOUNDLITERAL]], align 8 + // CHECK: %[[ARRAYINIT_ELEMENT:.*]] = getelementptr inbounds ptr, ptr %[[_COMPOUNDLITERAL]], i64 1 // CHECK: %[[V4:.*]] = load ptr, ptr @g1, align 8 // CHECK: %[[V5:.*]] = call ptr @llvm.objc.retain(ptr %[[V4]]) // CHECK: store ptr %[[V5]], ptr %[[ARRAYINIT_ELEMENT]], align 8 // CHECK: store i1 true, ptr %[[CLEANUP_COND]], align 1 // CHECK: %[[ARRAYDECAY:.*]] = getelementptr inbounds [2 x ptr], ptr %[[_COMPOUNDLITERAL]], i64 0, i64 0 - // CHECK: %[[ARRAYINIT_BEGIN2:.*]] = getelementptr inbounds [2 x ptr], ptr %[[_COMPOUNDLITERAL1]], i64 0, i64 0 // CHECK: %[[V6:.*]] = load ptr, ptr @g1, align 8 // CHECK: %[[V7:.*]] = call ptr @llvm.objc.retain(ptr %[[V6]]) - // CHECK: store ptr %[[V7]], ptr %[[ARRAYINIT_BEGIN2]], align 8 - // CHECK: %[[ARRAYINIT_ELEMENT3:.*]] = getelementptr inbounds ptr, ptr %[[ARRAYINIT_BEGIN2]], i64 1 + // CHECK: store ptr %[[V7]], ptr %[[_COMPOUNDLITERAL1]], align 8 + // CHECK: %[[ARRAYINIT_ELEMENT3:.*]] = getelementptr inbounds ptr, ptr %[[_COMPOUNDLITERAL1]], i64 1 // CHECK: %[[V8:.*]] = load ptr, ptr @g0, align 8 // CHECK: %[[V9:.*]] = call ptr @llvm.objc.retain(ptr %[[V8]]) // CHECK: store ptr %[[V9]], ptr %[[ARRAYINIT_ELEMENT3]], align 8 diff --git a/clang/test/CodeGenObjC/arc.m b/clang/test/CodeGenObjC/arc.m index aeead58f8131d70..48ca14af3b2abd9 100644 --- a/clang/test/CodeGenObjC/arc.m +++ b/clang/test/CodeGenObjC/arc.m @@ -1377,11 +1377,10 @@ void test71(void) { // CHECK: %[[T:.*]] = alloca [2 x ptr], align 16 // CHECK: %[[V0:.*]] = call ptr @llvm.objc.retain(ptr %[[A]]) // CHECK: %[[V1:.*]] = call ptr @llvm.objc.retain(ptr %[[B]]) #2 -// CHECK: %[[ARRAYINIT_BEGIN:.*]] = getelementptr inbounds [2 x ptr], ptr %[[T]], i64 0, i64 0 // CHECK: %[[V3:.*]] = load ptr, ptr %[[A_ADDR]], align 8, !tbaa !7 // CHECK: %[[V4:.*]] = call ptr @llvm.objc.retain(ptr %[[V3]]) #2 -// CHECK: store ptr %[[V4]], ptr %[[ARRAYINIT_BEGIN]], align 8, !tbaa !7 -// CHECK: %[[ARRAYINIT_ELEMENT:.*]] = getelementptr inbounds ptr, ptr %[[ARRAYINIT_BEGIN]], i64 1 +// CHECK: store ptr %[[V4]], ptr %[[T]], align 8, !tbaa !7 +// CHECK: %[[ARRAYINIT_ELEMENT:.*]] = getelementptr inbounds ptr, ptr %[[T]], i64 1 // CHECK: %[[V5:.*]] = load ptr, ptr %[[B_ADDR]], align 8, !tbaa !7 // CHECK: %[[V6:.*]] = call ptr @llvm.objc.retain(ptr %[[V5]]) #2 // CHECK: store ptr %[[V6]], ptr %[[ARRAYINIT_ELEMENT]], align 8, !tbaa !7 diff --git a/clang/test/CodeGenObjCXX/arc-exceptions.mm b/clang/test/CodeGenObjCXX/arc-exceptions.mm index 709afa32ac7d81f..3efe566a6a8b51a 100644 --- a/clang/test/CodeGenObjCXX/arc-exceptions.mm +++ b/clang/test/CodeGenObjCXX/arc-exceptions.mm @@ -115,23 +115,20 @@ void test5(void) { } // CHECK-LABEL: define{{.*}} void @_Z5test5v() // CHECK: [[ARRAY:%.*]] = alloca [2 x [2 x ptr]], align -// CHECK: [[A0:%.*]] = getelementptr inbounds [2 x [2 x ptr]], ptr [[ARRAY]], i64 0, i64 0 -// CHECK-NEXT: store ptr [[A0]], -// CHECK-NEXT: [[A00:%.*]] = getelementptr inbounds [2 x ptr], ptr [[A0]], i64 0, i64 0 -// CHECK-NEXT: store ptr [[A00]], +// CHECK: store ptr [[ARRAY]], +// CHECK-NEXT: store ptr [[ARRAY]], // CHECK-NEXT: [[T0:%.*]] = invoke noundef ptr @_Z12test5_helperj(i32 noundef 0) -// CHECK: store ptr [[T0]], ptr [[A00]], align -// CHECK-NEXT: [[A01:%.*]] = getelementptr inbounds ptr, ptr [[A00]], i64 1 +// CHECK: store ptr [[T0]], ptr [[ARRAY]], align +// CHECK-NEXT: [[A01:%.*]] = getelementptr inbounds ptr, ptr [[ARRAY]], i64 1 // CHECK-NEXT: store ptr [[A01]], // CHECK-NEXT: [[T0:%.*]] = invoke noundef ptr @_Z12test5_helperj(i32 noundef 1) // CHECK: store ptr [[T0]], ptr [[A01]], align -// CHECK-NEXT: [[A1:%.*]] = getelementptr inbounds [2 x ptr], ptr [[A0]], i64 1 +// CHECK-NEXT: [[A1:%.*]] = getelementptr inbounds [2 x ptr], ptr [[ARRAY]], i64 1 +// CHECK-NEXT: store ptr [[A1]], // CHECK-NEXT: store ptr [[A1]], -// CHECK-NEXT: [[A10:%.*]] = getelementptr inbounds [2 x ptr], ptr [[A1]], i64 0, i64 0 -// CHECK-NEXT: store ptr [[A10]], // CHECK-NEXT: [[T0:%.*]] = invoke noundef ptr @_Z12test5_helperj(i32 noundef 2) -// CHECK: store ptr [[T0]], ptr [[A10]], align -// CHECK-NEXT: [[A11:%.*]] = getelementptr inbounds ptr, ptr [[A10]], i64 1 +// CHECK: store ptr [[T0]], ptr [[A1]], align +// CHECK-NEXT: [[A11:%.*]] = getelementptr inbounds ptr, ptr [[A1]], i64 1 // CHECK-NEXT: store ptr [[A11]], // CHECK-NEXT: [[T0:%.*]] = invoke noundef ptr @_Z12test5_helperj(i32 noundef 3) // CHECK: store ptr [[T0]], ptr [[A11]], align diff --git a/clang/test/OpenMP/distribute_firstprivate_codegen.cpp b/clang/test/OpenMP/distribute_firstprivate_codegen.cpp index 567d12119db2187..6a15cae30762315 100644 --- a/clang/test/OpenMP/distribute_firstprivate_codegen.cpp +++ b/clang/test/OpenMP/distribute_firstprivate_codegen.cpp @@ -493,9 +493,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -798,9 +797,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1158,9 +1156,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -1461,9 +1458,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 diff --git a/clang/test/OpenMP/distribute_lastprivate_codegen.cpp b/clang/test/OpenMP/distribute_lastprivate_codegen.cpp index 5d1a12d27145e03..510c5d4d88efc24 100644 --- a/clang/test/OpenMP/distribute_lastprivate_codegen.cpp +++ b/clang/test/OpenMP/distribute_lastprivate_codegen.cpp @@ -478,9 +478,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -801,9 +800,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1180,9 +1178,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -1501,9 +1498,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 diff --git a/clang/test/OpenMP/distribute_parallel_for_firstprivate_codegen.cpp b/clang/test/OpenMP/distribute_parallel_for_firstprivate_codegen.cpp index faae599bb32afe7..11b18e871551eb6 100644 --- a/clang/test/OpenMP/distribute_parallel_for_firstprivate_codegen.cpp +++ b/clang/test/OpenMP/distribute_parallel_for_firstprivate_codegen.cpp @@ -776,9 +776,8 @@ int main() { // CHECK8-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK8-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK8-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK8-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK8-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK8-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK8-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK8-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK8-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK8-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK8-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1211,9 +1210,8 @@ int main() { // CHECK8-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK8-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK8-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK8-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK8-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK8-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK8-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK8-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK8-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK8-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK8-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1695,9 +1693,8 @@ int main() { // CHECK10-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK10-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK10-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK10-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK10-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK10-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK10-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK10-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK10-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK10-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK10-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -2124,9 +2121,8 @@ int main() { // CHECK10-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK10-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK10-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK10-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK10-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK10-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK10-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK10-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK10-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK10-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK10-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 diff --git a/clang/test/OpenMP/distribute_parallel_for_lastprivate_codegen.cpp b/clang/test/OpenMP/distribute_parallel_for_lastprivate_codegen.cpp index 18a86c5e5666399..29e1a35e90e0251 100644 --- a/clang/test/OpenMP/distribute_parallel_for_lastprivate_codegen.cpp +++ b/clang/test/OpenMP/distribute_parallel_for_lastprivate_codegen.cpp @@ -752,9 +752,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1223,9 +1222,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1744,9 +1742,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -2209,9 +2206,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 diff --git a/clang/test/OpenMP/distribute_parallel_for_private_codegen.cpp b/clang/test/OpenMP/distribute_parallel_for_private_codegen.cpp index 0969a0ca21e7ab5..03c62ce699b3031 100644 --- a/clang/test/OpenMP/distribute_parallel_for_private_codegen.cpp +++ b/clang/test/OpenMP/distribute_parallel_for_private_codegen.cpp @@ -514,9 +514,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: store ptr undef, ptr [[_TMP1]], align 8 @@ -835,9 +834,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: store ptr undef, ptr [[_TMP1]], align 8 @@ -1225,9 +1223,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: store ptr undef, ptr [[_TMP1]], align 4 @@ -1540,9 +1537,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: store ptr undef, ptr [[_TMP1]], align 4 diff --git a/clang/test/OpenMP/distribute_parallel_for_simd_firstprivate_codegen.cpp b/clang/test/OpenMP/distribute_parallel_for_simd_firstprivate_codegen.cpp index 545ea9f0f43fdf9..cb2cbb039a4e016 100644 --- a/clang/test/OpenMP/distribute_parallel_for_simd_firstprivate_codegen.cpp +++ b/clang/test/OpenMP/distribute_parallel_for_simd_firstprivate_codegen.cpp @@ -839,9 +839,8 @@ int main() { // CHECK8-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK8-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK8-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK8-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK8-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK8-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK8-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK8-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK8-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK8-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK8-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1288,9 +1287,8 @@ int main() { // CHECK8-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK8-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK8-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK8-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK8-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK8-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK8-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK8-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK8-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK8-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK8-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1786,9 +1784,8 @@ int main() { // CHECK10-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK10-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK10-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK10-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK10-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK10-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK10-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK10-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK10-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK10-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK10-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -2229,9 +2226,8 @@ int main() { // CHECK10-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK10-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK10-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK10-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK10-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK10-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK10-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK10-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK10-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK10-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK10-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -2721,9 +2717,8 @@ int main() { // CHECK12-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK12-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK12-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK12-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK12-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK12-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK12-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK12-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK12-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK12-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK12-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -2828,9 +2823,8 @@ int main() { // CHECK12-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK12-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK12-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK12-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK12-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK12-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK12-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK12-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK12-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK12-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK12-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -3026,9 +3020,8 @@ int main() { // CHECK14-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK14-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK14-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK14-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK14-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK14-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK14-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK14-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK14-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK14-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK14-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -3131,9 +3124,8 @@ int main() { // CHECK14-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK14-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK14-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK14-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK14-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK14-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK14-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK14-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK14-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK14-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK14-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 diff --git a/clang/test/OpenMP/distribute_parallel_for_simd_lastprivate_codegen.cpp b/clang/test/OpenMP/distribute_parallel_for_simd_lastprivate_codegen.cpp index 675f1313460e659..c8c230b40c50ca9 100644 --- a/clang/test/OpenMP/distribute_parallel_for_simd_lastprivate_codegen.cpp +++ b/clang/test/OpenMP/distribute_parallel_for_simd_lastprivate_codegen.cpp @@ -819,9 +819,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1304,9 +1303,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1839,9 +1837,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -2318,9 +2315,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -2853,9 +2849,8 @@ int main() { // CHECK13-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK13-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK13-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK13-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK13-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK13-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK13-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK13-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK13-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -3021,9 +3016,8 @@ int main() { // CHECK13-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK13-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK13-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK13-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK13-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK13-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK13-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK13-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK13-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -3260,9 +3254,8 @@ int main() { // CHECK15-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK15-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK15-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK15-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK15-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK15-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK15-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK15-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK15-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -3426,9 +3419,8 @@ int main() { // CHECK15-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK15-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK15-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK15-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK15-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK15-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK15-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK15-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK15-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 diff --git a/clang/test/OpenMP/distribute_parallel_for_simd_private_codegen.cpp b/clang/test/OpenMP/distribute_parallel_for_simd_private_codegen.cpp index a74af2d7c393204..f884babe7e2d90d 100644 --- a/clang/test/OpenMP/distribute_parallel_for_simd_private_codegen.cpp +++ b/clang/test/OpenMP/distribute_parallel_for_simd_private_codegen.cpp @@ -568,9 +568,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: store ptr undef, ptr [[_TMP1]], align 8 @@ -903,9 +902,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: store ptr undef, ptr [[_TMP1]], align 8 @@ -1307,9 +1305,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: store ptr undef, ptr [[_TMP1]], align 4 @@ -1636,9 +1633,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: store ptr undef, ptr [[_TMP1]], align 4 @@ -2043,9 +2039,8 @@ int main() { // CHECK13-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK13-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK13-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK13-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK13-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK13-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK13-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK13-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK13-NEXT: store ptr undef, ptr [[_TMP1]], align 8 @@ -2182,9 +2177,8 @@ int main() { // CHECK13-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK13-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK13-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK13-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK13-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK13-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK13-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK13-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK13-NEXT: store ptr undef, ptr [[_TMP1]], align 8 @@ -2393,9 +2387,8 @@ int main() { // CHECK15-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK15-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK15-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK15-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK15-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK15-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK15-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK15-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK15-NEXT: store ptr undef, ptr [[_TMP1]], align 4 @@ -2530,9 +2523,8 @@ int main() { // CHECK15-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK15-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK15-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK15-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK15-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK15-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK15-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK15-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK15-NEXT: store ptr undef, ptr [[_TMP1]], align 4 diff --git a/clang/test/OpenMP/distribute_private_codegen.cpp b/clang/test/OpenMP/distribute_private_codegen.cpp index b6e796fb8027c05..7544d429934241c 100644 --- a/clang/test/OpenMP/distribute_private_codegen.cpp +++ b/clang/test/OpenMP/distribute_private_codegen.cpp @@ -344,9 +344,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: store ptr undef, ptr [[_TMP1]], align 8 @@ -661,9 +660,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: store ptr undef, ptr [[_TMP1]], align 8 @@ -950,9 +948,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: store ptr undef, ptr [[_TMP1]], align 4 @@ -1265,9 +1262,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: store ptr undef, ptr [[_TMP1]], align 4 diff --git a/clang/test/OpenMP/distribute_simd_firstprivate_codegen.cpp b/clang/test/OpenMP/distribute_simd_firstprivate_codegen.cpp index 1b5950ce11d7fa2..35284a50c386667 100644 --- a/clang/test/OpenMP/distribute_simd_firstprivate_codegen.cpp +++ b/clang/test/OpenMP/distribute_simd_firstprivate_codegen.cpp @@ -546,9 +546,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -858,9 +857,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1225,9 +1223,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -1535,9 +1532,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -1900,9 +1896,8 @@ int main() { // CHECK13-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK13-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK13-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK13-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK13-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK13-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK13-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK13-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK13-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -2007,9 +2002,8 @@ int main() { // CHECK13-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK13-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK13-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK13-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK13-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK13-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK13-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK13-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK13-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -2205,9 +2199,8 @@ int main() { // CHECK15-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK15-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK15-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK15-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK15-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK15-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK15-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK15-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK15-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -2310,9 +2303,8 @@ int main() { // CHECK15-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK15-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK15-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK15-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK15-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK15-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK15-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK15-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK15-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 diff --git a/clang/test/OpenMP/distribute_simd_lastprivate_codegen.cpp b/clang/test/OpenMP/distribute_simd_lastprivate_codegen.cpp index 92b73b220bcc218..33c488eedf3a9c5 100644 --- a/clang/test/OpenMP/distribute_simd_lastprivate_codegen.cpp +++ b/clang/test/OpenMP/distribute_simd_lastprivate_codegen.cpp @@ -533,9 +533,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -863,9 +862,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1249,9 +1247,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -1577,9 +1574,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -1967,9 +1963,8 @@ int main() { // CHECK13-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK13-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK13-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK13-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK13-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK13-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK13-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK13-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK13-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -2135,9 +2130,8 @@ int main() { // CHECK13-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK13-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK13-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK13-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK13-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK13-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK13-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK13-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK13-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -2374,9 +2368,8 @@ int main() { // CHECK15-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK15-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK15-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK15-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK15-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK15-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK15-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK15-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK15-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -2540,9 +2533,8 @@ int main() { // CHECK15-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK15-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK15-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK15-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK15-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK15-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK15-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK15-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK15-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 diff --git a/clang/test/OpenMP/distribute_simd_private_codegen.cpp b/clang/test/OpenMP/distribute_simd_private_codegen.cpp index 93b2bd8b12a8ee3..534903e0afbe77e 100644 --- a/clang/test/OpenMP/distribute_simd_private_codegen.cpp +++ b/clang/test/OpenMP/distribute_simd_private_codegen.cpp @@ -389,9 +389,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: store ptr undef, ptr [[_TMP1]], align 8 @@ -735,9 +734,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: store ptr undef, ptr [[_TMP1]], align 8 @@ -1035,9 +1033,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: store ptr undef, ptr [[_TMP1]], align 4 @@ -1379,9 +1376,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: store ptr undef, ptr [[_TMP1]], align 4 @@ -1685,9 +1681,8 @@ int main() { // CHECK13-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK13-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK13-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK13-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK13-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK13-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK13-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK13-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK13-NEXT: store ptr undef, ptr [[_TMP1]], align 8 @@ -1849,9 +1844,8 @@ int main() { // CHECK13-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK13-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK13-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK13-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK13-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK13-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK13-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK13-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK13-NEXT: store ptr undef, ptr [[_TMP1]], align 8 @@ -2066,9 +2060,8 @@ int main() { // CHECK15-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK15-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK15-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK15-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK15-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK15-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK15-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK15-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK15-NEXT: store ptr undef, ptr [[_TMP1]], align 4 @@ -2228,9 +2221,8 @@ int main() { // CHECK15-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK15-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK15-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK15-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK15-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK15-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK15-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK15-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK15-NEXT: store ptr undef, ptr [[_TMP1]], align 4 diff --git a/clang/test/OpenMP/for_firstprivate_codegen.cpp b/clang/test/OpenMP/for_firstprivate_codegen.cpp index 79f76bd0f4482a1..6f51b82668be3ba 100644 --- a/clang/test/OpenMP/for_firstprivate_codegen.cpp +++ b/clang/test/OpenMP/for_firstprivate_codegen.cpp @@ -423,9 +423,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK1-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 diff --git a/clang/test/OpenMP/for_lastprivate_codegen.cpp b/clang/test/OpenMP/for_lastprivate_codegen.cpp index c7ef60afcfc0329..f89969d85a32047 100644 --- a/clang/test/OpenMP/for_lastprivate_codegen.cpp +++ b/clang/test/OpenMP/for_lastprivate_codegen.cpp @@ -374,9 +374,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[VAR]], float noundef 3.000000e+00) // CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB3:[0-9]+]], i32 5, ptr @main.omp_outlined, ptr [[T_VAR]], ptr [[VEC]], ptr [[S_ARR]], ptr [[VAR]], ptr @_ZZ4mainE5sivar) @@ -848,9 +847,8 @@ int main() { // CHECK1-NEXT: call void @_ZN3SSTIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[SST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 128 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 128 [[VEC]], ptr align 128 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK1-NEXT: store ptr [[TEST]], ptr [[VAR]], align 128 // CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 128 @@ -2798,9 +2796,8 @@ int main() { // CHECK5-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK5-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK5-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK5-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK5-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK5-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK5-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK5-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK5-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK5-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[VAR]], float noundef 3.000000e+00) // CHECK5-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB3:[0-9]+]], i32 5, ptr @main.omp_outlined, ptr [[T_VAR]], ptr [[VEC]], ptr [[S_ARR]], ptr [[VAR]], ptr @_ZZ4mainE5sivar) @@ -3290,9 +3287,8 @@ int main() { // CHECK5-NEXT: call void @_ZN3SSTIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[SST]]) // CHECK5-NEXT: store i32 0, ptr [[T_VAR]], align 128 // CHECK5-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 128 [[VEC]], ptr align 128 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK5-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK5-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK5-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK5-NEXT: store ptr [[TEST]], ptr [[VAR]], align 128 // CHECK5-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 128 diff --git a/clang/test/OpenMP/for_private_codegen.cpp b/clang/test/OpenMP/for_private_codegen.cpp index 463e5305e9fb667..d140f0f51cdf1e8 100644 --- a/clang/test/OpenMP/for_private_codegen.cpp +++ b/clang/test/OpenMP/for_private_codegen.cpp @@ -121,9 +121,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK1-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB3:[0-9]+]], i32 0, ptr @main.omp_outlined) @@ -360,9 +359,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK1-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB3]], i32 0, ptr @_Z5tmainIiET_v.omp_outlined) diff --git a/clang/test/OpenMP/for_reduction_codegen.cpp b/clang/test/OpenMP/for_reduction_codegen.cpp index d025c6d6e571f88..498845d1f4d15c1 100644 --- a/clang/test/OpenMP/for_reduction_codegen.cpp +++ b/clang/test/OpenMP/for_reduction_codegen.cpp @@ -556,13 +556,12 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store float 0.000000e+00, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [4 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT1:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_ELEMENT]], i64 1 +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT1:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 2 // CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT1]], float noundef 3.000000e+00) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT2:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_ELEMENT1]], i64 1 +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT2:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 3 // CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT2]], float noundef 4.000000e+00) // CHECK1-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK1-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[VAR1]]) @@ -3377,9 +3376,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiLi42EET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK1-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[VAR1]]) diff --git a/clang/test/OpenMP/for_reduction_codegen_UDR.cpp b/clang/test/OpenMP/for_reduction_codegen_UDR.cpp index 999e062bbe264d2..f0ab531e1d9924c 100644 --- a/clang/test/OpenMP/for_reduction_codegen_UDR.cpp +++ b/clang/test/OpenMP/for_reduction_codegen_UDR.cpp @@ -622,13 +622,12 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(12) [[TEST]]) // CHECK1-NEXT: store float 0.000000e+00, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [4 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(12) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(12) [[S_ARR]], float noundef 1.000000e+00) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(12) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT1:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_ELEMENT]], i64 1 +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT1:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 2 // CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(12) [[ARRAYINIT_ELEMENT1]], float noundef 3.000000e+00) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT2:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_ELEMENT1]], i64 1 +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT2:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 3 // CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(12) [[ARRAYINIT_ELEMENT2]], float noundef 4.000000e+00) // CHECK1-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK1-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(12) [[VAR1]]) @@ -2421,9 +2420,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(12) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiLi42EET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(12) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(12) [[S_ARR]], i32 noundef 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(12) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK1-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(12) [[VAR1]]) @@ -3318,13 +3316,12 @@ int main() { // CHECK3-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(12) [[TEST]]) // CHECK3-NEXT: store float 0.000000e+00, ptr [[T_VAR]], align 4 // CHECK3-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK3-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [4 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK3-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(12) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK3-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(12) [[S_ARR]], float noundef 1.000000e+00) +// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK3-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(12) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) -// CHECK3-NEXT: [[ARRAYINIT_ELEMENT1:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_ELEMENT]], i64 1 +// CHECK3-NEXT: [[ARRAYINIT_ELEMENT1:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 2 // CHECK3-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(12) [[ARRAYINIT_ELEMENT1]], float noundef 3.000000e+00) -// CHECK3-NEXT: [[ARRAYINIT_ELEMENT2:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_ELEMENT1]], i64 1 +// CHECK3-NEXT: [[ARRAYINIT_ELEMENT2:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 3 // CHECK3-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(12) [[ARRAYINIT_ELEMENT2]], float noundef 4.000000e+00) // CHECK3-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK3-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(12) [[VAR1]]) @@ -3584,9 +3581,8 @@ int main() { // CHECK3-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(12) [[TEST]]) // CHECK3-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK3-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiLi42EET_v.vec, i64 8, i1 false) -// CHECK3-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(12) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(12) [[S_ARR]], i32 noundef 1) +// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(12) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK3-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK3-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(12) [[VAR1]]) diff --git a/clang/test/OpenMP/parallel_copyin_codegen.cpp b/clang/test/OpenMP/parallel_copyin_codegen.cpp index 57a563d49b31e89..e653a7734161b72 100644 --- a/clang/test/OpenMP/parallel_copyin_codegen.cpp +++ b/clang/test/OpenMP/parallel_copyin_codegen.cpp @@ -297,9 +297,8 @@ void foo() { // CHECK1-NEXT: [[DOTADDR:%.*]] = alloca ptr, align 8 // CHECK1-NEXT: store ptr [[TMP0]], ptr [[DOTADDR]], align 8 // CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[DOTADDR]], align 8 -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[TMP1]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S:%.*]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[TMP1]], float noundef 1.000000e+00) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S:%.*]], ptr [[TMP1]], i64 1 // CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[DOTADDR]], align 8 // CHECK1-NEXT: ret ptr [[TMP2]] @@ -574,9 +573,8 @@ void foo() { // CHECK1-NEXT: [[DOTADDR:%.*]] = alloca ptr, align 8 // CHECK1-NEXT: store ptr [[TMP0]], ptr [[DOTADDR]], align 8 // CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[DOTADDR]], align 8 -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[TMP1]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0:%.*]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[TMP1]], i32 noundef 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0:%.*]], ptr [[TMP1]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[DOTADDR]], align 8 // CHECK1-NEXT: ret ptr [[TMP2]] diff --git a/clang/test/OpenMP/parallel_firstprivate_codegen.cpp b/clang/test/OpenMP/parallel_firstprivate_codegen.cpp index e5374c384d2ab24..1cfba8c3159584e 100644 --- a/clang/test/OpenMP/parallel_firstprivate_codegen.cpp +++ b/clang/test/OpenMP/parallel_firstprivate_codegen.cpp @@ -258,9 +258,8 @@ void array_func(float a[3], St s[2], int n, long double vla1[n]) { // CHECK1-NEXT: call void @_ZN1SIfEC1Ev(ptr nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float 1.000000e+00) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr nonnull align 4 dereferenceable(4) [[S_ARR]], float 1.000000e+00) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float 2.000000e+00) // CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr nonnull align 4 dereferenceable(4) [[VAR]], float 3.000000e+00) // CHECK1-NEXT: [[TMP0:%.*]] = load i32, ptr [[T_VAR]], align 4 @@ -473,9 +472,8 @@ void array_func(float a[3], St s[2], int n, long double vla1[n]) { // CHECK1-NEXT: call void @_ZN3SSTIiEC1Ev(ptr nonnull align 4 dereferenceable(4) [[SST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 128 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 128 [[VEC]], ptr align 128 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr nonnull align 4 dereferenceable(4) [[S_ARR]], i32 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 2) // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr nonnull align 4 dereferenceable(4) [[VAR]], i32 3) // CHECK1-NEXT: [[TMP0:%.*]] = load i32, ptr [[T_VAR]], align 128 @@ -1440,9 +1438,8 @@ void array_func(float a[3], St s[2], int n, long double vla1[n]) { // CHECK9-NEXT: call void @_ZN1SIfEC1Ev(ptr nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float 1.000000e+00) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr nonnull align 4 dereferenceable(4) [[S_ARR]], float 1.000000e+00) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float 2.000000e+00) // CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr nonnull align 4 dereferenceable(4) [[VAR]], float 3.000000e+00) // CHECK9-NEXT: [[TMP0:%.*]] = load i32, ptr [[T_VAR]], align 4 @@ -1655,9 +1652,8 @@ void array_func(float a[3], St s[2], int n, long double vla1[n]) { // CHECK9-NEXT: call void @_ZN3SSTIiEC1Ev(ptr nonnull align 4 dereferenceable(4) [[SST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 128 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 128 [[VEC]], ptr align 128 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 1) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr nonnull align 4 dereferenceable(4) [[S_ARR]], i32 1) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 2) // CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr nonnull align 4 dereferenceable(4) [[VAR]], i32 3) // CHECK9-NEXT: [[TMP0:%.*]] = load i32, ptr [[T_VAR]], align 128 diff --git a/clang/test/OpenMP/parallel_master_taskloop_firstprivate_codegen.cpp b/clang/test/OpenMP/parallel_master_taskloop_firstprivate_codegen.cpp index fa631854ccae887..bfb31c3003c3beb 100644 --- a/clang/test/OpenMP/parallel_master_taskloop_firstprivate_codegen.cpp +++ b/clang/test/OpenMP/parallel_master_taskloop_firstprivate_codegen.cpp @@ -206,9 +206,8 @@ void array_func(int n, float a[n], St s[2]) { // CHECK-NEXT: call void @_ZN1SIdEC1ERKS0_d(ptr noundef nonnull align 8 dereferenceable(8) [[TEST]], ptr noundef nonnull align 8 dereferenceable(8) [[TTT]], double noundef 0.000000e+00) // CHECK-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[ARRAYINIT_BEGIN]], double noundef 1.000000e+00) -// CHECK-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[S_ARR]], double noundef 1.000000e+00) +// CHECK-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[ARRAYINIT_ELEMENT]], double noundef 2.000000e+00) // CHECK-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[VAR]], double noundef 3.000000e+00) // CHECK-NEXT: [[TMP0:%.*]] = load i32, ptr [[T_VAR]], align 4 @@ -569,9 +568,8 @@ void array_func(int n, float a[n], St s[2]) { // CHECK-NEXT: call void @_ZN1SIiEC1ERKS0_i(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]], ptr noundef nonnull align 4 dereferenceable(4) [[TTT]], i32 noundef 0) // CHECK-NEXT: store i32 0, ptr [[T_VAR]], align 128 // CHECK-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[VAR]], i32 noundef 3) // CHECK-NEXT: [[TMP0:%.*]] = load i32, ptr [[T_VAR]], align 128 @@ -1519,9 +1517,8 @@ void array_func(int n, float a[n], St s[2]) { // SIMD-ONLY0-NEXT: call void @_ZN1SIdEC1ERKS0_d(ptr noundef nonnull align 8 dereferenceable(8) [[TEST]], ptr noundef nonnull align 8 dereferenceable(8) [[TTT]], double noundef 0.000000e+00) // SIMD-ONLY0-NEXT: store i32 0, ptr [[T_VAR]], align 4 // SIMD-ONLY0-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// SIMD-ONLY0-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// SIMD-ONLY0-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[ARRAYINIT_BEGIN]], double noundef 1.000000e+00) -// SIMD-ONLY0-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// SIMD-ONLY0-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[S_ARR]], double noundef 1.000000e+00) +// SIMD-ONLY0-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // SIMD-ONLY0-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[ARRAYINIT_ELEMENT]], double noundef 2.000000e+00) // SIMD-ONLY0-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[VAR]], double noundef 3.000000e+00) // SIMD-ONLY0-NEXT: store i32 0, ptr [[I]], align 4 @@ -1617,9 +1614,8 @@ void array_func(int n, float a[n], St s[2]) { // SIMD-ONLY0-NEXT: call void @_ZN1SIiEC1ERKS0_i(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]], ptr noundef nonnull align 4 dereferenceable(4) [[TTT]], i32 noundef 0) // SIMD-ONLY0-NEXT: store i32 0, ptr [[T_VAR]], align 128 // SIMD-ONLY0-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// SIMD-ONLY0-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// SIMD-ONLY0-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// SIMD-ONLY0-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// SIMD-ONLY0-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// SIMD-ONLY0-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // SIMD-ONLY0-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // SIMD-ONLY0-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[VAR]], i32 noundef 3) // SIMD-ONLY0-NEXT: store i32 0, ptr [[I]], align 4 @@ -1842,9 +1838,8 @@ void array_func(int n, float a[n], St s[2]) { // SIMD-ONLY1-NEXT: call void @_ZN1SIdEC1ERKS0_d(ptr noundef nonnull align 8 dereferenceable(8) [[TEST]], ptr noundef nonnull align 8 dereferenceable(8) [[TTT]], double noundef 0.000000e+00) // SIMD-ONLY1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // SIMD-ONLY1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// SIMD-ONLY1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// SIMD-ONLY1-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[ARRAYINIT_BEGIN]], double noundef 1.000000e+00) -// SIMD-ONLY1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// SIMD-ONLY1-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[S_ARR]], double noundef 1.000000e+00) +// SIMD-ONLY1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // SIMD-ONLY1-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[ARRAYINIT_ELEMENT]], double noundef 2.000000e+00) // SIMD-ONLY1-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[VAR]], double noundef 3.000000e+00) // SIMD-ONLY1-NEXT: store i32 0, ptr [[I]], align 4 @@ -1940,9 +1935,8 @@ void array_func(int n, float a[n], St s[2]) { // SIMD-ONLY1-NEXT: call void @_ZN1SIiEC1ERKS0_i(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]], ptr noundef nonnull align 4 dereferenceable(4) [[TTT]], i32 noundef 0) // SIMD-ONLY1-NEXT: store i32 0, ptr [[T_VAR]], align 128 // SIMD-ONLY1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// SIMD-ONLY1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// SIMD-ONLY1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// SIMD-ONLY1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// SIMD-ONLY1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// SIMD-ONLY1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // SIMD-ONLY1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // SIMD-ONLY1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[VAR]], i32 noundef 3) // SIMD-ONLY1-NEXT: store i32 0, ptr [[I]], align 4 diff --git a/clang/test/OpenMP/parallel_master_taskloop_lastprivate_codegen.cpp b/clang/test/OpenMP/parallel_master_taskloop_lastprivate_codegen.cpp index 409a983f3121d94..f86d53777bd4a18 100644 --- a/clang/test/OpenMP/parallel_master_taskloop_lastprivate_codegen.cpp +++ b/clang/test/OpenMP/parallel_master_taskloop_lastprivate_codegen.cpp @@ -212,9 +212,8 @@ void loop() { // CHECK1-NEXT: call void @_ZN1SIdEC1Ev(ptr noundef nonnull align 8 dereferenceable(8) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[ARRAYINIT_BEGIN]], double noundef 1.000000e+00) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[S_ARR]], double noundef 1.000000e+00) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[ARRAYINIT_ELEMENT]], double noundef 2.000000e+00) // CHECK1-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[VAR]], double noundef 3.000000e+00) // CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB1:[0-9]+]], i32 4, ptr @main.omp_outlined, ptr [[VEC]], ptr [[T_VAR]], ptr [[S_ARR]], ptr [[VAR]]) @@ -580,9 +579,8 @@ void loop() { // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 128 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[VAR]], i32 noundef 3) // CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB1]], i32 4, ptr @_Z5tmainIiET_v.omp_outlined, ptr [[VEC]], ptr [[T_VAR]], ptr [[S_ARR]], ptr [[VAR]]) diff --git a/clang/test/OpenMP/parallel_master_taskloop_simd_firstprivate_codegen.cpp b/clang/test/OpenMP/parallel_master_taskloop_simd_firstprivate_codegen.cpp index 46c35833a4e635c..68a7257a019c28c 100644 --- a/clang/test/OpenMP/parallel_master_taskloop_simd_firstprivate_codegen.cpp +++ b/clang/test/OpenMP/parallel_master_taskloop_simd_firstprivate_codegen.cpp @@ -206,9 +206,8 @@ void array_func(int n, float a[n], St s[2]) { // CHECK-NEXT: call void @_ZN1SIdEC1ERKS0_d(ptr noundef nonnull align 8 dereferenceable(8) [[TEST]], ptr noundef nonnull align 8 dereferenceable(8) [[TTT]], double noundef 0.000000e+00) // CHECK-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[ARRAYINIT_BEGIN]], double noundef 1.000000e+00) -// CHECK-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[S_ARR]], double noundef 1.000000e+00) +// CHECK-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[ARRAYINIT_ELEMENT]], double noundef 2.000000e+00) // CHECK-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[VAR]], double noundef 3.000000e+00) // CHECK-NEXT: [[TMP0:%.*]] = load i32, ptr [[T_VAR]], align 4 @@ -569,9 +568,8 @@ void array_func(int n, float a[n], St s[2]) { // CHECK-NEXT: call void @_ZN1SIiEC1ERKS0_i(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]], ptr noundef nonnull align 4 dereferenceable(4) [[TTT]], i32 noundef 0) // CHECK-NEXT: store i32 0, ptr [[T_VAR]], align 128 // CHECK-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[VAR]], i32 noundef 3) // CHECK-NEXT: [[TMP0:%.*]] = load i32, ptr [[T_VAR]], align 128 @@ -1523,9 +1521,8 @@ void array_func(int n, float a[n], St s[2]) { // SIMD-ONLY0-NEXT: call void @_ZN1SIdEC1ERKS0_d(ptr noundef nonnull align 8 dereferenceable(8) [[TEST]], ptr noundef nonnull align 8 dereferenceable(8) [[TTT]], double noundef 0.000000e+00) // SIMD-ONLY0-NEXT: store i32 0, ptr [[T_VAR]], align 4 // SIMD-ONLY0-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// SIMD-ONLY0-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// SIMD-ONLY0-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[ARRAYINIT_BEGIN]], double noundef 1.000000e+00) -// SIMD-ONLY0-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// SIMD-ONLY0-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[S_ARR]], double noundef 1.000000e+00) +// SIMD-ONLY0-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // SIMD-ONLY0-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[ARRAYINIT_ELEMENT]], double noundef 2.000000e+00) // SIMD-ONLY0-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[VAR]], double noundef 3.000000e+00) // SIMD-ONLY0-NEXT: store i64 0, ptr [[DOTOMP_LB]], align 8 @@ -1638,9 +1635,8 @@ void array_func(int n, float a[n], St s[2]) { // SIMD-ONLY0-NEXT: call void @_ZN1SIiEC1ERKS0_i(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]], ptr noundef nonnull align 4 dereferenceable(4) [[TTT]], i32 noundef 0) // SIMD-ONLY0-NEXT: store i32 0, ptr [[T_VAR]], align 128 // SIMD-ONLY0-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// SIMD-ONLY0-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// SIMD-ONLY0-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// SIMD-ONLY0-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// SIMD-ONLY0-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// SIMD-ONLY0-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // SIMD-ONLY0-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // SIMD-ONLY0-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[VAR]], i32 noundef 3) // SIMD-ONLY0-NEXT: store i64 0, ptr [[DOTOMP_LB]], align 8 @@ -1880,9 +1876,8 @@ void array_func(int n, float a[n], St s[2]) { // SIMD-ONLY1-NEXT: call void @_ZN1SIdEC1ERKS0_d(ptr noundef nonnull align 8 dereferenceable(8) [[TEST]], ptr noundef nonnull align 8 dereferenceable(8) [[TTT]], double noundef 0.000000e+00) // SIMD-ONLY1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // SIMD-ONLY1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// SIMD-ONLY1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// SIMD-ONLY1-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[ARRAYINIT_BEGIN]], double noundef 1.000000e+00) -// SIMD-ONLY1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// SIMD-ONLY1-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[S_ARR]], double noundef 1.000000e+00) +// SIMD-ONLY1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // SIMD-ONLY1-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[ARRAYINIT_ELEMENT]], double noundef 2.000000e+00) // SIMD-ONLY1-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[VAR]], double noundef 3.000000e+00) // SIMD-ONLY1-NEXT: store i64 0, ptr [[DOTOMP_LB]], align 8 @@ -1995,9 +1990,8 @@ void array_func(int n, float a[n], St s[2]) { // SIMD-ONLY1-NEXT: call void @_ZN1SIiEC1ERKS0_i(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]], ptr noundef nonnull align 4 dereferenceable(4) [[TTT]], i32 noundef 0) // SIMD-ONLY1-NEXT: store i32 0, ptr [[T_VAR]], align 128 // SIMD-ONLY1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// SIMD-ONLY1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// SIMD-ONLY1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// SIMD-ONLY1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// SIMD-ONLY1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// SIMD-ONLY1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // SIMD-ONLY1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // SIMD-ONLY1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[VAR]], i32 noundef 3) // SIMD-ONLY1-NEXT: store i64 0, ptr [[DOTOMP_LB]], align 8 diff --git a/clang/test/OpenMP/parallel_master_taskloop_simd_lastprivate_codegen.cpp b/clang/test/OpenMP/parallel_master_taskloop_simd_lastprivate_codegen.cpp index 5e2c192497b5b93..8495cfa2e027e55 100644 --- a/clang/test/OpenMP/parallel_master_taskloop_simd_lastprivate_codegen.cpp +++ b/clang/test/OpenMP/parallel_master_taskloop_simd_lastprivate_codegen.cpp @@ -212,9 +212,8 @@ void loop() { // CHECK1-NEXT: call void @_ZN1SIdEC1Ev(ptr noundef nonnull align 8 dereferenceable(8) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[ARRAYINIT_BEGIN]], double noundef 1.000000e+00) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[S_ARR]], double noundef 1.000000e+00) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[ARRAYINIT_ELEMENT]], double noundef 2.000000e+00) // CHECK1-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[VAR]], double noundef 3.000000e+00) // CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB1:[0-9]+]], i32 4, ptr @main.omp_outlined, ptr [[VEC]], ptr [[T_VAR]], ptr [[S_ARR]], ptr [[VAR]]) @@ -580,9 +579,8 @@ void loop() { // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 128 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[VAR]], i32 noundef 3) // CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB1]], i32 4, ptr @_Z5tmainIiET_v.omp_outlined, ptr [[VEC]], ptr [[T_VAR]], ptr [[S_ARR]], ptr [[VAR]]) @@ -1755,9 +1753,8 @@ void loop() { // CHECK7-NEXT: call void @_ZN1SIdEC1Ev(ptr noundef nonnull align 8 dereferenceable(8) [[TEST]]) // CHECK7-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK7-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK7-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK7-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[ARRAYINIT_BEGIN]], double noundef 1.000000e+00) -// CHECK7-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK7-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[S_ARR]], double noundef 1.000000e+00) +// CHECK7-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK7-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[ARRAYINIT_ELEMENT]], double noundef 2.000000e+00) // CHECK7-NEXT: call void @_ZN1SIdEC1Ed(ptr noundef nonnull align 8 dereferenceable(8) [[VAR]], double noundef 3.000000e+00) // CHECK7-NEXT: store i64 0, ptr [[DOTOMP_LB]], align 8 @@ -1911,9 +1908,8 @@ void loop() { // CHECK7-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK7-NEXT: store i32 0, ptr [[T_VAR]], align 128 // CHECK7-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK7-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK7-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK7-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[VAR]], i32 noundef 3) // CHECK7-NEXT: store i64 0, ptr [[DOTOMP_LB]], align 8 diff --git a/clang/test/OpenMP/parallel_private_codegen.cpp b/clang/test/OpenMP/parallel_private_codegen.cpp index 586399ba89a6bfa..b86f4e48a335277 100644 --- a/clang/test/OpenMP/parallel_private_codegen.cpp +++ b/clang/test/OpenMP/parallel_private_codegen.cpp @@ -182,9 +182,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[VAR]], float noundef 3.000000e+00) // CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB1:[0-9]+]], i32 0, ptr @main.omp_outlined) @@ -309,9 +308,8 @@ int main() { // CHECK1-NEXT: call void @_ZN3SSTIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[SST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 128 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 128 [[VEC]], ptr align 128 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[VAR]], i32 noundef 3) // CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB1]], i32 0, ptr @_Z5tmainIiET_v.omp_outlined) diff --git a/clang/test/OpenMP/parallel_reduction_codegen.cpp b/clang/test/OpenMP/parallel_reduction_codegen.cpp index 88a10a16615d0a0..d27ef4913a38649 100644 --- a/clang/test/OpenMP/parallel_reduction_codegen.cpp +++ b/clang/test/OpenMP/parallel_reduction_codegen.cpp @@ -511,9 +511,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store float 0.000000e+00, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[VAR]], float noundef 3.000000e+00) // CHECK1-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[VAR1]]) @@ -997,9 +996,8 @@ int main() { // CHECK1-NEXT: call void @_ZN3SSTIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[SST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 128 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[VAR]], i32 noundef 3) // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[VAR1]]) diff --git a/clang/test/OpenMP/sections_firstprivate_codegen.cpp b/clang/test/OpenMP/sections_firstprivate_codegen.cpp index af67c429ac1125d..3cf6cc29d56ae62 100644 --- a/clang/test/OpenMP/sections_firstprivate_codegen.cpp +++ b/clang/test/OpenMP/sections_firstprivate_codegen.cpp @@ -402,9 +402,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[VAR]], i32 noundef 3) // CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB2]], i32 4, ptr @_Z5tmainIiET_v.omp_outlined, ptr [[T_VAR]], ptr [[VEC]], ptr [[S_ARR]], ptr [[VAR]]) diff --git a/clang/test/OpenMP/sections_lastprivate_codegen.cpp b/clang/test/OpenMP/sections_lastprivate_codegen.cpp index fb7f29077bb50c6..126a3ba87e8182a 100644 --- a/clang/test/OpenMP/sections_lastprivate_codegen.cpp +++ b/clang/test/OpenMP/sections_lastprivate_codegen.cpp @@ -201,9 +201,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[VAR]], float noundef 3.000000e+00) // CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB3:[0-9]+]], i32 5, ptr @main.omp_outlined, ptr [[T_VAR]], ptr [[VEC]], ptr [[S_ARR]], ptr [[VAR]], ptr @_ZZ4mainE5sivar) @@ -465,9 +464,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[VAR]], i32 noundef 3) // CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB3]], i32 4, ptr @_Z5tmainIiET_v.omp_outlined, ptr [[T_VAR]], ptr [[VEC]], ptr [[S_ARR]], ptr [[VAR]]) @@ -879,9 +877,8 @@ int main() { // CHECK5-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK5-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK5-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK5-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK5-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK5-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK5-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK5-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK5-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK5-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[VAR]], float noundef 3.000000e+00) // CHECK5-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB3:[0-9]+]], i32 5, ptr @main.omp_outlined, ptr [[T_VAR]], ptr [[VEC]], ptr [[S_ARR]], ptr [[VAR]], ptr @_ZZ4mainE5sivar) @@ -1161,9 +1158,8 @@ int main() { // CHECK5-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK5-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK5-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK5-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK5-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK5-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[VAR]], i32 noundef 3) // CHECK5-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB3]], i32 4, ptr @_Z5tmainIiET_v.omp_outlined, ptr [[T_VAR]], ptr [[VEC]], ptr [[S_ARR]], ptr [[VAR]]) diff --git a/clang/test/OpenMP/sections_private_codegen.cpp b/clang/test/OpenMP/sections_private_codegen.cpp index 034a8b38dac4c9d..ee29fd953f97aee 100644 --- a/clang/test/OpenMP/sections_private_codegen.cpp +++ b/clang/test/OpenMP/sections_private_codegen.cpp @@ -117,9 +117,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[VAR]], float noundef 3.000000e+00) // CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB3:[0-9]+]], i32 0, ptr @main.omp_outlined) @@ -274,9 +273,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[VAR]], i32 noundef 3) // CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB3]], i32 0, ptr @_Z5tmainIiET_v.omp_outlined) diff --git a/clang/test/OpenMP/sections_reduction_codegen.cpp b/clang/test/OpenMP/sections_reduction_codegen.cpp index 6166b5b3f9486c8..3eb7a1f8c54e5a8 100644 --- a/clang/test/OpenMP/sections_reduction_codegen.cpp +++ b/clang/test/OpenMP/sections_reduction_codegen.cpp @@ -196,9 +196,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store float 0.000000e+00, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[VAR]], float noundef 3.000000e+00) // CHECK1-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[VAR1]]) @@ -546,9 +545,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[VAR]], i32 noundef 3) // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[VAR1]]) diff --git a/clang/test/OpenMP/simd_private_taskloop_codegen.cpp b/clang/test/OpenMP/simd_private_taskloop_codegen.cpp index 50726b201523a0f..7e81f6f4dac600d 100644 --- a/clang/test/OpenMP/simd_private_taskloop_codegen.cpp +++ b/clang/test/OpenMP/simd_private_taskloop_codegen.cpp @@ -470,9 +470,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: call void @__kmpc_taskgroup(ptr @[[GLOB1]], i32 [[TMP0]]) @@ -850,9 +849,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.2], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_2]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_2]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: call void @__kmpc_taskgroup(ptr @[[GLOB1]], i32 [[TMP0]]) @@ -1209,9 +1207,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: call void @__kmpc_taskgroup(ptr @[[GLOB1]], i32 [[TMP0]]) @@ -1587,9 +1584,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.2], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_2]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_2]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: call void @__kmpc_taskgroup(ptr @[[GLOB1]], i32 [[TMP0]]) @@ -1955,9 +1951,8 @@ int main() { // CHECK13-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK13-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK13-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK13-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK13-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK13-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK13-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK13-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK13-NEXT: store ptr undef, ptr [[_TMP1]], align 8 @@ -2123,9 +2118,8 @@ int main() { // CHECK13-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK13-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK13-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK13-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK13-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK13-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK13-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK13-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK13-NEXT: store ptr undef, ptr [[_TMP1]], align 8 @@ -2342,9 +2336,8 @@ int main() { // CHECK15-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK15-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK15-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK15-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK15-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK15-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK15-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK15-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK15-NEXT: store ptr undef, ptr [[_TMP1]], align 4 @@ -2508,9 +2501,8 @@ int main() { // CHECK15-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK15-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK15-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK15-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK15-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK15-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK15-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK15-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK15-NEXT: store ptr undef, ptr [[_TMP1]], align 4 diff --git a/clang/test/OpenMP/single_firstprivate_codegen.cpp b/clang/test/OpenMP/single_firstprivate_codegen.cpp index f7156bf1c54d67c..4b4c96655fccbff 100644 --- a/clang/test/OpenMP/single_firstprivate_codegen.cpp +++ b/clang/test/OpenMP/single_firstprivate_codegen.cpp @@ -357,9 +357,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr nonnull align 4 dereferenceable(4) [[S_ARR]], i32 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 2) // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr nonnull align 4 dereferenceable(4) [[VAR]], i32 3) // CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB1]], i32 4, ptr @_Z5tmainIiET_v.omp_outlined, ptr [[T_VAR]], ptr [[VEC]], ptr [[S_ARR]], ptr [[VAR]]) diff --git a/clang/test/OpenMP/single_private_codegen.cpp b/clang/test/OpenMP/single_private_codegen.cpp index e657d7bcbe8881c..ce79cfd1bd2690c 100644 --- a/clang/test/OpenMP/single_private_codegen.cpp +++ b/clang/test/OpenMP/single_private_codegen.cpp @@ -104,9 +104,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK1-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[VAR]], float noundef 3.000000e+00) // CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB1:[0-9]+]], i32 0, ptr @main.omp_outlined) @@ -226,9 +225,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[VAR]], i32 noundef 3) // CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB1]], i32 0, ptr @_Z5tmainIiET_v.omp_outlined) diff --git a/clang/test/OpenMP/target_teams_distribute_firstprivate_codegen.cpp b/clang/test/OpenMP/target_teams_distribute_firstprivate_codegen.cpp index da69e19a890efa7..ab0a35f791a3168 100644 --- a/clang/test/OpenMP/target_teams_distribute_firstprivate_codegen.cpp +++ b/clang/test/OpenMP/target_teams_distribute_firstprivate_codegen.cpp @@ -546,9 +546,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK1-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1346,9 +1345,8 @@ int main() { // CHECK3-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK3-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK3-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK3-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK3-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK3-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 diff --git a/clang/test/OpenMP/target_teams_distribute_lastprivate_codegen.cpp b/clang/test/OpenMP/target_teams_distribute_lastprivate_codegen.cpp index e23435d13e9aecd..adbab5d8775f7f3 100644 --- a/clang/test/OpenMP/target_teams_distribute_lastprivate_codegen.cpp +++ b/clang/test/OpenMP/target_teams_distribute_lastprivate_codegen.cpp @@ -475,9 +475,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -799,9 +798,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1176,9 +1174,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -1498,9 +1495,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 diff --git a/clang/test/OpenMP/target_teams_distribute_parallel_for_firstprivate_codegen.cpp b/clang/test/OpenMP/target_teams_distribute_parallel_for_firstprivate_codegen.cpp index bf1fc1d7cbae5f6..0806a71f95bcc52 100644 --- a/clang/test/OpenMP/target_teams_distribute_parallel_for_firstprivate_codegen.cpp +++ b/clang/test/OpenMP/target_teams_distribute_parallel_for_firstprivate_codegen.cpp @@ -737,9 +737,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK1-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1793,9 +1792,8 @@ int main() { // CHECK3-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK3-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK3-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK3-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK3-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK3-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 diff --git a/clang/test/OpenMP/target_teams_distribute_parallel_for_lastprivate_codegen.cpp b/clang/test/OpenMP/target_teams_distribute_parallel_for_lastprivate_codegen.cpp index ae900d2dc53b9c1..6f14f5113ee274e 100644 --- a/clang/test/OpenMP/target_teams_distribute_parallel_for_lastprivate_codegen.cpp +++ b/clang/test/OpenMP/target_teams_distribute_parallel_for_lastprivate_codegen.cpp @@ -736,9 +736,8 @@ int main() { // CHECK5-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK5-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK5-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK5-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK5-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK5-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK5-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK5-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK5-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK5-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK5-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1214,9 +1213,8 @@ int main() { // CHECK5-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK5-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK5-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK5-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK5-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK5-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK5-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK5-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1736,9 +1734,8 @@ int main() { // CHECK7-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK7-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK7-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK7-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK7-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK7-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK7-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK7-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK7-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK7-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK7-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -2208,9 +2205,8 @@ int main() { // CHECK7-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK7-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK7-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK7-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK7-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK7-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK7-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK7-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 diff --git a/clang/test/OpenMP/target_teams_distribute_parallel_for_private_codegen.cpp b/clang/test/OpenMP/target_teams_distribute_parallel_for_private_codegen.cpp index 49e4681c051530b..05ab657a3c83ea3 100644 --- a/clang/test/OpenMP/target_teams_distribute_parallel_for_private_codegen.cpp +++ b/clang/test/OpenMP/target_teams_distribute_parallel_for_private_codegen.cpp @@ -564,9 +564,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK1-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK1-NEXT: store ptr undef, ptr [[_TMP1]], align 8 @@ -1291,9 +1290,8 @@ int main() { // CHECK3-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK3-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK3-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK3-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK3-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK3-NEXT: store ptr undef, ptr [[_TMP1]], align 4 diff --git a/clang/test/OpenMP/target_teams_distribute_parallel_for_simd_firstprivate_codegen.cpp b/clang/test/OpenMP/target_teams_distribute_parallel_for_simd_firstprivate_codegen.cpp index ea05aecebf3713c..fbf0b842542258b 100644 --- a/clang/test/OpenMP/target_teams_distribute_parallel_for_simd_firstprivate_codegen.cpp +++ b/clang/test/OpenMP/target_teams_distribute_parallel_for_simd_firstprivate_codegen.cpp @@ -749,9 +749,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK1-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1833,9 +1832,8 @@ int main() { // CHECK3-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK3-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK3-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK3-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK3-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK3-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -2856,9 +2854,8 @@ int main() { // CHECK7-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK7-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK7-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK7-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK7-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK7-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK7-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK7-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -3177,9 +3174,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK9-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 diff --git a/clang/test/OpenMP/target_teams_distribute_parallel_for_simd_lastprivate_codegen.cpp b/clang/test/OpenMP/target_teams_distribute_parallel_for_simd_lastprivate_codegen.cpp index b2b9e2082fa3a8d..cb8e091e8b02cdf 100644 --- a/clang/test/OpenMP/target_teams_distribute_parallel_for_simd_lastprivate_codegen.cpp +++ b/clang/test/OpenMP/target_teams_distribute_parallel_for_simd_lastprivate_codegen.cpp @@ -764,9 +764,8 @@ int main() { // CHECK5-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK5-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK5-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK5-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK5-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK5-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK5-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK5-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK5-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK5-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK5-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1256,9 +1255,8 @@ int main() { // CHECK5-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK5-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK5-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK5-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK5-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK5-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK5-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK5-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1792,9 +1790,8 @@ int main() { // CHECK7-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK7-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK7-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK7-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK7-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK7-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK7-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK7-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK7-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK7-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK7-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -2278,9 +2275,8 @@ int main() { // CHECK7-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK7-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK7-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK7-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK7-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK7-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK7-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK7-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -2848,9 +2844,8 @@ int main() { // CHECK13-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK13-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK13-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK13-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK13-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK13-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK13-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK13-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK13-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -3012,9 +3007,8 @@ int main() { // CHECK13-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK13-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK13-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK13-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK13-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK13-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK13-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK13-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK13-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -3247,9 +3241,8 @@ int main() { // CHECK15-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK15-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK15-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK15-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK15-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK15-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK15-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK15-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK15-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -3409,9 +3402,8 @@ int main() { // CHECK15-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK15-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK15-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK15-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK15-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK15-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK15-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK15-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK15-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 diff --git a/clang/test/OpenMP/target_teams_distribute_parallel_for_simd_private_codegen.cpp b/clang/test/OpenMP/target_teams_distribute_parallel_for_simd_private_codegen.cpp index 0cc07c86aaea7d7..c7b919cfbb443a0 100644 --- a/clang/test/OpenMP/target_teams_distribute_parallel_for_simd_private_codegen.cpp +++ b/clang/test/OpenMP/target_teams_distribute_parallel_for_simd_private_codegen.cpp @@ -578,9 +578,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK1-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK1-NEXT: store ptr undef, ptr [[_TMP1]], align 8 @@ -1333,9 +1332,8 @@ int main() { // CHECK3-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK3-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK3-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK3-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK3-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK3-NEXT: store ptr undef, ptr [[_TMP1]], align 4 @@ -2178,9 +2176,8 @@ int main() { // CHECK7-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK7-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK7-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK7-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK7-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK7-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK7-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK7-NEXT: store ptr undef, ptr [[_TMP1]], align 8 @@ -2554,9 +2551,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK9-NEXT: store ptr undef, ptr [[_TMP1]], align 4 diff --git a/clang/test/OpenMP/target_teams_distribute_private_codegen.cpp b/clang/test/OpenMP/target_teams_distribute_private_codegen.cpp index 3cb2d2c2f9b43ef..b718173246f383f 100644 --- a/clang/test/OpenMP/target_teams_distribute_private_codegen.cpp +++ b/clang/test/OpenMP/target_teams_distribute_private_codegen.cpp @@ -402,9 +402,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK1-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK1-NEXT: store ptr undef, ptr [[_TMP1]], align 8 @@ -928,9 +927,8 @@ int main() { // CHECK3-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK3-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK3-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK3-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK3-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK3-NEXT: store ptr undef, ptr [[_TMP1]], align 4 diff --git a/clang/test/OpenMP/target_teams_distribute_simd_firstprivate_codegen.cpp b/clang/test/OpenMP/target_teams_distribute_simd_firstprivate_codegen.cpp index d22aeee120d5d5c..3e1cfb00209041b 100644 --- a/clang/test/OpenMP/target_teams_distribute_simd_firstprivate_codegen.cpp +++ b/clang/test/OpenMP/target_teams_distribute_simd_firstprivate_codegen.cpp @@ -553,9 +553,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK1-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1367,9 +1366,8 @@ int main() { // CHECK3-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK3-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK3-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK3-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK3-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK3-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -1911,9 +1909,8 @@ int main() { // CHECK5-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK5-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK5-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK5-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK5-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK5-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK5-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK5-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -2231,9 +2228,8 @@ int main() { // CHECK7-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK7-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK7-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK7-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK7-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK7-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK7-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK7-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 diff --git a/clang/test/OpenMP/target_teams_distribute_simd_lastprivate_codegen.cpp b/clang/test/OpenMP/target_teams_distribute_simd_lastprivate_codegen.cpp index 93c570ac0604ff4..fc17caae823282d 100644 --- a/clang/test/OpenMP/target_teams_distribute_simd_lastprivate_codegen.cpp +++ b/clang/test/OpenMP/target_teams_distribute_simd_lastprivate_codegen.cpp @@ -525,9 +525,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -856,9 +855,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1240,9 +1238,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -1569,9 +1566,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -1955,9 +1951,8 @@ int main() { // CHECK13-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK13-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK13-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK13-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK13-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK13-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK13-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK13-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK13-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -2118,9 +2113,8 @@ int main() { // CHECK13-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK13-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK13-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK13-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK13-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK13-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK13-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK13-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK13-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -2352,9 +2346,8 @@ int main() { // CHECK15-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK15-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK15-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK15-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK15-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK15-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK15-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK15-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK15-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -2513,9 +2506,8 @@ int main() { // CHECK15-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK15-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK15-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK15-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK15-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK15-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK15-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK15-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK15-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 diff --git a/clang/test/OpenMP/target_teams_distribute_simd_private_codegen.cpp b/clang/test/OpenMP/target_teams_distribute_simd_private_codegen.cpp index 4e9b4e1027e6863..c1f3530e484c27c 100644 --- a/clang/test/OpenMP/target_teams_distribute_simd_private_codegen.cpp +++ b/clang/test/OpenMP/target_teams_distribute_simd_private_codegen.cpp @@ -409,9 +409,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK1-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK1-NEXT: store ptr undef, ptr [[_TMP1]], align 8 @@ -949,9 +948,8 @@ int main() { // CHECK3-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK3-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK3-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK3-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK3-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK3-NEXT: store ptr undef, ptr [[_TMP1]], align 4 @@ -1377,9 +1375,8 @@ int main() { // CHECK5-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK5-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK5-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK5-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK5-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK5-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK5-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK5-NEXT: store ptr undef, ptr [[_TMP1]], align 8 @@ -1753,9 +1750,8 @@ int main() { // CHECK7-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK7-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK7-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK7-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK7-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK7-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK7-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK7-NEXT: store ptr undef, ptr [[_TMP1]], align 4 diff --git a/clang/test/OpenMP/target_teams_generic_loop_private_codegen.cpp b/clang/test/OpenMP/target_teams_generic_loop_private_codegen.cpp index 9b3d77f5b0adcf1..5efad5b4327cfc6 100644 --- a/clang/test/OpenMP/target_teams_generic_loop_private_codegen.cpp +++ b/clang/test/OpenMP/target_teams_generic_loop_private_codegen.cpp @@ -463,9 +463,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK1-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK1-NEXT: store ptr undef, ptr [[_TMP1]], align 8 @@ -989,9 +988,8 @@ int main() { // CHECK3-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK3-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK3-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK3-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK3-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK3-NEXT: store ptr undef, ptr [[_TMP1]], align 4 diff --git a/clang/test/OpenMP/teams_distribute_firstprivate_codegen.cpp b/clang/test/OpenMP/teams_distribute_firstprivate_codegen.cpp index 8b3e657428ec642..2ebedb8dd03eeec 100644 --- a/clang/test/OpenMP/teams_distribute_firstprivate_codegen.cpp +++ b/clang/test/OpenMP/teams_distribute_firstprivate_codegen.cpp @@ -549,9 +549,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK1-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1351,9 +1350,8 @@ int main() { // CHECK3-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK3-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK3-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK3-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK3-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK3-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 diff --git a/clang/test/OpenMP/teams_distribute_lastprivate_codegen.cpp b/clang/test/OpenMP/teams_distribute_lastprivate_codegen.cpp index 66c952f3281fb54..96ba444d402efc4 100644 --- a/clang/test/OpenMP/teams_distribute_lastprivate_codegen.cpp +++ b/clang/test/OpenMP/teams_distribute_lastprivate_codegen.cpp @@ -469,9 +469,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -789,9 +788,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1165,9 +1163,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -1483,9 +1480,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 diff --git a/clang/test/OpenMP/teams_distribute_parallel_for_firstprivate_codegen.cpp b/clang/test/OpenMP/teams_distribute_parallel_for_firstprivate_codegen.cpp index 0726fb659e7d15d..8c411dee8a97adb 100644 --- a/clang/test/OpenMP/teams_distribute_parallel_for_firstprivate_codegen.cpp +++ b/clang/test/OpenMP/teams_distribute_parallel_for_firstprivate_codegen.cpp @@ -713,9 +713,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK1-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1771,9 +1770,8 @@ int main() { // CHECK3-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK3-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK3-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK3-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK3-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK3-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 diff --git a/clang/test/OpenMP/teams_distribute_parallel_for_lastprivate_codegen.cpp b/clang/test/OpenMP/teams_distribute_parallel_for_lastprivate_codegen.cpp index 06fc87d7da4c6c3..3e4afe969013e94 100644 --- a/clang/test/OpenMP/teams_distribute_parallel_for_lastprivate_codegen.cpp +++ b/clang/test/OpenMP/teams_distribute_parallel_for_lastprivate_codegen.cpp @@ -717,9 +717,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1185,9 +1184,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1703,9 +1701,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -2165,9 +2162,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 diff --git a/clang/test/OpenMP/teams_distribute_parallel_for_private_codegen.cpp b/clang/test/OpenMP/teams_distribute_parallel_for_private_codegen.cpp index 4d10d16f47ed871..c9535d156d2ccb2 100644 --- a/clang/test/OpenMP/teams_distribute_parallel_for_private_codegen.cpp +++ b/clang/test/OpenMP/teams_distribute_parallel_for_private_codegen.cpp @@ -526,9 +526,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK1-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK1-NEXT: store ptr undef, ptr [[_TMP1]], align 8 @@ -1253,9 +1252,8 @@ int main() { // CHECK3-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK3-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK3-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK3-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK3-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK3-NEXT: store ptr undef, ptr [[_TMP1]], align 4 diff --git a/clang/test/OpenMP/teams_distribute_parallel_for_simd_firstprivate_codegen.cpp b/clang/test/OpenMP/teams_distribute_parallel_for_simd_firstprivate_codegen.cpp index 7d9e2ab89676f68..7c13f557c1c13ac 100644 --- a/clang/test/OpenMP/teams_distribute_parallel_for_simd_firstprivate_codegen.cpp +++ b/clang/test/OpenMP/teams_distribute_parallel_for_simd_firstprivate_codegen.cpp @@ -730,9 +730,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK1-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1816,9 +1815,8 @@ int main() { // CHECK3-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK3-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK3-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK3-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK3-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK3-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -2496,9 +2494,8 @@ int main() { // CHECK5-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK5-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK5-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK5-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK5-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK5-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK5-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK5-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -2819,9 +2816,8 @@ int main() { // CHECK7-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK7-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK7-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK7-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK7-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK7-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK7-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK7-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 diff --git a/clang/test/OpenMP/teams_distribute_parallel_for_simd_lastprivate_codegen.cpp b/clang/test/OpenMP/teams_distribute_parallel_for_simd_lastprivate_codegen.cpp index 8643cb9bb84a8bd..77a51286ff4970c 100644 --- a/clang/test/OpenMP/teams_distribute_parallel_for_simd_lastprivate_codegen.cpp +++ b/clang/test/OpenMP/teams_distribute_parallel_for_simd_lastprivate_codegen.cpp @@ -790,9 +790,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1272,9 +1271,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1804,9 +1802,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -2280,9 +2277,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -2811,9 +2807,8 @@ int main() { // CHECK13-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK13-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK13-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK13-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK13-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK13-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK13-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK13-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK13-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -2977,9 +2972,8 @@ int main() { // CHECK13-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK13-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK13-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK13-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK13-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK13-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK13-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK13-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK13-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -3214,9 +3208,8 @@ int main() { // CHECK15-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK15-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK15-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK15-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK15-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK15-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK15-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK15-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK15-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -3378,9 +3371,8 @@ int main() { // CHECK15-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK15-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK15-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK15-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK15-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK15-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK15-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK15-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK15-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 diff --git a/clang/test/OpenMP/teams_distribute_parallel_for_simd_private_codegen.cpp b/clang/test/OpenMP/teams_distribute_parallel_for_simd_private_codegen.cpp index f40acb42f9dae46..4d0c7f00d316a2e 100644 --- a/clang/test/OpenMP/teams_distribute_parallel_for_simd_private_codegen.cpp +++ b/clang/test/OpenMP/teams_distribute_parallel_for_simd_private_codegen.cpp @@ -542,9 +542,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK1-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK1-NEXT: store ptr undef, ptr [[_TMP1]], align 8 @@ -1297,9 +1296,8 @@ int main() { // CHECK3-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK3-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK3-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK3-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK3-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK3-NEXT: store ptr undef, ptr [[_TMP1]], align 4 @@ -1832,9 +1830,8 @@ int main() { // CHECK5-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK5-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK5-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK5-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK5-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK5-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK5-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK5-NEXT: store ptr undef, ptr [[_TMP1]], align 8 @@ -2208,9 +2205,8 @@ int main() { // CHECK7-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK7-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK7-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK7-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK7-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK7-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK7-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK7-NEXT: store ptr undef, ptr [[_TMP1]], align 4 diff --git a/clang/test/OpenMP/teams_distribute_private_codegen.cpp b/clang/test/OpenMP/teams_distribute_private_codegen.cpp index 78b42e3194c7960..1f49d9334f31f70 100644 --- a/clang/test/OpenMP/teams_distribute_private_codegen.cpp +++ b/clang/test/OpenMP/teams_distribute_private_codegen.cpp @@ -405,9 +405,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK1-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK1-NEXT: store ptr undef, ptr [[_TMP1]], align 8 @@ -931,9 +930,8 @@ int main() { // CHECK3-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK3-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK3-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK3-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK3-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK3-NEXT: store ptr undef, ptr [[_TMP1]], align 4 diff --git a/clang/test/OpenMP/teams_distribute_simd_firstprivate_codegen.cpp b/clang/test/OpenMP/teams_distribute_simd_firstprivate_codegen.cpp index 2d3fccd90c0a5d1..9e6825d57cf9ad5 100644 --- a/clang/test/OpenMP/teams_distribute_simd_firstprivate_codegen.cpp +++ b/clang/test/OpenMP/teams_distribute_simd_firstprivate_codegen.cpp @@ -556,9 +556,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK1-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1372,9 +1371,8 @@ int main() { // CHECK3-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK3-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK3-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK3-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK3-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK3-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -1919,9 +1917,8 @@ int main() { // CHECK5-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK5-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK5-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK5-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK5-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK5-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK5-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK5-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -2241,9 +2238,8 @@ int main() { // CHECK7-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK7-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK7-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK7-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK7-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK7-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK7-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK7-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 diff --git a/clang/test/OpenMP/teams_distribute_simd_lastprivate_codegen.cpp b/clang/test/OpenMP/teams_distribute_simd_lastprivate_codegen.cpp index ec95ab55b1552a9..3f01eab5a7352bc 100644 --- a/clang/test/OpenMP/teams_distribute_simd_lastprivate_codegen.cpp +++ b/clang/test/OpenMP/teams_distribute_simd_lastprivate_codegen.cpp @@ -519,9 +519,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -846,9 +845,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK9-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK9-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -1229,9 +1227,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -1554,9 +1551,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK11-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK11-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -1940,9 +1936,8 @@ int main() { // CHECK13-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK13-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK13-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK13-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK13-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK13-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK13-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK13-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK13-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -2105,9 +2100,8 @@ int main() { // CHECK13-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK13-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK13-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK13-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK13-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK13-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK13-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK13-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK13-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK13-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8 @@ -2341,9 +2335,8 @@ int main() { // CHECK15-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK15-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK15-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK15-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK15-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK15-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK15-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK15-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK15-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 @@ -2504,9 +2497,8 @@ int main() { // CHECK15-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK15-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK15-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK15-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK15-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK15-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK15-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK15-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK15-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK15-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 4 diff --git a/clang/test/OpenMP/teams_distribute_simd_private_codegen.cpp b/clang/test/OpenMP/teams_distribute_simd_private_codegen.cpp index c839268ada75348..4c87b2dcc111be8 100644 --- a/clang/test/OpenMP/teams_distribute_simd_private_codegen.cpp +++ b/clang/test/OpenMP/teams_distribute_simd_private_codegen.cpp @@ -413,9 +413,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK1-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK1-NEXT: store ptr undef, ptr [[_TMP1]], align 8 @@ -953,9 +952,8 @@ int main() { // CHECK3-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK3-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK3-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK3-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK3-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK3-NEXT: store ptr undef, ptr [[_TMP1]], align 4 @@ -1381,9 +1379,8 @@ int main() { // CHECK5-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK5-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK5-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK5-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK5-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK5-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK5-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK5-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK5-NEXT: store ptr undef, ptr [[_TMP1]], align 8 @@ -1757,9 +1754,8 @@ int main() { // CHECK7-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK7-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK7-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK7-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK7-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK7-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK7-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK7-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK7-NEXT: store ptr undef, ptr [[_TMP1]], align 4 diff --git a/clang/test/OpenMP/teams_firstprivate_codegen.cpp b/clang/test/OpenMP/teams_firstprivate_codegen.cpp index 649fae99b21ac35..304bd5dbb50ef38 100644 --- a/clang/test/OpenMP/teams_firstprivate_codegen.cpp +++ b/clang/test/OpenMP/teams_firstprivate_codegen.cpp @@ -283,9 +283,8 @@ void array_func(float a[3], St s[2], int n, long double vla1[n]) { // CHECK9-NEXT: call void @_ZN1SIfEC1Ev(ptr nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float 1.000000e+00) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr nonnull align 4 dereferenceable(4) [[S_ARR]], float 1.000000e+00) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float 2.000000e+00) // CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr nonnull align 4 dereferenceable(4) [[VAR]], float 3.000000e+00) // CHECK9-NEXT: [[TMP0:%.*]] = load i32, ptr [[T_VAR]], align 4 @@ -628,9 +627,8 @@ void array_func(float a[3], St s[2], int n, long double vla1[n]) { // CHECK9-NEXT: call void @_ZN1SIiEC1Ev(ptr nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 128 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 128 [[VEC]], ptr align 128 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 signext 1) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr nonnull align 4 dereferenceable(4) [[S_ARR]], i32 signext 1) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 signext 2) // CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr nonnull align 4 dereferenceable(4) [[VAR]], i32 signext 3) // CHECK9-NEXT: [[TMP0:%.*]] = load i32, ptr [[T_VAR]], align 128 @@ -1077,9 +1075,8 @@ void array_func(float a[3], St s[2], int n, long double vla1[n]) { // CHECK11-NEXT: call void @_ZN1SIfEC1Ev(ptr nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float 1.000000e+00) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr nonnull align 4 dereferenceable(4) [[S_ARR]], float 1.000000e+00) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float 2.000000e+00) // CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr nonnull align 4 dereferenceable(4) [[VAR]], float 3.000000e+00) // CHECK11-NEXT: [[TMP0:%.*]] = load i32, ptr [[T_VAR]], align 4 @@ -1422,9 +1419,8 @@ void array_func(float a[3], St s[2], int n, long double vla1[n]) { // CHECK11-NEXT: call void @_ZN1SIiEC1Ev(ptr nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 128 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 128 [[VEC]], ptr align 128 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 1) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr nonnull align 4 dereferenceable(4) [[S_ARR]], i32 1) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 2) // CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr nonnull align 4 dereferenceable(4) [[VAR]], i32 3) // CHECK11-NEXT: [[TMP0:%.*]] = load i32, ptr [[T_VAR]], align 128 diff --git a/clang/test/OpenMP/teams_generic_loop_private_codegen.cpp b/clang/test/OpenMP/teams_generic_loop_private_codegen.cpp index e955db129d1da60..f44acf211b5e3bb 100644 --- a/clang/test/OpenMP/teams_generic_loop_private_codegen.cpp +++ b/clang/test/OpenMP/teams_generic_loop_private_codegen.cpp @@ -425,9 +425,8 @@ int main() { // CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK1-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK1-NEXT: store ptr [[TEST]], ptr [[VAR]], align 8 // CHECK1-NEXT: store ptr undef, ptr [[_TMP1]], align 8 @@ -951,9 +950,8 @@ int main() { // CHECK3-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK3-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK3-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK3-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK3-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK3-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK3-NEXT: store ptr [[TEST]], ptr [[VAR]], align 4 // CHECK3-NEXT: store ptr undef, ptr [[_TMP1]], align 4 diff --git a/clang/test/OpenMP/teams_private_codegen.cpp b/clang/test/OpenMP/teams_private_codegen.cpp index 1d0b2435ff00581..81c0ea77f1b0487 100644 --- a/clang/test/OpenMP/teams_private_codegen.cpp +++ b/clang/test/OpenMP/teams_private_codegen.cpp @@ -551,9 +551,8 @@ int main() { // CHECK9-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK9-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[VAR]], float noundef 3.000000e+00) // CHECK9-NEXT: [[TMP0:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 0 @@ -718,9 +717,8 @@ int main() { // CHECK9-NEXT: call void @_ZN3SSTIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[SST]]) // CHECK9-NEXT: store i32 0, ptr [[T_VAR]], align 128 // CHECK9-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 128 [[VEC]], ptr align 128 @__const._Z5tmainIiET_v.vec, i64 8, i1 false) -// CHECK9-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i64 0, i64 0 -// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef signext 1) -// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i64 1 +// CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef signext 1) +// CHECK9-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i64 1 // CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef signext 2) // CHECK9-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[VAR]], i32 noundef signext 3) // CHECK9-NEXT: [[TMP0:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 0 @@ -1141,9 +1139,8 @@ int main() { // CHECK11-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[TEST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 4 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const.main.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float noundef 1.000000e+00) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], float noundef 1.000000e+00) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float noundef 2.000000e+00) // CHECK11-NEXT: call void @_ZN1SIfEC1Ef(ptr noundef nonnull align 4 dereferenceable(4) [[VAR]], float noundef 3.000000e+00) // CHECK11-NEXT: [[TMP0:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 0 @@ -1308,9 +1305,8 @@ int main() { // CHECK11-NEXT: call void @_ZN3SSTIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[SST]]) // CHECK11-NEXT: store i32 0, ptr [[T_VAR]], align 128 // CHECK11-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 128 [[VEC]], ptr align 128 @__const._Z5tmainIiET_v.vec, i32 8, i1 false) -// CHECK11-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], ptr [[S_ARR]], i32 0, i32 0 -// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 noundef 1) -// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYINIT_BEGIN]], i32 1 +// CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[S_ARR]], i32 noundef 1) +// CHECK11-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[S_ARR]], i32 1 // CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) // CHECK11-NEXT: call void @_ZN1SIiEC1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[VAR]], i32 noundef 3) // CHECK11-NEXT: [[TMP0:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 0 diff --git a/clang/test/OpenMP/threadprivate_codegen.cpp b/clang/test/OpenMP/threadprivate_codegen.cpp index 5087451b944b981..2ee328008cc4bf0 100644 --- a/clang/test/OpenMP/threadprivate_codegen.cpp +++ b/clang/test/OpenMP/threadprivate_codegen.cpp @@ -1028,46 +1028,43 @@ int foobar() { // CHECK1-NEXT: entry: // CHECK1-NEXT: [[DOTADDR:%.*]] = alloca ptr, align 8 // CHECK1-NEXT: [[ARRAYINIT_ENDOFINIT:%.*]] = alloca ptr, align 8 -// CHECK1-NEXT: [[ARRAYINIT_ENDOFINIT2:%.*]] = alloca ptr, align 8 +// CHECK1-NEXT: [[ARRAYINIT_ENDOFINIT1:%.*]] = alloca ptr, align 8 // CHECK1-NEXT: [[EXN_SLOT:%.*]] = alloca ptr, align 8 // CHECK1-NEXT: [[EHSELECTOR_SLOT:%.*]] = alloca i32, align 4 -// CHECK1-NEXT: [[ARRAYINIT_ENDOFINIT9:%.*]] = alloca ptr, align 8 +// CHECK1-NEXT: [[ARRAYINIT_ENDOFINIT7:%.*]] = alloca ptr, align 8 // CHECK1-NEXT: store ptr [[TMP0]], ptr [[DOTADDR]], align 8 // CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[DOTADDR]], align 8 -// CHECK1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x [3 x %struct.S1]], ptr [[TMP1]], i64 0, i64 0 -// CHECK1-NEXT: store ptr [[ARRAYINIT_BEGIN]], ptr [[ARRAYINIT_ENDOFINIT]], align 8 -// CHECK1-NEXT: [[ARRAYINIT_BEGIN1:%.*]] = getelementptr inbounds [3 x %struct.S1], ptr [[ARRAYINIT_BEGIN]], i64 0, i64 0 -// CHECK1-NEXT: store ptr [[ARRAYINIT_BEGIN1]], ptr [[ARRAYINIT_ENDOFINIT2]], align 8 -// CHECK1-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN1]], i32 noundef 1) +// CHECK1-NEXT: store ptr [[TMP1]], ptr [[ARRAYINIT_ENDOFINIT]], align 8 +// CHECK1-NEXT: store ptr [[TMP1]], ptr [[ARRAYINIT_ENDOFINIT1]], align 8 +// CHECK1-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[TMP1]], i32 noundef 1) // CHECK1-NEXT: to label [[INVOKE_CONT:%.*]] unwind label [[LPAD:%.*]] // CHECK1: invoke.cont: -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S1:%.*]], ptr [[ARRAYINIT_BEGIN1]], i64 1 -// CHECK1-NEXT: store ptr [[ARRAYINIT_ELEMENT]], ptr [[ARRAYINIT_ENDOFINIT2]], align 8 +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S1:%.*]], ptr [[TMP1]], i64 1 +// CHECK1-NEXT: store ptr [[ARRAYINIT_ELEMENT]], ptr [[ARRAYINIT_ENDOFINIT1]], align 8 // CHECK1-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) -// CHECK1-NEXT: to label [[INVOKE_CONT3:%.*]] unwind label [[LPAD]] -// CHECK1: invoke.cont3: -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT4:%.*]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYINIT_ELEMENT]], i64 1 -// CHECK1-NEXT: store ptr [[ARRAYINIT_ELEMENT4]], ptr [[ARRAYINIT_ENDOFINIT2]], align 8 -// CHECK1-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT4]], i32 noundef 3) -// CHECK1-NEXT: to label [[INVOKE_CONT5:%.*]] unwind label [[LPAD]] -// CHECK1: invoke.cont5: -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT7:%.*]] = getelementptr inbounds [3 x %struct.S1], ptr [[ARRAYINIT_BEGIN]], i64 1 -// CHECK1-NEXT: store ptr [[ARRAYINIT_ELEMENT7]], ptr [[ARRAYINIT_ENDOFINIT]], align 8 -// CHECK1-NEXT: [[ARRAYINIT_BEGIN8:%.*]] = getelementptr inbounds [3 x %struct.S1], ptr [[ARRAYINIT_ELEMENT7]], i64 0, i64 0 -// CHECK1-NEXT: store ptr [[ARRAYINIT_BEGIN8]], ptr [[ARRAYINIT_ENDOFINIT9]], align 8 -// CHECK1-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN8]], i32 noundef 4) -// CHECK1-NEXT: to label [[INVOKE_CONT11:%.*]] unwind label [[LPAD10:%.*]] +// CHECK1-NEXT: to label [[INVOKE_CONT2:%.*]] unwind label [[LPAD]] +// CHECK1: invoke.cont2: +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT3:%.*]] = getelementptr inbounds [[STRUCT_S1]], ptr [[TMP1]], i64 2 +// CHECK1-NEXT: store ptr [[ARRAYINIT_ELEMENT3]], ptr [[ARRAYINIT_ENDOFINIT1]], align 8 +// CHECK1-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT3]], i32 noundef 3) +// CHECK1-NEXT: to label [[INVOKE_CONT4:%.*]] unwind label [[LPAD]] +// CHECK1: invoke.cont4: +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT6:%.*]] = getelementptr inbounds [3 x %struct.S1], ptr [[TMP1]], i64 1 +// CHECK1-NEXT: store ptr [[ARRAYINIT_ELEMENT6]], ptr [[ARRAYINIT_ENDOFINIT]], align 8 +// CHECK1-NEXT: store ptr [[ARRAYINIT_ELEMENT6]], ptr [[ARRAYINIT_ENDOFINIT7]], align 8 +// CHECK1-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT6]], i32 noundef 4) +// CHECK1-NEXT: to label [[INVOKE_CONT9:%.*]] unwind label [[LPAD8:%.*]] +// CHECK1: invoke.cont9: +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT10:%.*]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYINIT_ELEMENT6]], i64 1 +// CHECK1-NEXT: store ptr [[ARRAYINIT_ELEMENT10]], ptr [[ARRAYINIT_ENDOFINIT7]], align 8 +// CHECK1-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT10]], i32 noundef 5) +// CHECK1-NEXT: to label [[INVOKE_CONT11:%.*]] unwind label [[LPAD8]] // CHECK1: invoke.cont11: -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT12:%.*]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYINIT_BEGIN8]], i64 1 -// CHECK1-NEXT: store ptr [[ARRAYINIT_ELEMENT12]], ptr [[ARRAYINIT_ENDOFINIT9]], align 8 -// CHECK1-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT12]], i32 noundef 5) -// CHECK1-NEXT: to label [[INVOKE_CONT13:%.*]] unwind label [[LPAD10]] +// CHECK1-NEXT: [[ARRAYINIT_ELEMENT12:%.*]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYINIT_ELEMENT6]], i64 2 +// CHECK1-NEXT: store ptr [[ARRAYINIT_ELEMENT12]], ptr [[ARRAYINIT_ENDOFINIT7]], align 8 +// CHECK1-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT12]], i32 noundef 6) +// CHECK1-NEXT: to label [[INVOKE_CONT13:%.*]] unwind label [[LPAD8]] // CHECK1: invoke.cont13: -// CHECK1-NEXT: [[ARRAYINIT_ELEMENT14:%.*]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYINIT_ELEMENT12]], i64 1 -// CHECK1-NEXT: store ptr [[ARRAYINIT_ELEMENT14]], ptr [[ARRAYINIT_ENDOFINIT9]], align 8 -// CHECK1-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT14]], i32 noundef 6) -// CHECK1-NEXT: to label [[INVOKE_CONT15:%.*]] unwind label [[LPAD10]] -// CHECK1: invoke.cont15: // CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[DOTADDR]], align 8 // CHECK1-NEXT: ret ptr [[TMP2]] // CHECK1: lpad: @@ -1077,55 +1074,55 @@ int foobar() { // CHECK1-NEXT: store ptr [[TMP4]], ptr [[EXN_SLOT]], align 8 // CHECK1-NEXT: [[TMP5:%.*]] = extractvalue { ptr, i32 } [[TMP3]], 1 // CHECK1-NEXT: store i32 [[TMP5]], ptr [[EHSELECTOR_SLOT]], align 4 -// CHECK1-NEXT: [[TMP6:%.*]] = load ptr, ptr [[ARRAYINIT_ENDOFINIT2]], align 8 -// CHECK1-NEXT: [[ARRAYDESTROY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYINIT_BEGIN1]], [[TMP6]] -// CHECK1-NEXT: br i1 [[ARRAYDESTROY_ISEMPTY]], label [[ARRAYDESTROY_DONE6:%.*]], label [[ARRAYDESTROY_BODY:%.*]] +// CHECK1-NEXT: [[TMP6:%.*]] = load ptr, ptr [[ARRAYINIT_ENDOFINIT1]], align 8 +// CHECK1-NEXT: [[ARRAYDESTROY_ISEMPTY:%.*]] = icmp eq ptr [[TMP1]], [[TMP6]] +// CHECK1-NEXT: br i1 [[ARRAYDESTROY_ISEMPTY]], label [[ARRAYDESTROY_DONE5:%.*]], label [[ARRAYDESTROY_BODY:%.*]] // CHECK1: arraydestroy.body: // CHECK1-NEXT: [[ARRAYDESTROY_ELEMENTPAST:%.*]] = phi ptr [ [[TMP6]], [[LPAD]] ], [ [[ARRAYDESTROY_ELEMENT:%.*]], [[ARRAYDESTROY_BODY]] ] // CHECK1-NEXT: [[ARRAYDESTROY_ELEMENT]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYDESTROY_ELEMENTPAST]], i64 -1 // CHECK1-NEXT: call void @_ZN2S1D1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYDESTROY_ELEMENT]]) #[[ATTR3]] -// CHECK1-NEXT: [[ARRAYDESTROY_DONE:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT]], [[ARRAYINIT_BEGIN1]] -// CHECK1-NEXT: br i1 [[ARRAYDESTROY_DONE]], label [[ARRAYDESTROY_DONE6]], label [[ARRAYDESTROY_BODY]] -// CHECK1: arraydestroy.done6: +// CHECK1-NEXT: [[ARRAYDESTROY_DONE:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT]], [[TMP1]] +// CHECK1-NEXT: br i1 [[ARRAYDESTROY_DONE]], label [[ARRAYDESTROY_DONE5]], label [[ARRAYDESTROY_BODY]] +// CHECK1: arraydestroy.done5: // CHECK1-NEXT: br label [[EHCLEANUP:%.*]] -// CHECK1: lpad10: +// CHECK1: lpad8: // CHECK1-NEXT: [[TMP7:%.*]] = landingpad { ptr, i32 } // CHECK1-NEXT: cleanup // CHECK1-NEXT: [[TMP8:%.*]] = extractvalue { ptr, i32 } [[TMP7]], 0 // CHECK1-NEXT: store ptr [[TMP8]], ptr [[EXN_SLOT]], align 8 // CHECK1-NEXT: [[TMP9:%.*]] = extractvalue { ptr, i32 } [[TMP7]], 1 // CHECK1-NEXT: store i32 [[TMP9]], ptr [[EHSELECTOR_SLOT]], align 4 -// CHECK1-NEXT: [[TMP10:%.*]] = load ptr, ptr [[ARRAYINIT_ENDOFINIT9]], align 8 -// CHECK1-NEXT: [[ARRAYDESTROY_ISEMPTY16:%.*]] = icmp eq ptr [[ARRAYINIT_BEGIN8]], [[TMP10]] -// CHECK1-NEXT: br i1 [[ARRAYDESTROY_ISEMPTY16]], label [[ARRAYDESTROY_DONE21:%.*]], label [[ARRAYDESTROY_BODY17:%.*]] -// CHECK1: arraydestroy.body17: -// CHECK1-NEXT: [[ARRAYDESTROY_ELEMENTPAST18:%.*]] = phi ptr [ [[TMP10]], [[LPAD10]] ], [ [[ARRAYDESTROY_ELEMENT19:%.*]], [[ARRAYDESTROY_BODY17]] ] -// CHECK1-NEXT: [[ARRAYDESTROY_ELEMENT19]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYDESTROY_ELEMENTPAST18]], i64 -1 -// CHECK1-NEXT: call void @_ZN2S1D1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYDESTROY_ELEMENT19]]) #[[ATTR3]] -// CHECK1-NEXT: [[ARRAYDESTROY_DONE20:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT19]], [[ARRAYINIT_BEGIN8]] -// CHECK1-NEXT: br i1 [[ARRAYDESTROY_DONE20]], label [[ARRAYDESTROY_DONE21]], label [[ARRAYDESTROY_BODY17]] -// CHECK1: arraydestroy.done21: +// CHECK1-NEXT: [[TMP10:%.*]] = load ptr, ptr [[ARRAYINIT_ENDOFINIT7]], align 8 +// CHECK1-NEXT: [[ARRAYDESTROY_ISEMPTY14:%.*]] = icmp eq ptr [[ARRAYINIT_ELEMENT6]], [[TMP10]] +// CHECK1-NEXT: br i1 [[ARRAYDESTROY_ISEMPTY14]], label [[ARRAYDESTROY_DONE19:%.*]], label [[ARRAYDESTROY_BODY15:%.*]] +// CHECK1: arraydestroy.body15: +// CHECK1-NEXT: [[ARRAYDESTROY_ELEMENTPAST16:%.*]] = phi ptr [ [[TMP10]], [[LPAD8]] ], [ [[ARRAYDESTROY_ELEMENT17:%.*]], [[ARRAYDESTROY_BODY15]] ] +// CHECK1-NEXT: [[ARRAYDESTROY_ELEMENT17]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYDESTROY_ELEMENTPAST16]], i64 -1 +// CHECK1-NEXT: call void @_ZN2S1D1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYDESTROY_ELEMENT17]]) #[[ATTR3]] +// CHECK1-NEXT: [[ARRAYDESTROY_DONE18:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT17]], [[ARRAYINIT_ELEMENT6]] +// CHECK1-NEXT: br i1 [[ARRAYDESTROY_DONE18]], label [[ARRAYDESTROY_DONE19]], label [[ARRAYDESTROY_BODY15]] +// CHECK1: arraydestroy.done19: // CHECK1-NEXT: br label [[EHCLEANUP]] // CHECK1: ehcleanup: // CHECK1-NEXT: [[TMP11:%.*]] = load ptr, ptr [[ARRAYINIT_ENDOFINIT]], align 8 -// CHECK1-NEXT: [[PAD_ARRAYBEGIN:%.*]] = getelementptr inbounds [3 x %struct.S1], ptr [[ARRAYINIT_BEGIN]], i64 0, i64 0 +// CHECK1-NEXT: [[PAD_ARRAYBEGIN:%.*]] = getelementptr inbounds [3 x %struct.S1], ptr [[TMP1]], i64 0, i64 0 // CHECK1-NEXT: [[PAD_ARRAYEND:%.*]] = getelementptr inbounds [3 x %struct.S1], ptr [[TMP11]], i64 0, i64 0 -// CHECK1-NEXT: [[ARRAYDESTROY_ISEMPTY22:%.*]] = icmp eq ptr [[PAD_ARRAYBEGIN]], [[PAD_ARRAYEND]] -// CHECK1-NEXT: br i1 [[ARRAYDESTROY_ISEMPTY22]], label [[ARRAYDESTROY_DONE27:%.*]], label [[ARRAYDESTROY_BODY23:%.*]] -// CHECK1: arraydestroy.body23: -// CHECK1-NEXT: [[ARRAYDESTROY_ELEMENTPAST24:%.*]] = phi ptr [ [[PAD_ARRAYEND]], [[EHCLEANUP]] ], [ [[ARRAYDESTROY_ELEMENT25:%.*]], [[ARRAYDESTROY_BODY23]] ] -// CHECK1-NEXT: [[ARRAYDESTROY_ELEMENT25]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYDESTROY_ELEMENTPAST24]], i64 -1 -// CHECK1-NEXT: call void @_ZN2S1D1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYDESTROY_ELEMENT25]]) #[[ATTR3]] -// CHECK1-NEXT: [[ARRAYDESTROY_DONE26:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT25]], [[PAD_ARRAYBEGIN]] -// CHECK1-NEXT: br i1 [[ARRAYDESTROY_DONE26]], label [[ARRAYDESTROY_DONE27]], label [[ARRAYDESTROY_BODY23]] -// CHECK1: arraydestroy.done27: +// CHECK1-NEXT: [[ARRAYDESTROY_ISEMPTY20:%.*]] = icmp eq ptr [[PAD_ARRAYBEGIN]], [[PAD_ARRAYEND]] +// CHECK1-NEXT: br i1 [[ARRAYDESTROY_ISEMPTY20]], label [[ARRAYDESTROY_DONE25:%.*]], label [[ARRAYDESTROY_BODY21:%.*]] +// CHECK1: arraydestroy.body21: +// CHECK1-NEXT: [[ARRAYDESTROY_ELEMENTPAST22:%.*]] = phi ptr [ [[PAD_ARRAYEND]], [[EHCLEANUP]] ], [ [[ARRAYDESTROY_ELEMENT23:%.*]], [[ARRAYDESTROY_BODY21]] ] +// CHECK1-NEXT: [[ARRAYDESTROY_ELEMENT23]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYDESTROY_ELEMENTPAST22]], i64 -1 +// CHECK1-NEXT: call void @_ZN2S1D1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYDESTROY_ELEMENT23]]) #[[ATTR3]] +// CHECK1-NEXT: [[ARRAYDESTROY_DONE24:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT23]], [[PAD_ARRAYBEGIN]] +// CHECK1-NEXT: br i1 [[ARRAYDESTROY_DONE24]], label [[ARRAYDESTROY_DONE25]], label [[ARRAYDESTROY_BODY21]] +// CHECK1: arraydestroy.done25: // CHECK1-NEXT: br label [[EH_RESUME:%.*]] // CHECK1: eh.resume: // CHECK1-NEXT: [[EXN:%.*]] = load ptr, ptr [[EXN_SLOT]], align 8 // CHECK1-NEXT: [[SEL:%.*]] = load i32, ptr [[EHSELECTOR_SLOT]], align 4 // CHECK1-NEXT: [[LPAD_VAL:%.*]] = insertvalue { ptr, i32 } poison, ptr [[EXN]], 0 -// CHECK1-NEXT: [[LPAD_VAL28:%.*]] = insertvalue { ptr, i32 } [[LPAD_VAL]], i32 [[SEL]], 1 -// CHECK1-NEXT: resume { ptr, i32 } [[LPAD_VAL28]] +// CHECK1-NEXT: [[LPAD_VAL26:%.*]] = insertvalue { ptr, i32 } [[LPAD_VAL]], i32 [[SEL]], 1 +// CHECK1-NEXT: resume { ptr, i32 } [[LPAD_VAL26]] // // // CHECK1-LABEL: define {{[^@]+}}@.__kmpc_global_dtor_..2 @@ -1880,46 +1877,43 @@ int foobar() { // CHECK2-NEXT: entry: // CHECK2-NEXT: [[DOTADDR:%.*]] = alloca ptr, align 8 // CHECK2-NEXT: [[ARRAYINIT_ENDOFINIT:%.*]] = alloca ptr, align 8 -// CHECK2-NEXT: [[ARRAYINIT_ENDOFINIT2:%.*]] = alloca ptr, align 8 +// CHECK2-NEXT: [[ARRAYINIT_ENDOFINIT1:%.*]] = alloca ptr, align 8 // CHECK2-NEXT: [[EXN_SLOT:%.*]] = alloca ptr, align 8 // CHECK2-NEXT: [[EHSELECTOR_SLOT:%.*]] = alloca i32, align 4 -// CHECK2-NEXT: [[ARRAYINIT_ENDOFINIT9:%.*]] = alloca ptr, align 8 +// CHECK2-NEXT: [[ARRAYINIT_ENDOFINIT7:%.*]] = alloca ptr, align 8 // CHECK2-NEXT: store ptr [[TMP0]], ptr [[DOTADDR]], align 8 // CHECK2-NEXT: [[TMP1:%.*]] = load ptr, ptr [[DOTADDR]], align 8 -// CHECK2-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x [3 x %struct.S1]], ptr [[TMP1]], i64 0, i64 0 -// CHECK2-NEXT: store ptr [[ARRAYINIT_BEGIN]], ptr [[ARRAYINIT_ENDOFINIT]], align 8 -// CHECK2-NEXT: [[ARRAYINIT_BEGIN1:%.*]] = getelementptr inbounds [3 x %struct.S1], ptr [[ARRAYINIT_BEGIN]], i64 0, i64 0 -// CHECK2-NEXT: store ptr [[ARRAYINIT_BEGIN1]], ptr [[ARRAYINIT_ENDOFINIT2]], align 8 -// CHECK2-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN1]], i32 noundef 1) +// CHECK2-NEXT: store ptr [[TMP1]], ptr [[ARRAYINIT_ENDOFINIT]], align 8 +// CHECK2-NEXT: store ptr [[TMP1]], ptr [[ARRAYINIT_ENDOFINIT1]], align 8 +// CHECK2-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[TMP1]], i32 noundef 1) // CHECK2-NEXT: to label [[INVOKE_CONT:%.*]] unwind label [[LPAD:%.*]] // CHECK2: invoke.cont: -// CHECK2-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S1:%.*]], ptr [[ARRAYINIT_BEGIN1]], i64 1 -// CHECK2-NEXT: store ptr [[ARRAYINIT_ELEMENT]], ptr [[ARRAYINIT_ENDOFINIT2]], align 8 +// CHECK2-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S1:%.*]], ptr [[TMP1]], i64 1 +// CHECK2-NEXT: store ptr [[ARRAYINIT_ELEMENT]], ptr [[ARRAYINIT_ENDOFINIT1]], align 8 // CHECK2-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) -// CHECK2-NEXT: to label [[INVOKE_CONT3:%.*]] unwind label [[LPAD]] -// CHECK2: invoke.cont3: -// CHECK2-NEXT: [[ARRAYINIT_ELEMENT4:%.*]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYINIT_ELEMENT]], i64 1 -// CHECK2-NEXT: store ptr [[ARRAYINIT_ELEMENT4]], ptr [[ARRAYINIT_ENDOFINIT2]], align 8 -// CHECK2-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT4]], i32 noundef 3) -// CHECK2-NEXT: to label [[INVOKE_CONT5:%.*]] unwind label [[LPAD]] -// CHECK2: invoke.cont5: -// CHECK2-NEXT: [[ARRAYINIT_ELEMENT7:%.*]] = getelementptr inbounds [3 x %struct.S1], ptr [[ARRAYINIT_BEGIN]], i64 1 -// CHECK2-NEXT: store ptr [[ARRAYINIT_ELEMENT7]], ptr [[ARRAYINIT_ENDOFINIT]], align 8 -// CHECK2-NEXT: [[ARRAYINIT_BEGIN8:%.*]] = getelementptr inbounds [3 x %struct.S1], ptr [[ARRAYINIT_ELEMENT7]], i64 0, i64 0 -// CHECK2-NEXT: store ptr [[ARRAYINIT_BEGIN8]], ptr [[ARRAYINIT_ENDOFINIT9]], align 8 -// CHECK2-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN8]], i32 noundef 4) -// CHECK2-NEXT: to label [[INVOKE_CONT11:%.*]] unwind label [[LPAD10:%.*]] +// CHECK2-NEXT: to label [[INVOKE_CONT2:%.*]] unwind label [[LPAD]] +// CHECK2: invoke.cont2: +// CHECK2-NEXT: [[ARRAYINIT_ELEMENT3:%.*]] = getelementptr inbounds [[STRUCT_S1]], ptr [[TMP1]], i64 2 +// CHECK2-NEXT: store ptr [[ARRAYINIT_ELEMENT3]], ptr [[ARRAYINIT_ENDOFINIT1]], align 8 +// CHECK2-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT3]], i32 noundef 3) +// CHECK2-NEXT: to label [[INVOKE_CONT4:%.*]] unwind label [[LPAD]] +// CHECK2: invoke.cont4: +// CHECK2-NEXT: [[ARRAYINIT_ELEMENT6:%.*]] = getelementptr inbounds [3 x %struct.S1], ptr [[TMP1]], i64 1 +// CHECK2-NEXT: store ptr [[ARRAYINIT_ELEMENT6]], ptr [[ARRAYINIT_ENDOFINIT]], align 8 +// CHECK2-NEXT: store ptr [[ARRAYINIT_ELEMENT6]], ptr [[ARRAYINIT_ENDOFINIT7]], align 8 +// CHECK2-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT6]], i32 noundef 4) +// CHECK2-NEXT: to label [[INVOKE_CONT9:%.*]] unwind label [[LPAD8:%.*]] +// CHECK2: invoke.cont9: +// CHECK2-NEXT: [[ARRAYINIT_ELEMENT10:%.*]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYINIT_ELEMENT6]], i64 1 +// CHECK2-NEXT: store ptr [[ARRAYINIT_ELEMENT10]], ptr [[ARRAYINIT_ENDOFINIT7]], align 8 +// CHECK2-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT10]], i32 noundef 5) +// CHECK2-NEXT: to label [[INVOKE_CONT11:%.*]] unwind label [[LPAD8]] // CHECK2: invoke.cont11: -// CHECK2-NEXT: [[ARRAYINIT_ELEMENT12:%.*]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYINIT_BEGIN8]], i64 1 -// CHECK2-NEXT: store ptr [[ARRAYINIT_ELEMENT12]], ptr [[ARRAYINIT_ENDOFINIT9]], align 8 -// CHECK2-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT12]], i32 noundef 5) -// CHECK2-NEXT: to label [[INVOKE_CONT13:%.*]] unwind label [[LPAD10]] +// CHECK2-NEXT: [[ARRAYINIT_ELEMENT12:%.*]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYINIT_ELEMENT6]], i64 2 +// CHECK2-NEXT: store ptr [[ARRAYINIT_ELEMENT12]], ptr [[ARRAYINIT_ENDOFINIT7]], align 8 +// CHECK2-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT12]], i32 noundef 6) +// CHECK2-NEXT: to label [[INVOKE_CONT13:%.*]] unwind label [[LPAD8]] // CHECK2: invoke.cont13: -// CHECK2-NEXT: [[ARRAYINIT_ELEMENT14:%.*]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYINIT_ELEMENT12]], i64 1 -// CHECK2-NEXT: store ptr [[ARRAYINIT_ELEMENT14]], ptr [[ARRAYINIT_ENDOFINIT9]], align 8 -// CHECK2-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT14]], i32 noundef 6) -// CHECK2-NEXT: to label [[INVOKE_CONT15:%.*]] unwind label [[LPAD10]] -// CHECK2: invoke.cont15: // CHECK2-NEXT: [[TMP2:%.*]] = load ptr, ptr [[DOTADDR]], align 8 // CHECK2-NEXT: ret ptr [[TMP2]] // CHECK2: lpad: @@ -1929,55 +1923,55 @@ int foobar() { // CHECK2-NEXT: store ptr [[TMP4]], ptr [[EXN_SLOT]], align 8 // CHECK2-NEXT: [[TMP5:%.*]] = extractvalue { ptr, i32 } [[TMP3]], 1 // CHECK2-NEXT: store i32 [[TMP5]], ptr [[EHSELECTOR_SLOT]], align 4 -// CHECK2-NEXT: [[TMP6:%.*]] = load ptr, ptr [[ARRAYINIT_ENDOFINIT2]], align 8 -// CHECK2-NEXT: [[ARRAYDESTROY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYINIT_BEGIN1]], [[TMP6]] -// CHECK2-NEXT: br i1 [[ARRAYDESTROY_ISEMPTY]], label [[ARRAYDESTROY_DONE6:%.*]], label [[ARRAYDESTROY_BODY:%.*]] +// CHECK2-NEXT: [[TMP6:%.*]] = load ptr, ptr [[ARRAYINIT_ENDOFINIT1]], align 8 +// CHECK2-NEXT: [[ARRAYDESTROY_ISEMPTY:%.*]] = icmp eq ptr [[TMP1]], [[TMP6]] +// CHECK2-NEXT: br i1 [[ARRAYDESTROY_ISEMPTY]], label [[ARRAYDESTROY_DONE5:%.*]], label [[ARRAYDESTROY_BODY:%.*]] // CHECK2: arraydestroy.body: // CHECK2-NEXT: [[ARRAYDESTROY_ELEMENTPAST:%.*]] = phi ptr [ [[TMP6]], [[LPAD]] ], [ [[ARRAYDESTROY_ELEMENT:%.*]], [[ARRAYDESTROY_BODY]] ] // CHECK2-NEXT: [[ARRAYDESTROY_ELEMENT]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYDESTROY_ELEMENTPAST]], i64 -1 // CHECK2-NEXT: call void @_ZN2S1D1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYDESTROY_ELEMENT]]) #[[ATTR3]] -// CHECK2-NEXT: [[ARRAYDESTROY_DONE:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT]], [[ARRAYINIT_BEGIN1]] -// CHECK2-NEXT: br i1 [[ARRAYDESTROY_DONE]], label [[ARRAYDESTROY_DONE6]], label [[ARRAYDESTROY_BODY]] -// CHECK2: arraydestroy.done6: +// CHECK2-NEXT: [[ARRAYDESTROY_DONE:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT]], [[TMP1]] +// CHECK2-NEXT: br i1 [[ARRAYDESTROY_DONE]], label [[ARRAYDESTROY_DONE5]], label [[ARRAYDESTROY_BODY]] +// CHECK2: arraydestroy.done5: // CHECK2-NEXT: br label [[EHCLEANUP:%.*]] -// CHECK2: lpad10: +// CHECK2: lpad8: // CHECK2-NEXT: [[TMP7:%.*]] = landingpad { ptr, i32 } // CHECK2-NEXT: cleanup // CHECK2-NEXT: [[TMP8:%.*]] = extractvalue { ptr, i32 } [[TMP7]], 0 // CHECK2-NEXT: store ptr [[TMP8]], ptr [[EXN_SLOT]], align 8 // CHECK2-NEXT: [[TMP9:%.*]] = extractvalue { ptr, i32 } [[TMP7]], 1 // CHECK2-NEXT: store i32 [[TMP9]], ptr [[EHSELECTOR_SLOT]], align 4 -// CHECK2-NEXT: [[TMP10:%.*]] = load ptr, ptr [[ARRAYINIT_ENDOFINIT9]], align 8 -// CHECK2-NEXT: [[ARRAYDESTROY_ISEMPTY16:%.*]] = icmp eq ptr [[ARRAYINIT_BEGIN8]], [[TMP10]] -// CHECK2-NEXT: br i1 [[ARRAYDESTROY_ISEMPTY16]], label [[ARRAYDESTROY_DONE21:%.*]], label [[ARRAYDESTROY_BODY17:%.*]] -// CHECK2: arraydestroy.body17: -// CHECK2-NEXT: [[ARRAYDESTROY_ELEMENTPAST18:%.*]] = phi ptr [ [[TMP10]], [[LPAD10]] ], [ [[ARRAYDESTROY_ELEMENT19:%.*]], [[ARRAYDESTROY_BODY17]] ] -// CHECK2-NEXT: [[ARRAYDESTROY_ELEMENT19]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYDESTROY_ELEMENTPAST18]], i64 -1 -// CHECK2-NEXT: call void @_ZN2S1D1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYDESTROY_ELEMENT19]]) #[[ATTR3]] -// CHECK2-NEXT: [[ARRAYDESTROY_DONE20:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT19]], [[ARRAYINIT_BEGIN8]] -// CHECK2-NEXT: br i1 [[ARRAYDESTROY_DONE20]], label [[ARRAYDESTROY_DONE21]], label [[ARRAYDESTROY_BODY17]] -// CHECK2: arraydestroy.done21: +// CHECK2-NEXT: [[TMP10:%.*]] = load ptr, ptr [[ARRAYINIT_ENDOFINIT7]], align 8 +// CHECK2-NEXT: [[ARRAYDESTROY_ISEMPTY14:%.*]] = icmp eq ptr [[ARRAYINIT_ELEMENT6]], [[TMP10]] +// CHECK2-NEXT: br i1 [[ARRAYDESTROY_ISEMPTY14]], label [[ARRAYDESTROY_DONE19:%.*]], label [[ARRAYDESTROY_BODY15:%.*]] +// CHECK2: arraydestroy.body15: +// CHECK2-NEXT: [[ARRAYDESTROY_ELEMENTPAST16:%.*]] = phi ptr [ [[TMP10]], [[LPAD8]] ], [ [[ARRAYDESTROY_ELEMENT17:%.*]], [[ARRAYDESTROY_BODY15]] ] +// CHECK2-NEXT: [[ARRAYDESTROY_ELEMENT17]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYDESTROY_ELEMENTPAST16]], i64 -1 +// CHECK2-NEXT: call void @_ZN2S1D1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYDESTROY_ELEMENT17]]) #[[ATTR3]] +// CHECK2-NEXT: [[ARRAYDESTROY_DONE18:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT17]], [[ARRAYINIT_ELEMENT6]] +// CHECK2-NEXT: br i1 [[ARRAYDESTROY_DONE18]], label [[ARRAYDESTROY_DONE19]], label [[ARRAYDESTROY_BODY15]] +// CHECK2: arraydestroy.done19: // CHECK2-NEXT: br label [[EHCLEANUP]] // CHECK2: ehcleanup: // CHECK2-NEXT: [[TMP11:%.*]] = load ptr, ptr [[ARRAYINIT_ENDOFINIT]], align 8 -// CHECK2-NEXT: [[PAD_ARRAYBEGIN:%.*]] = getelementptr inbounds [3 x %struct.S1], ptr [[ARRAYINIT_BEGIN]], i64 0, i64 0 +// CHECK2-NEXT: [[PAD_ARRAYBEGIN:%.*]] = getelementptr inbounds [3 x %struct.S1], ptr [[TMP1]], i64 0, i64 0 // CHECK2-NEXT: [[PAD_ARRAYEND:%.*]] = getelementptr inbounds [3 x %struct.S1], ptr [[TMP11]], i64 0, i64 0 -// CHECK2-NEXT: [[ARRAYDESTROY_ISEMPTY22:%.*]] = icmp eq ptr [[PAD_ARRAYBEGIN]], [[PAD_ARRAYEND]] -// CHECK2-NEXT: br i1 [[ARRAYDESTROY_ISEMPTY22]], label [[ARRAYDESTROY_DONE27:%.*]], label [[ARRAYDESTROY_BODY23:%.*]] -// CHECK2: arraydestroy.body23: -// CHECK2-NEXT: [[ARRAYDESTROY_ELEMENTPAST24:%.*]] = phi ptr [ [[PAD_ARRAYEND]], [[EHCLEANUP]] ], [ [[ARRAYDESTROY_ELEMENT25:%.*]], [[ARRAYDESTROY_BODY23]] ] -// CHECK2-NEXT: [[ARRAYDESTROY_ELEMENT25]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYDESTROY_ELEMENTPAST24]], i64 -1 -// CHECK2-NEXT: call void @_ZN2S1D1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYDESTROY_ELEMENT25]]) #[[ATTR3]] -// CHECK2-NEXT: [[ARRAYDESTROY_DONE26:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT25]], [[PAD_ARRAYBEGIN]] -// CHECK2-NEXT: br i1 [[ARRAYDESTROY_DONE26]], label [[ARRAYDESTROY_DONE27]], label [[ARRAYDESTROY_BODY23]] -// CHECK2: arraydestroy.done27: +// CHECK2-NEXT: [[ARRAYDESTROY_ISEMPTY20:%.*]] = icmp eq ptr [[PAD_ARRAYBEGIN]], [[PAD_ARRAYEND]] +// CHECK2-NEXT: br i1 [[ARRAYDESTROY_ISEMPTY20]], label [[ARRAYDESTROY_DONE25:%.*]], label [[ARRAYDESTROY_BODY21:%.*]] +// CHECK2: arraydestroy.body21: +// CHECK2-NEXT: [[ARRAYDESTROY_ELEMENTPAST22:%.*]] = phi ptr [ [[PAD_ARRAYEND]], [[EHCLEANUP]] ], [ [[ARRAYDESTROY_ELEMENT23:%.*]], [[ARRAYDESTROY_BODY21]] ] +// CHECK2-NEXT: [[ARRAYDESTROY_ELEMENT23]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYDESTROY_ELEMENTPAST22]], i64 -1 +// CHECK2-NEXT: call void @_ZN2S1D1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYDESTROY_ELEMENT23]]) #[[ATTR3]] +// CHECK2-NEXT: [[ARRAYDESTROY_DONE24:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT23]], [[PAD_ARRAYBEGIN]] +// CHECK2-NEXT: br i1 [[ARRAYDESTROY_DONE24]], label [[ARRAYDESTROY_DONE25]], label [[ARRAYDESTROY_BODY21]] +// CHECK2: arraydestroy.done25: // CHECK2-NEXT: br label [[EH_RESUME:%.*]] // CHECK2: eh.resume: // CHECK2-NEXT: [[EXN:%.*]] = load ptr, ptr [[EXN_SLOT]], align 8 // CHECK2-NEXT: [[SEL:%.*]] = load i32, ptr [[EHSELECTOR_SLOT]], align 4 // CHECK2-NEXT: [[LPAD_VAL:%.*]] = insertvalue { ptr, i32 } poison, ptr [[EXN]], 0 -// CHECK2-NEXT: [[LPAD_VAL28:%.*]] = insertvalue { ptr, i32 } [[LPAD_VAL]], i32 [[SEL]], 1 -// CHECK2-NEXT: resume { ptr, i32 } [[LPAD_VAL28]] +// CHECK2-NEXT: [[LPAD_VAL26:%.*]] = insertvalue { ptr, i32 } [[LPAD_VAL]], i32 [[SEL]], 1 +// CHECK2-NEXT: resume { ptr, i32 } [[LPAD_VAL26]] // // // CHECK2-LABEL: define {{[^@]+}}@.__kmpc_global_dtor_..4 @@ -6505,47 +6499,44 @@ int foobar() { // DEBUG1-NEXT: entry: // DEBUG1-NEXT: [[DOTADDR:%.*]] = alloca ptr, align 8 // DEBUG1-NEXT: [[ARRAYINIT_ENDOFINIT:%.*]] = alloca ptr, align 8 -// DEBUG1-NEXT: [[ARRAYINIT_ENDOFINIT2:%.*]] = alloca ptr, align 8 +// DEBUG1-NEXT: [[ARRAYINIT_ENDOFINIT1:%.*]] = alloca ptr, align 8 // DEBUG1-NEXT: [[EXN_SLOT:%.*]] = alloca ptr, align 8 // DEBUG1-NEXT: [[EHSELECTOR_SLOT:%.*]] = alloca i32, align 4 -// DEBUG1-NEXT: [[ARRAYINIT_ENDOFINIT9:%.*]] = alloca ptr, align 8 +// DEBUG1-NEXT: [[ARRAYINIT_ENDOFINIT7:%.*]] = alloca ptr, align 8 // DEBUG1-NEXT: store ptr [[TMP0]], ptr [[DOTADDR]], align 8 // DEBUG1-NEXT: tail call void @llvm.dbg.declare(metadata ptr [[DOTADDR]], metadata [[META143:![0-9]+]], metadata !DIExpression()), !dbg [[DBG144:![0-9]+]] // DEBUG1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[DOTADDR]], align 8, !dbg [[DBG145:![0-9]+]] -// DEBUG1-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x [3 x %struct.S1]], ptr [[TMP1]], i64 0, i64 0, !dbg [[DBG146:![0-9]+]] -// DEBUG1-NEXT: store ptr [[ARRAYINIT_BEGIN]], ptr [[ARRAYINIT_ENDOFINIT]], align 8, !dbg [[DBG146]] -// DEBUG1-NEXT: [[ARRAYINIT_BEGIN1:%.*]] = getelementptr inbounds [3 x %struct.S1], ptr [[ARRAYINIT_BEGIN]], i64 0, i64 0, !dbg [[DBG147:![0-9]+]] -// DEBUG1-NEXT: store ptr [[ARRAYINIT_BEGIN1]], ptr [[ARRAYINIT_ENDOFINIT2]], align 8, !dbg [[DBG147]] -// DEBUG1-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN1]], i32 noundef 1) +// DEBUG1-NEXT: store ptr [[TMP1]], ptr [[ARRAYINIT_ENDOFINIT]], align 8, !dbg [[DBG146:![0-9]+]] +// DEBUG1-NEXT: store ptr [[TMP1]], ptr [[ARRAYINIT_ENDOFINIT1]], align 8, !dbg [[DBG147:![0-9]+]] +// DEBUG1-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[TMP1]], i32 noundef 1) // DEBUG1-NEXT: to label [[INVOKE_CONT:%.*]] unwind label [[LPAD:%.*]], !dbg [[DBG148:![0-9]+]] // DEBUG1: invoke.cont: -// DEBUG1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S1:%.*]], ptr [[ARRAYINIT_BEGIN1]], i64 1, !dbg [[DBG147]] -// DEBUG1-NEXT: store ptr [[ARRAYINIT_ELEMENT]], ptr [[ARRAYINIT_ENDOFINIT2]], align 8, !dbg [[DBG147]] +// DEBUG1-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S1:%.*]], ptr [[TMP1]], i64 1, !dbg [[DBG147]] +// DEBUG1-NEXT: store ptr [[ARRAYINIT_ELEMENT]], ptr [[ARRAYINIT_ENDOFINIT1]], align 8, !dbg [[DBG147]] // DEBUG1-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) -// DEBUG1-NEXT: to label [[INVOKE_CONT3:%.*]] unwind label [[LPAD]], !dbg [[DBG149:![0-9]+]] -// DEBUG1: invoke.cont3: -// DEBUG1-NEXT: [[ARRAYINIT_ELEMENT4:%.*]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYINIT_ELEMENT]], i64 1, !dbg [[DBG147]] -// DEBUG1-NEXT: store ptr [[ARRAYINIT_ELEMENT4]], ptr [[ARRAYINIT_ENDOFINIT2]], align 8, !dbg [[DBG147]] -// DEBUG1-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT4]], i32 noundef 3) -// DEBUG1-NEXT: to label [[INVOKE_CONT5:%.*]] unwind label [[LPAD]], !dbg [[DBG150:![0-9]+]] -// DEBUG1: invoke.cont5: -// DEBUG1-NEXT: [[ARRAYINIT_ELEMENT7:%.*]] = getelementptr inbounds [3 x %struct.S1], ptr [[ARRAYINIT_BEGIN]], i64 1, !dbg [[DBG146]] -// DEBUG1-NEXT: store ptr [[ARRAYINIT_ELEMENT7]], ptr [[ARRAYINIT_ENDOFINIT]], align 8, !dbg [[DBG146]] -// DEBUG1-NEXT: [[ARRAYINIT_BEGIN8:%.*]] = getelementptr inbounds [3 x %struct.S1], ptr [[ARRAYINIT_ELEMENT7]], i64 0, i64 0, !dbg [[DBG151:![0-9]+]] -// DEBUG1-NEXT: store ptr [[ARRAYINIT_BEGIN8]], ptr [[ARRAYINIT_ENDOFINIT9]], align 8, !dbg [[DBG151]] -// DEBUG1-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN8]], i32 noundef 4) -// DEBUG1-NEXT: to label [[INVOKE_CONT11:%.*]] unwind label [[LPAD10:%.*]], !dbg [[DBG152:![0-9]+]] +// DEBUG1-NEXT: to label [[INVOKE_CONT2:%.*]] unwind label [[LPAD]], !dbg [[DBG149:![0-9]+]] +// DEBUG1: invoke.cont2: +// DEBUG1-NEXT: [[ARRAYINIT_ELEMENT3:%.*]] = getelementptr inbounds [[STRUCT_S1]], ptr [[TMP1]], i64 2, !dbg [[DBG147]] +// DEBUG1-NEXT: store ptr [[ARRAYINIT_ELEMENT3]], ptr [[ARRAYINIT_ENDOFINIT1]], align 8, !dbg [[DBG147]] +// DEBUG1-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT3]], i32 noundef 3) +// DEBUG1-NEXT: to label [[INVOKE_CONT4:%.*]] unwind label [[LPAD]], !dbg [[DBG150:![0-9]+]] +// DEBUG1: invoke.cont4: +// DEBUG1-NEXT: [[ARRAYINIT_ELEMENT6:%.*]] = getelementptr inbounds [3 x %struct.S1], ptr [[TMP1]], i64 1, !dbg [[DBG146]] +// DEBUG1-NEXT: store ptr [[ARRAYINIT_ELEMENT6]], ptr [[ARRAYINIT_ENDOFINIT]], align 8, !dbg [[DBG146]] +// DEBUG1-NEXT: store ptr [[ARRAYINIT_ELEMENT6]], ptr [[ARRAYINIT_ENDOFINIT7]], align 8, !dbg [[DBG151:![0-9]+]] +// DEBUG1-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT6]], i32 noundef 4) +// DEBUG1-NEXT: to label [[INVOKE_CONT9:%.*]] unwind label [[LPAD8:%.*]], !dbg [[DBG152:![0-9]+]] +// DEBUG1: invoke.cont9: +// DEBUG1-NEXT: [[ARRAYINIT_ELEMENT10:%.*]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYINIT_ELEMENT6]], i64 1, !dbg [[DBG151]] +// DEBUG1-NEXT: store ptr [[ARRAYINIT_ELEMENT10]], ptr [[ARRAYINIT_ENDOFINIT7]], align 8, !dbg [[DBG151]] +// DEBUG1-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT10]], i32 noundef 5) +// DEBUG1-NEXT: to label [[INVOKE_CONT11:%.*]] unwind label [[LPAD8]], !dbg [[DBG153:![0-9]+]] // DEBUG1: invoke.cont11: -// DEBUG1-NEXT: [[ARRAYINIT_ELEMENT12:%.*]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYINIT_BEGIN8]], i64 1, !dbg [[DBG151]] -// DEBUG1-NEXT: store ptr [[ARRAYINIT_ELEMENT12]], ptr [[ARRAYINIT_ENDOFINIT9]], align 8, !dbg [[DBG151]] -// DEBUG1-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT12]], i32 noundef 5) -// DEBUG1-NEXT: to label [[INVOKE_CONT13:%.*]] unwind label [[LPAD10]], !dbg [[DBG153:![0-9]+]] +// DEBUG1-NEXT: [[ARRAYINIT_ELEMENT12:%.*]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYINIT_ELEMENT6]], i64 2, !dbg [[DBG151]] +// DEBUG1-NEXT: store ptr [[ARRAYINIT_ELEMENT12]], ptr [[ARRAYINIT_ENDOFINIT7]], align 8, !dbg [[DBG151]] +// DEBUG1-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT12]], i32 noundef 6) +// DEBUG1-NEXT: to label [[INVOKE_CONT13:%.*]] unwind label [[LPAD8]], !dbg [[DBG154:![0-9]+]] // DEBUG1: invoke.cont13: -// DEBUG1-NEXT: [[ARRAYINIT_ELEMENT14:%.*]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYINIT_ELEMENT12]], i64 1, !dbg [[DBG151]] -// DEBUG1-NEXT: store ptr [[ARRAYINIT_ELEMENT14]], ptr [[ARRAYINIT_ENDOFINIT9]], align 8, !dbg [[DBG151]] -// DEBUG1-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT14]], i32 noundef 6) -// DEBUG1-NEXT: to label [[INVOKE_CONT15:%.*]] unwind label [[LPAD10]], !dbg [[DBG154:![0-9]+]] -// DEBUG1: invoke.cont15: // DEBUG1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[DOTADDR]], align 8, !dbg [[DBG145]] // DEBUG1-NEXT: ret ptr [[TMP2]], !dbg [[DBG145]] // DEBUG1: lpad: @@ -6555,55 +6546,55 @@ int foobar() { // DEBUG1-NEXT: store ptr [[TMP4]], ptr [[EXN_SLOT]], align 8, !dbg [[DBG144]] // DEBUG1-NEXT: [[TMP5:%.*]] = extractvalue { ptr, i32 } [[TMP3]], 1, !dbg [[DBG144]] // DEBUG1-NEXT: store i32 [[TMP5]], ptr [[EHSELECTOR_SLOT]], align 4, !dbg [[DBG144]] -// DEBUG1-NEXT: [[TMP6:%.*]] = load ptr, ptr [[ARRAYINIT_ENDOFINIT2]], align 8, !dbg [[DBG147]] -// DEBUG1-NEXT: [[ARRAYDESTROY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYINIT_BEGIN1]], [[TMP6]], !dbg [[DBG147]] -// DEBUG1-NEXT: br i1 [[ARRAYDESTROY_ISEMPTY]], label [[ARRAYDESTROY_DONE6:%.*]], label [[ARRAYDESTROY_BODY:%.*]], !dbg [[DBG147]] +// DEBUG1-NEXT: [[TMP6:%.*]] = load ptr, ptr [[ARRAYINIT_ENDOFINIT1]], align 8, !dbg [[DBG147]] +// DEBUG1-NEXT: [[ARRAYDESTROY_ISEMPTY:%.*]] = icmp eq ptr [[TMP1]], [[TMP6]], !dbg [[DBG147]] +// DEBUG1-NEXT: br i1 [[ARRAYDESTROY_ISEMPTY]], label [[ARRAYDESTROY_DONE5:%.*]], label [[ARRAYDESTROY_BODY:%.*]], !dbg [[DBG147]] // DEBUG1: arraydestroy.body: // DEBUG1-NEXT: [[ARRAYDESTROY_ELEMENTPAST:%.*]] = phi ptr [ [[TMP6]], [[LPAD]] ], [ [[ARRAYDESTROY_ELEMENT:%.*]], [[ARRAYDESTROY_BODY]] ], !dbg [[DBG147]] // DEBUG1-NEXT: [[ARRAYDESTROY_ELEMENT]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYDESTROY_ELEMENTPAST]], i64 -1, !dbg [[DBG147]] // DEBUG1-NEXT: call void @_ZN2S1D1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYDESTROY_ELEMENT]]) #[[ATTR4]], !dbg [[DBG147]] -// DEBUG1-NEXT: [[ARRAYDESTROY_DONE:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT]], [[ARRAYINIT_BEGIN1]], !dbg [[DBG147]] -// DEBUG1-NEXT: br i1 [[ARRAYDESTROY_DONE]], label [[ARRAYDESTROY_DONE6]], label [[ARRAYDESTROY_BODY]], !dbg [[DBG147]] -// DEBUG1: arraydestroy.done6: +// DEBUG1-NEXT: [[ARRAYDESTROY_DONE:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT]], [[TMP1]], !dbg [[DBG147]] +// DEBUG1-NEXT: br i1 [[ARRAYDESTROY_DONE]], label [[ARRAYDESTROY_DONE5]], label [[ARRAYDESTROY_BODY]], !dbg [[DBG147]] +// DEBUG1: arraydestroy.done5: // DEBUG1-NEXT: br label [[EHCLEANUP:%.*]], !dbg [[DBG147]] -// DEBUG1: lpad10: +// DEBUG1: lpad8: // DEBUG1-NEXT: [[TMP7:%.*]] = landingpad { ptr, i32 } // DEBUG1-NEXT: cleanup, !dbg [[DBG144]] // DEBUG1-NEXT: [[TMP8:%.*]] = extractvalue { ptr, i32 } [[TMP7]], 0, !dbg [[DBG144]] // DEBUG1-NEXT: store ptr [[TMP8]], ptr [[EXN_SLOT]], align 8, !dbg [[DBG144]] // DEBUG1-NEXT: [[TMP9:%.*]] = extractvalue { ptr, i32 } [[TMP7]], 1, !dbg [[DBG144]] // DEBUG1-NEXT: store i32 [[TMP9]], ptr [[EHSELECTOR_SLOT]], align 4, !dbg [[DBG144]] -// DEBUG1-NEXT: [[TMP10:%.*]] = load ptr, ptr [[ARRAYINIT_ENDOFINIT9]], align 8, !dbg [[DBG151]] -// DEBUG1-NEXT: [[ARRAYDESTROY_ISEMPTY16:%.*]] = icmp eq ptr [[ARRAYINIT_BEGIN8]], [[TMP10]], !dbg [[DBG151]] -// DEBUG1-NEXT: br i1 [[ARRAYDESTROY_ISEMPTY16]], label [[ARRAYDESTROY_DONE21:%.*]], label [[ARRAYDESTROY_BODY17:%.*]], !dbg [[DBG151]] -// DEBUG1: arraydestroy.body17: -// DEBUG1-NEXT: [[ARRAYDESTROY_ELEMENTPAST18:%.*]] = phi ptr [ [[TMP10]], [[LPAD10]] ], [ [[ARRAYDESTROY_ELEMENT19:%.*]], [[ARRAYDESTROY_BODY17]] ], !dbg [[DBG151]] -// DEBUG1-NEXT: [[ARRAYDESTROY_ELEMENT19]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYDESTROY_ELEMENTPAST18]], i64 -1, !dbg [[DBG151]] -// DEBUG1-NEXT: call void @_ZN2S1D1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYDESTROY_ELEMENT19]]) #[[ATTR4]], !dbg [[DBG151]] -// DEBUG1-NEXT: [[ARRAYDESTROY_DONE20:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT19]], [[ARRAYINIT_BEGIN8]], !dbg [[DBG151]] -// DEBUG1-NEXT: br i1 [[ARRAYDESTROY_DONE20]], label [[ARRAYDESTROY_DONE21]], label [[ARRAYDESTROY_BODY17]], !dbg [[DBG151]] -// DEBUG1: arraydestroy.done21: +// DEBUG1-NEXT: [[TMP10:%.*]] = load ptr, ptr [[ARRAYINIT_ENDOFINIT7]], align 8, !dbg [[DBG151]] +// DEBUG1-NEXT: [[ARRAYDESTROY_ISEMPTY14:%.*]] = icmp eq ptr [[ARRAYINIT_ELEMENT6]], [[TMP10]], !dbg [[DBG151]] +// DEBUG1-NEXT: br i1 [[ARRAYDESTROY_ISEMPTY14]], label [[ARRAYDESTROY_DONE19:%.*]], label [[ARRAYDESTROY_BODY15:%.*]], !dbg [[DBG151]] +// DEBUG1: arraydestroy.body15: +// DEBUG1-NEXT: [[ARRAYDESTROY_ELEMENTPAST16:%.*]] = phi ptr [ [[TMP10]], [[LPAD8]] ], [ [[ARRAYDESTROY_ELEMENT17:%.*]], [[ARRAYDESTROY_BODY15]] ], !dbg [[DBG151]] +// DEBUG1-NEXT: [[ARRAYDESTROY_ELEMENT17]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYDESTROY_ELEMENTPAST16]], i64 -1, !dbg [[DBG151]] +// DEBUG1-NEXT: call void @_ZN2S1D1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYDESTROY_ELEMENT17]]) #[[ATTR4]], !dbg [[DBG151]] +// DEBUG1-NEXT: [[ARRAYDESTROY_DONE18:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT17]], [[ARRAYINIT_ELEMENT6]], !dbg [[DBG151]] +// DEBUG1-NEXT: br i1 [[ARRAYDESTROY_DONE18]], label [[ARRAYDESTROY_DONE19]], label [[ARRAYDESTROY_BODY15]], !dbg [[DBG151]] +// DEBUG1: arraydestroy.done19: // DEBUG1-NEXT: br label [[EHCLEANUP]], !dbg [[DBG151]] // DEBUG1: ehcleanup: // DEBUG1-NEXT: [[TMP11:%.*]] = load ptr, ptr [[ARRAYINIT_ENDOFINIT]], align 8, !dbg [[DBG146]] -// DEBUG1-NEXT: [[PAD_ARRAYBEGIN:%.*]] = getelementptr inbounds [3 x %struct.S1], ptr [[ARRAYINIT_BEGIN]], i64 0, i64 0, !dbg [[DBG146]] +// DEBUG1-NEXT: [[PAD_ARRAYBEGIN:%.*]] = getelementptr inbounds [3 x %struct.S1], ptr [[TMP1]], i64 0, i64 0, !dbg [[DBG146]] // DEBUG1-NEXT: [[PAD_ARRAYEND:%.*]] = getelementptr inbounds [3 x %struct.S1], ptr [[TMP11]], i64 0, i64 0, !dbg [[DBG146]] -// DEBUG1-NEXT: [[ARRAYDESTROY_ISEMPTY22:%.*]] = icmp eq ptr [[PAD_ARRAYBEGIN]], [[PAD_ARRAYEND]], !dbg [[DBG146]] -// DEBUG1-NEXT: br i1 [[ARRAYDESTROY_ISEMPTY22]], label [[ARRAYDESTROY_DONE27:%.*]], label [[ARRAYDESTROY_BODY23:%.*]], !dbg [[DBG146]] -// DEBUG1: arraydestroy.body23: -// DEBUG1-NEXT: [[ARRAYDESTROY_ELEMENTPAST24:%.*]] = phi ptr [ [[PAD_ARRAYEND]], [[EHCLEANUP]] ], [ [[ARRAYDESTROY_ELEMENT25:%.*]], [[ARRAYDESTROY_BODY23]] ], !dbg [[DBG146]] -// DEBUG1-NEXT: [[ARRAYDESTROY_ELEMENT25]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYDESTROY_ELEMENTPAST24]], i64 -1, !dbg [[DBG146]] -// DEBUG1-NEXT: call void @_ZN2S1D1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYDESTROY_ELEMENT25]]) #[[ATTR4]], !dbg [[DBG146]] -// DEBUG1-NEXT: [[ARRAYDESTROY_DONE26:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT25]], [[PAD_ARRAYBEGIN]], !dbg [[DBG146]] -// DEBUG1-NEXT: br i1 [[ARRAYDESTROY_DONE26]], label [[ARRAYDESTROY_DONE27]], label [[ARRAYDESTROY_BODY23]], !dbg [[DBG146]] -// DEBUG1: arraydestroy.done27: +// DEBUG1-NEXT: [[ARRAYDESTROY_ISEMPTY20:%.*]] = icmp eq ptr [[PAD_ARRAYBEGIN]], [[PAD_ARRAYEND]], !dbg [[DBG146]] +// DEBUG1-NEXT: br i1 [[ARRAYDESTROY_ISEMPTY20]], label [[ARRAYDESTROY_DONE25:%.*]], label [[ARRAYDESTROY_BODY21:%.*]], !dbg [[DBG146]] +// DEBUG1: arraydestroy.body21: +// DEBUG1-NEXT: [[ARRAYDESTROY_ELEMENTPAST22:%.*]] = phi ptr [ [[PAD_ARRAYEND]], [[EHCLEANUP]] ], [ [[ARRAYDESTROY_ELEMENT23:%.*]], [[ARRAYDESTROY_BODY21]] ], !dbg [[DBG146]] +// DEBUG1-NEXT: [[ARRAYDESTROY_ELEMENT23]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYDESTROY_ELEMENTPAST22]], i64 -1, !dbg [[DBG146]] +// DEBUG1-NEXT: call void @_ZN2S1D1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYDESTROY_ELEMENT23]]) #[[ATTR4]], !dbg [[DBG146]] +// DEBUG1-NEXT: [[ARRAYDESTROY_DONE24:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT23]], [[PAD_ARRAYBEGIN]], !dbg [[DBG146]] +// DEBUG1-NEXT: br i1 [[ARRAYDESTROY_DONE24]], label [[ARRAYDESTROY_DONE25]], label [[ARRAYDESTROY_BODY21]], !dbg [[DBG146]] +// DEBUG1: arraydestroy.done25: // DEBUG1-NEXT: br label [[EH_RESUME:%.*]], !dbg [[DBG146]] // DEBUG1: eh.resume: // DEBUG1-NEXT: [[EXN:%.*]] = load ptr, ptr [[EXN_SLOT]], align 8, !dbg [[DBG146]] // DEBUG1-NEXT: [[SEL:%.*]] = load i32, ptr [[EHSELECTOR_SLOT]], align 4, !dbg [[DBG146]] // DEBUG1-NEXT: [[LPAD_VAL:%.*]] = insertvalue { ptr, i32 } poison, ptr [[EXN]], 0, !dbg [[DBG146]] -// DEBUG1-NEXT: [[LPAD_VAL28:%.*]] = insertvalue { ptr, i32 } [[LPAD_VAL]], i32 [[SEL]], 1, !dbg [[DBG146]] -// DEBUG1-NEXT: resume { ptr, i32 } [[LPAD_VAL28]], !dbg [[DBG146]] +// DEBUG1-NEXT: [[LPAD_VAL26:%.*]] = insertvalue { ptr, i32 } [[LPAD_VAL]], i32 [[SEL]], 1, !dbg [[DBG146]] +// DEBUG1-NEXT: resume { ptr, i32 } [[LPAD_VAL26]], !dbg [[DBG146]] // // // DEBUG1-LABEL: define {{[^@]+}}@.__kmpc_global_dtor_..2 @@ -7375,47 +7366,44 @@ int foobar() { // DEBUG2-NEXT: entry: // DEBUG2-NEXT: [[DOTADDR:%.*]] = alloca ptr, align 8 // DEBUG2-NEXT: [[ARRAYINIT_ENDOFINIT:%.*]] = alloca ptr, align 8 -// DEBUG2-NEXT: [[ARRAYINIT_ENDOFINIT2:%.*]] = alloca ptr, align 8 +// DEBUG2-NEXT: [[ARRAYINIT_ENDOFINIT1:%.*]] = alloca ptr, align 8 // DEBUG2-NEXT: [[EXN_SLOT:%.*]] = alloca ptr, align 8 // DEBUG2-NEXT: [[EHSELECTOR_SLOT:%.*]] = alloca i32, align 4 -// DEBUG2-NEXT: [[ARRAYINIT_ENDOFINIT9:%.*]] = alloca ptr, align 8 +// DEBUG2-NEXT: [[ARRAYINIT_ENDOFINIT7:%.*]] = alloca ptr, align 8 // DEBUG2-NEXT: store ptr [[TMP0]], ptr [[DOTADDR]], align 8 // DEBUG2-NEXT: tail call void @llvm.dbg.declare(metadata ptr [[DOTADDR]], metadata [[META179:![0-9]+]], metadata !DIExpression()), !dbg [[DBG180:![0-9]+]] // DEBUG2-NEXT: [[TMP1:%.*]] = load ptr, ptr [[DOTADDR]], align 8, !dbg [[DBG181:![0-9]+]] -// DEBUG2-NEXT: [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x [3 x %struct.S1]], ptr [[TMP1]], i64 0, i64 0, !dbg [[DBG182:![0-9]+]] -// DEBUG2-NEXT: store ptr [[ARRAYINIT_BEGIN]], ptr [[ARRAYINIT_ENDOFINIT]], align 8, !dbg [[DBG182]] -// DEBUG2-NEXT: [[ARRAYINIT_BEGIN1:%.*]] = getelementptr inbounds [3 x %struct.S1], ptr [[ARRAYINIT_BEGIN]], i64 0, i64 0, !dbg [[DBG183:![0-9]+]] -// DEBUG2-NEXT: store ptr [[ARRAYINIT_BEGIN1]], ptr [[ARRAYINIT_ENDOFINIT2]], align 8, !dbg [[DBG183]] -// DEBUG2-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN1]], i32 noundef 1) +// DEBUG2-NEXT: store ptr [[TMP1]], ptr [[ARRAYINIT_ENDOFINIT]], align 8, !dbg [[DBG182:![0-9]+]] +// DEBUG2-NEXT: store ptr [[TMP1]], ptr [[ARRAYINIT_ENDOFINIT1]], align 8, !dbg [[DBG183:![0-9]+]] +// DEBUG2-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[TMP1]], i32 noundef 1) // DEBUG2-NEXT: to label [[INVOKE_CONT:%.*]] unwind label [[LPAD:%.*]], !dbg [[DBG184:![0-9]+]] // DEBUG2: invoke.cont: -// DEBUG2-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S1:%.*]], ptr [[ARRAYINIT_BEGIN1]], i64 1, !dbg [[DBG183]] -// DEBUG2-NEXT: store ptr [[ARRAYINIT_ELEMENT]], ptr [[ARRAYINIT_ENDOFINIT2]], align 8, !dbg [[DBG183]] +// DEBUG2-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S1:%.*]], ptr [[TMP1]], i64 1, !dbg [[DBG183]] +// DEBUG2-NEXT: store ptr [[ARRAYINIT_ELEMENT]], ptr [[ARRAYINIT_ENDOFINIT1]], align 8, !dbg [[DBG183]] // DEBUG2-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 noundef 2) -// DEBUG2-NEXT: to label [[INVOKE_CONT3:%.*]] unwind label [[LPAD]], !dbg [[DBG185:![0-9]+]] -// DEBUG2: invoke.cont3: -// DEBUG2-NEXT: [[ARRAYINIT_ELEMENT4:%.*]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYINIT_ELEMENT]], i64 1, !dbg [[DBG183]] -// DEBUG2-NEXT: store ptr [[ARRAYINIT_ELEMENT4]], ptr [[ARRAYINIT_ENDOFINIT2]], align 8, !dbg [[DBG183]] -// DEBUG2-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT4]], i32 noundef 3) -// DEBUG2-NEXT: to label [[INVOKE_CONT5:%.*]] unwind label [[LPAD]], !dbg [[DBG186:![0-9]+]] -// DEBUG2: invoke.cont5: -// DEBUG2-NEXT: [[ARRAYINIT_ELEMENT7:%.*]] = getelementptr inbounds [3 x %struct.S1], ptr [[ARRAYINIT_BEGIN]], i64 1, !dbg [[DBG182]] -// DEBUG2-NEXT: store ptr [[ARRAYINIT_ELEMENT7]], ptr [[ARRAYINIT_ENDOFINIT]], align 8, !dbg [[DBG182]] -// DEBUG2-NEXT: [[ARRAYINIT_BEGIN8:%.*]] = getelementptr inbounds [3 x %struct.S1], ptr [[ARRAYINIT_ELEMENT7]], i64 0, i64 0, !dbg [[DBG187:![0-9]+]] -// DEBUG2-NEXT: store ptr [[ARRAYINIT_BEGIN8]], ptr [[ARRAYINIT_ENDOFINIT9]], align 8, !dbg [[DBG187]] -// DEBUG2-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN8]], i32 noundef 4) -// DEBUG2-NEXT: to label [[INVOKE_CONT11:%.*]] unwind label [[LPAD10:%.*]], !dbg [[DBG188:![0-9]+]] +// DEBUG2-NEXT: to label [[INVOKE_CONT2:%.*]] unwind label [[LPAD]], !dbg [[DBG185:![0-9]+]] +// DEBUG2: invoke.cont2: +// DEBUG2-NEXT: [[ARRAYINIT_ELEMENT3:%.*]] = getelementptr inbounds [[STRUCT_S1]], ptr [[TMP1]], i64 2, !dbg [[DBG183]] +// DEBUG2-NEXT: store ptr [[ARRAYINIT_ELEMENT3]], ptr [[ARRAYINIT_ENDOFINIT1]], align 8, !dbg [[DBG183]] +// DEBUG2-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT3]], i32 noundef 3) +// DEBUG2-NEXT: to label [[INVOKE_CONT4:%.*]] unwind label [[LPAD]], !dbg [[DBG186:![0-9]+]] +// DEBUG2: invoke.cont4: +// DEBUG2-NEXT: [[ARRAYINIT_ELEMENT6:%.*]] = getelementptr inbounds [3 x %struct.S1], ptr [[TMP1]], i64 1, !dbg [[DBG182]] +// DEBUG2-NEXT: store ptr [[ARRAYINIT_ELEMENT6]], ptr [[ARRAYINIT_ENDOFINIT]], align 8, !dbg [[DBG182]] +// DEBUG2-NEXT: store ptr [[ARRAYINIT_ELEMENT6]], ptr [[ARRAYINIT_ENDOFINIT7]], align 8, !dbg [[DBG187:![0-9]+]] +// DEBUG2-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT6]], i32 noundef 4) +// DEBUG2-NEXT: to label [[INVOKE_CONT9:%.*]] unwind label [[LPAD8:%.*]], !dbg [[DBG188:![0-9]+]] +// DEBUG2: invoke.cont9: +// DEBUG2-NEXT: [[ARRAYINIT_ELEMENT10:%.*]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYINIT_ELEMENT6]], i64 1, !dbg [[DBG187]] +// DEBUG2-NEXT: store ptr [[ARRAYINIT_ELEMENT10]], ptr [[ARRAYINIT_ENDOFINIT7]], align 8, !dbg [[DBG187]] +// DEBUG2-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT10]], i32 noundef 5) +// DEBUG2-NEXT: to label [[INVOKE_CONT11:%.*]] unwind label [[LPAD8]], !dbg [[DBG189:![0-9]+]] // DEBUG2: invoke.cont11: -// DEBUG2-NEXT: [[ARRAYINIT_ELEMENT12:%.*]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYINIT_BEGIN8]], i64 1, !dbg [[DBG187]] -// DEBUG2-NEXT: store ptr [[ARRAYINIT_ELEMENT12]], ptr [[ARRAYINIT_ENDOFINIT9]], align 8, !dbg [[DBG187]] -// DEBUG2-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT12]], i32 noundef 5) -// DEBUG2-NEXT: to label [[INVOKE_CONT13:%.*]] unwind label [[LPAD10]], !dbg [[DBG189:![0-9]+]] +// DEBUG2-NEXT: [[ARRAYINIT_ELEMENT12:%.*]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYINIT_ELEMENT6]], i64 2, !dbg [[DBG187]] +// DEBUG2-NEXT: store ptr [[ARRAYINIT_ELEMENT12]], ptr [[ARRAYINIT_ENDOFINIT7]], align 8, !dbg [[DBG187]] +// DEBUG2-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT12]], i32 noundef 6) +// DEBUG2-NEXT: to label [[INVOKE_CONT13:%.*]] unwind label [[LPAD8]], !dbg [[DBG190:![0-9]+]] // DEBUG2: invoke.cont13: -// DEBUG2-NEXT: [[ARRAYINIT_ELEMENT14:%.*]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYINIT_ELEMENT12]], i64 1, !dbg [[DBG187]] -// DEBUG2-NEXT: store ptr [[ARRAYINIT_ELEMENT14]], ptr [[ARRAYINIT_ENDOFINIT9]], align 8, !dbg [[DBG187]] -// DEBUG2-NEXT: invoke void @_ZN2S1C1Ei(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT14]], i32 noundef 6) -// DEBUG2-NEXT: to label [[INVOKE_CONT15:%.*]] unwind label [[LPAD10]], !dbg [[DBG190:![0-9]+]] -// DEBUG2: invoke.cont15: // DEBUG2-NEXT: [[TMP2:%.*]] = load ptr, ptr [[DOTADDR]], align 8, !dbg [[DBG181]] // DEBUG2-NEXT: ret ptr [[TMP2]], !dbg [[DBG181]] // DEBUG2: lpad: @@ -7425,55 +7413,55 @@ int foobar() { // DEBUG2-NEXT: store ptr [[TMP4]], ptr [[EXN_SLOT]], align 8, !dbg [[DBG180]] // DEBUG2-NEXT: [[TMP5:%.*]] = extractvalue { ptr, i32 } [[TMP3]], 1, !dbg [[DBG180]] // DEBUG2-NEXT: store i32 [[TMP5]], ptr [[EHSELECTOR_SLOT]], align 4, !dbg [[DBG180]] -// DEBUG2-NEXT: [[TMP6:%.*]] = load ptr, ptr [[ARRAYINIT_ENDOFINIT2]], align 8, !dbg [[DBG183]] -// DEBUG2-NEXT: [[ARRAYDESTROY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYINIT_BEGIN1]], [[TMP6]], !dbg [[DBG183]] -// DEBUG2-NEXT: br i1 [[ARRAYDESTROY_ISEMPTY]], label [[ARRAYDESTROY_DONE6:%.*]], label [[ARRAYDESTROY_BODY:%.*]], !dbg [[DBG183]] +// DEBUG2-NEXT: [[TMP6:%.*]] = load ptr, ptr [[ARRAYINIT_ENDOFINIT1]], align 8, !dbg [[DBG183]] +// DEBUG2-NEXT: [[ARRAYDESTROY_ISEMPTY:%.*]] = icmp eq ptr [[TMP1]], [[TMP6]], !dbg [[DBG183]] +// DEBUG2-NEXT: br i1 [[ARRAYDESTROY_ISEMPTY]], label [[ARRAYDESTROY_DONE5:%.*]], label [[ARRAYDESTROY_BODY:%.*]], !dbg [[DBG183]] // DEBUG2: arraydestroy.body: // DEBUG2-NEXT: [[ARRAYDESTROY_ELEMENTPAST:%.*]] = phi ptr [ [[TMP6]], [[LPAD]] ], [ [[ARRAYDESTROY_ELEMENT:%.*]], [[ARRAYDESTROY_BODY]] ], !dbg [[DBG183]] // DEBUG2-NEXT: [[ARRAYDESTROY_ELEMENT]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYDESTROY_ELEMENTPAST]], i64 -1, !dbg [[DBG183]] // DEBUG2-NEXT: call void @_ZN2S1D1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYDESTROY_ELEMENT]]) #[[ATTR4]], !dbg [[DBG183]] -// DEBUG2-NEXT: [[ARRAYDESTROY_DONE:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT]], [[ARRAYINIT_BEGIN1]], !dbg [[DBG183]] -// DEBUG2-NEXT: br i1 [[ARRAYDESTROY_DONE]], label [[ARRAYDESTROY_DONE6]], label [[ARRAYDESTROY_BODY]], !dbg [[DBG183]] -// DEBUG2: arraydestroy.done6: +// DEBUG2-NEXT: [[ARRAYDESTROY_DONE:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT]], [[TMP1]], !dbg [[DBG183]] +// DEBUG2-NEXT: br i1 [[ARRAYDESTROY_DONE]], label [[ARRAYDESTROY_DONE5]], label [[ARRAYDESTROY_BODY]], !dbg [[DBG183]] +// DEBUG2: arraydestroy.done5: // DEBUG2-NEXT: br label [[EHCLEANUP:%.*]], !dbg [[DBG183]] -// DEBUG2: lpad10: +// DEBUG2: lpad8: // DEBUG2-NEXT: [[TMP7:%.*]] = landingpad { ptr, i32 } // DEBUG2-NEXT: cleanup, !dbg [[DBG180]] // DEBUG2-NEXT: [[TMP8:%.*]] = extractvalue { ptr, i32 } [[TMP7]], 0, !dbg [[DBG180]] // DEBUG2-NEXT: store ptr [[TMP8]], ptr [[EXN_SLOT]], align 8, !dbg [[DBG180]] // DEBUG2-NEXT: [[TMP9:%.*]] = extractvalue { ptr, i32 } [[TMP7]], 1, !dbg [[DBG180]] // DEBUG2-NEXT: store i32 [[TMP9]], ptr [[EHSELECTOR_SLOT]], align 4, !dbg [[DBG180]] -// DEBUG2-NEXT: [[TMP10:%.*]] = load ptr, ptr [[ARRAYINIT_ENDOFINIT9]], align 8, !dbg [[DBG187]] -// DEBUG2-NEXT: [[ARRAYDESTROY_ISEMPTY16:%.*]] = icmp eq ptr [[ARRAYINIT_BEGIN8]], [[TMP10]], !dbg [[DBG187]] -// DEBUG2-NEXT: br i1 [[ARRAYDESTROY_ISEMPTY16]], label [[ARRAYDESTROY_DONE21:%.*]], label [[ARRAYDESTROY_BODY17:%.*]], !dbg [[DBG187]] -// DEBUG2: arraydestroy.body17: -// DEBUG2-NEXT: [[ARRAYDESTROY_ELEMENTPAST18:%.*]] = phi ptr [ [[TMP10]], [[LPAD10]] ], [ [[ARRAYDESTROY_ELEMENT19:%.*]], [[ARRAYDESTROY_BODY17]] ], !dbg [[DBG187]] -// DEBUG2-NEXT: [[ARRAYDESTROY_ELEMENT19]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYDESTROY_ELEMENTPAST18]], i64 -1, !dbg [[DBG187]] -// DEBUG2-NEXT: call void @_ZN2S1D1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYDESTROY_ELEMENT19]]) #[[ATTR4]], !dbg [[DBG187]] -// DEBUG2-NEXT: [[ARRAYDESTROY_DONE20:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT19]], [[ARRAYINIT_BEGIN8]], !dbg [[DBG187]] -// DEBUG2-NEXT: br i1 [[ARRAYDESTROY_DONE20]], label [[ARRAYDESTROY_DONE21]], label [[ARRAYDESTROY_BODY17]], !dbg [[DBG187]] -// DEBUG2: arraydestroy.done21: +// DEBUG2-NEXT: [[TMP10:%.*]] = load ptr, ptr [[ARRAYINIT_ENDOFINIT7]], align 8, !dbg [[DBG187]] +// DEBUG2-NEXT: [[ARRAYDESTROY_ISEMPTY14:%.*]] = icmp eq ptr [[ARRAYINIT_ELEMENT6]], [[TMP10]], !dbg [[DBG187]] +// DEBUG2-NEXT: br i1 [[ARRAYDESTROY_ISEMPTY14]], label [[ARRAYDESTROY_DONE19:%.*]], label [[ARRAYDESTROY_BODY15:%.*]], !dbg [[DBG187]] +// DEBUG2: arraydestroy.body15: +// DEBUG2-NEXT: [[ARRAYDESTROY_ELEMENTPAST16:%.*]] = phi ptr [ [[TMP10]], [[LPAD8]] ], [ [[ARRAYDESTROY_ELEMENT17:%.*]], [[ARRAYDESTROY_BODY15]] ], !dbg [[DBG187]] +// DEBUG2-NEXT: [[ARRAYDESTROY_ELEMENT17]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYDESTROY_ELEMENTPAST16]], i64 -1, !dbg [[DBG187]] +// DEBUG2-NEXT: call void @_ZN2S1D1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYDESTROY_ELEMENT17]]) #[[ATTR4]], !dbg [[DBG187]] +// DEBUG2-NEXT: [[ARRAYDESTROY_DONE18:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT17]], [[ARRAYINIT_ELEMENT6]], !dbg [[DBG187]] +// DEBUG2-NEXT: br i1 [[ARRAYDESTROY_DONE18]], label [[ARRAYDESTROY_DONE19]], label [[ARRAYDESTROY_BODY15]], !dbg [[DBG187]] +// DEBUG2: arraydestroy.done19: // DEBUG2-NEXT: br label [[EHCLEANUP]], !dbg [[DBG187]] // DEBUG2: ehcleanup: // DEBUG2-NEXT: [[TMP11:%.*]] = load ptr, ptr [[ARRAYINIT_ENDOFINIT]], align 8, !dbg [[DBG182]] -// DEBUG2-NEXT: [[PAD_ARRAYBEGIN:%.*]] = getelementptr inbounds [3 x %struct.S1], ptr [[ARRAYINIT_BEGIN]], i64 0, i64 0, !dbg [[DBG182]] +// DEBUG2-NEXT: [[PAD_ARRAYBEGIN:%.*]] = getelementptr inbounds [3 x %struct.S1], ptr [[TMP1]], i64 0, i64 0, !dbg [[DBG182]] // DEBUG2-NEXT: [[PAD_ARRAYEND:%.*]] = getelementptr inbounds [3 x %struct.S1], ptr [[TMP11]], i64 0, i64 0, !dbg [[DBG182]] -// DEBUG2-NEXT: [[ARRAYDESTROY_ISEMPTY22:%.*]] = icmp eq ptr [[PAD_ARRAYBEGIN]], [[PAD_ARRAYEND]], !dbg [[DBG182]] -// DEBUG2-NEXT: br i1 [[ARRAYDESTROY_ISEMPTY22]], label [[ARRAYDESTROY_DONE27:%.*]], label [[ARRAYDESTROY_BODY23:%.*]], !dbg [[DBG182]] -// DEBUG2: arraydestroy.body23: -// DEBUG2-NEXT: [[ARRAYDESTROY_ELEMENTPAST24:%.*]] = phi ptr [ [[PAD_ARRAYEND]], [[EHCLEANUP]] ], [ [[ARRAYDESTROY_ELEMENT25:%.*]], [[ARRAYDESTROY_BODY23]] ], !dbg [[DBG182]] -// DEBUG2-NEXT: [[ARRAYDESTROY_ELEMENT25]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYDESTROY_ELEMENTPAST24]], i64 -1, !dbg [[DBG182]] -// DEBUG2-NEXT: call void @_ZN2S1D1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYDESTROY_ELEMENT25]]) #[[ATTR4]], !dbg [[DBG182]] -// DEBUG2-NEXT: [[ARRAYDESTROY_DONE26:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT25]], [[PAD_ARRAYBEGIN]], !dbg [[DBG182]] -// DEBUG2-NEXT: br i1 [[ARRAYDESTROY_DONE26]], label [[ARRAYDESTROY_DONE27]], label [[ARRAYDESTROY_BODY23]], !dbg [[DBG182]] -// DEBUG2: arraydestroy.done27: +// DEBUG2-NEXT: [[ARRAYDESTROY_ISEMPTY20:%.*]] = icmp eq ptr [[PAD_ARRAYBEGIN]], [[PAD_ARRAYEND]], !dbg [[DBG182]] +// DEBUG2-NEXT: br i1 [[ARRAYDESTROY_ISEMPTY20]], label [[ARRAYDESTROY_DONE25:%.*]], label [[ARRAYDESTROY_BODY21:%.*]], !dbg [[DBG182]] +// DEBUG2: arraydestroy.body21: +// DEBUG2-NEXT: [[ARRAYDESTROY_ELEMENTPAST22:%.*]] = phi ptr [ [[PAD_ARRAYEND]], [[EHCLEANUP]] ], [ [[ARRAYDESTROY_ELEMENT23:%.*]], [[ARRAYDESTROY_BODY21]] ], !dbg [[DBG182]] +// DEBUG2-NEXT: [[ARRAYDESTROY_ELEMENT23]] = getelementptr inbounds [[STRUCT_S1]], ptr [[ARRAYDESTROY_ELEMENTPAST22]], i64 -1, !dbg [[DBG182]] +// DEBUG2-NEXT: call void @_ZN2S1D1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[ARRAYDESTROY_ELEMENT23]]) #[[ATTR4]], !dbg [[DBG182]] +// DEBUG2-NEXT: [[ARRAYDESTROY_DONE24:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT23]], [[PAD_ARRAYBEGIN]], !dbg [[DBG182]] +// DEBUG2-NEXT: br i1 [[ARRAYDESTROY_DONE24]], label [[ARRAYDESTROY_DONE25]], label [[ARRAYDESTROY_BODY21]], !dbg [[DBG182]] +// DEBUG2: arraydestroy.done25: // DEBUG2-NEXT: br label [[EH_RESUME:%.*]], !dbg [[DBG182]] // DEBUG2: eh.resume: // DEBUG2-NEXT: [[EXN:%.*]] = load ptr, ptr [[EXN_SLOT]], align 8, !dbg [[DBG182]] // DEBUG2-NEXT: [[SEL:%.*]] = load i32, ptr [[EHSELECTOR_SLOT]], align 4, !dbg [[DBG182]] // DEBUG2-NEXT: [[LPAD_VAL:%.*]] = insertvalue { ptr, i32 } poison, ptr [[EXN]], 0, !dbg [[DBG182]] -// DEBUG2-NEXT: [[LPAD_VAL28:%.*]] = insertvalue { ptr, i32 } [[LPAD_VAL]], i32 [[SEL]], 1, !dbg [[DBG182]] -// DEBUG2-NEXT: resume { ptr, i32 } [[LPAD_VAL28]], !dbg [[DBG182]] +// DEBUG2-NEXT: [[LPAD_VAL26:%.*]] = insertvalue { ptr, i32 } [[LPAD_VAL]], i32 [[SEL]], 1, !dbg [[DBG182]] +// DEBUG2-NEXT: resume { ptr, i32 } [[LPAD_VAL26]], !dbg [[DBG182]] // // // DEBUG2-LABEL: define {{[^@]+}}@.__kmpc_global_dtor_..4 diff --git a/clang/test/PCH/cxx_paren_init.cpp b/clang/test/PCH/cxx_paren_init.cpp index 9731ea7737c1983..298e3161007ce21 100644 --- a/clang/test/PCH/cxx_paren_init.cpp +++ b/clang/test/PCH/cxx_paren_init.cpp @@ -16,14 +16,13 @@ U u = baz(3); // CHECK-NEXT: [[ARR:%.*]] = alloca [4 x i32], align 16 // CHECK-NEXT: store i32 [[A:%.*]], ptr [[I_ADDR]], align 4 // CHECK-NEXT: store i32 [[B:%.*]], ptr [[J_ADDR]], align 4 -// CHECK-NEXT: [[ARRINIT_BEGIN:%.*]] = getelementptr inbounds [4 x i32], ptr [[ARR]], i64 0, i64 0 // CHECK-NEXT: [[TMP_0:%.*]] = load i32, ptr [[I_ADDR]], align 4 -// CHECK-NEXT: store i32 [[TMP_0]], ptr [[ARRINIT_BEGIN]], align 4 -// CHECK-NEXT: [[ARRINIT_ELEM:%.*]] = getelementptr inbounds i32, ptr [[ARRINIT_BEGIN]], i64 1 +// CHECK-NEXT: store i32 [[TMP_0]], ptr [[ARR]], align 4 +// CHECK-NEXT: [[ARRINIT_ELEM:%.*]] = getelementptr inbounds i32, ptr [[ARR]], i64 1 // CHECK-NEXT: [[TMP_1:%.*]] = load i32, ptr [[J_ADDR]], align 4 // CHECK-NEXT: store i32 [[TMP_1]], ptr [[ARRINIT_ELEM]], align 4 -// CHECK-NEXT: [[ARRINIT_START:%.*]] = getelementptr inbounds i32, ptr [[ARRINIT_ELEM]], i64 1 -// CHECK-NEXT: [[ARRINIT_END:%.*]] = getelementptr inbounds i32, ptr [[ARRINIT_BEGIN]], i64 4 +// CHECK-NEXT: [[ARRINIT_START:%.*]] = getelementptr inbounds i32, ptr [[ARR]], i64 2 +// CHECK-NEXT: [[ARRINIT_END:%.*]] = getelementptr inbounds i32, ptr [[ARR]], i64 4 // CHECK-NEXT: br label [[ARRINIT_BODY:%.*]] // CHECK: [[ARRINIT_CUR:%.*]] = phi ptr [ [[ARRINIT_START]], %entry ], [ [[ARRINIT_NEXT:%.*]], [[ARRINIT_BODY]] ] // CHECK-NEXT: store i32 0, ptr [[ARRINIT_CUR]], align 4 From e4790ce2096c805b7598c353f9e7fa513701f843 Mon Sep 17 00:00:00 2001 From: Paschalis Mpeis Date: Mon, 10 Jun 2024 08:37:00 +0100 Subject: [PATCH 14/36] [TLI] ReplaceWithVecLib: drop Instruction support (#94365) Refactor the pass to only support `IntrinsicInst` calls. `ReplaceWithVecLib` used to support instructions, as AArch64 was using this pass to replace a vectorized frem instruction to the fmod vector library call (through TLI). As this replacement is now done by the codegen (#83859), there is no need for this pass to support instructions. Additionally, removed 'frem' tests from: - AArch64/replace-with-veclib-armpl.ll - AArch64/replace-with-veclib-sleef-scalable.ll - AArch64/replace-with-veclib-sleef.ll Such testing is done at codegen level: - #83859 --- llvm/lib/CodeGen/ReplaceWithVeclib.cpp | 174 +++++++----------- .../AArch64/replace-with-veclib-armpl.ll | 42 +---- .../replace-with-veclib-sleef-scalable.ll | 20 +- .../AArch64/replace-with-veclib-sleef.ll | 20 +- .../Analysis/ReplaceWithVecLibTest.cpp | 2 +- 5 files changed, 75 insertions(+), 183 deletions(-) diff --git a/llvm/lib/CodeGen/ReplaceWithVeclib.cpp b/llvm/lib/CodeGen/ReplaceWithVeclib.cpp index c67cc57ba675dea..9fbb7b461364b10 100644 --- a/llvm/lib/CodeGen/ReplaceWithVeclib.cpp +++ b/llvm/lib/CodeGen/ReplaceWithVeclib.cpp @@ -6,9 +6,8 @@ // //===----------------------------------------------------------------------===// // -// Replaces LLVM IR instructions with vector operands (i.e., the frem -// instruction or calls to LLVM intrinsics) with matching calls to functions -// from a vector library (e.g libmvec, SVML) using TargetLibraryInfo interface. +// Replaces calls to LLVM Intrinsics with matching calls to functions from a +// vector library (e.g libmvec, SVML) using TargetLibraryInfo interface. // //===----------------------------------------------------------------------===// @@ -25,6 +24,7 @@ #include "llvm/IR/DerivedTypes.h" #include "llvm/IR/IRBuilder.h" #include "llvm/IR/InstIterator.h" +#include "llvm/IR/IntrinsicInst.h" #include "llvm/IR/VFABIDemangler.h" #include "llvm/Support/TypeSize.h" #include "llvm/Transforms/Utils/ModuleUtils.h" @@ -70,84 +70,68 @@ Function *getTLIFunction(Module *M, FunctionType *VectorFTy, return TLIFunc; } -/// Replace the instruction \p I with a call to the corresponding function from -/// the vector library (\p TLIVecFunc). -static void replaceWithTLIFunction(Instruction &I, VFInfo &Info, +/// Replace the intrinsic call \p II to \p TLIVecFunc, which is the +/// corresponding function from the vector library. +static void replaceWithTLIFunction(IntrinsicInst *II, VFInfo &Info, Function *TLIVecFunc) { - IRBuilder<> IRBuilder(&I); - auto *CI = dyn_cast(&I); - SmallVector Args(CI ? CI->args() : I.operands()); + IRBuilder<> IRBuilder(II); + SmallVector Args(II->args()); if (auto OptMaskpos = Info.getParamIndexForOptionalMask()) { auto *MaskTy = - VectorType::get(Type::getInt1Ty(I.getContext()), Info.Shape.VF); + VectorType::get(Type::getInt1Ty(II->getContext()), Info.Shape.VF); Args.insert(Args.begin() + OptMaskpos.value(), Constant::getAllOnesValue(MaskTy)); } - // If it is a call instruction, preserve the operand bundles. + // Preserve the operand bundles. SmallVector OpBundles; - if (CI) - CI->getOperandBundlesAsDefs(OpBundles); + II->getOperandBundlesAsDefs(OpBundles); auto *Replacement = IRBuilder.CreateCall(TLIVecFunc, Args, OpBundles); - I.replaceAllUsesWith(Replacement); + II->replaceAllUsesWith(Replacement); // Preserve fast math flags for FP math. if (isa(Replacement)) - Replacement->copyFastMathFlags(&I); + Replacement->copyFastMathFlags(II); } -/// Returns true when successfully replaced \p I with a suitable function taking -/// vector arguments, based on available mappings in the \p TLI. Currently only -/// works when \p I is a call to vectorized intrinsic or the frem instruction. +/// Returns true when successfully replaced \p II, which is a call to a +/// vectorized intrinsic, with a suitable function taking vector arguments, +/// based on available mappings in the \p TLI. static bool replaceWithCallToVeclib(const TargetLibraryInfo &TLI, - Instruction &I) { + IntrinsicInst *II) { + assert(II != nullptr && "Intrinsic cannot be null"); // At the moment VFABI assumes the return type is always widened unless it is // a void type. - auto *VTy = dyn_cast(I.getType()); + auto *VTy = dyn_cast(II->getType()); ElementCount EC(VTy ? VTy->getElementCount() : ElementCount::getFixed(0)); - - // Compute the argument types of the corresponding scalar call and the scalar - // function name. For calls, it additionally finds the function to replace - // and checks that all vector operands match the previously found EC. + // Compute the argument types of the corresponding scalar call and check that + // all vector operands match the previously found EC. SmallVector ScalarArgTypes; - std::string ScalarName; - Function *FuncToReplace = nullptr; - auto *CI = dyn_cast(&I); - if (CI) { - FuncToReplace = CI->getCalledFunction(); - Intrinsic::ID IID = FuncToReplace->getIntrinsicID(); - assert(IID != Intrinsic::not_intrinsic && "Not an intrinsic"); - for (auto Arg : enumerate(CI->args())) { - auto *ArgTy = Arg.value()->getType(); - if (isVectorIntrinsicWithScalarOpAtArg(IID, Arg.index())) { - ScalarArgTypes.push_back(ArgTy); - } else if (auto *VectorArgTy = dyn_cast(ArgTy)) { - ScalarArgTypes.push_back(VectorArgTy->getElementType()); - // When return type is void, set EC to the first vector argument, and - // disallow vector arguments with different ECs. - if (EC.isZero()) - EC = VectorArgTy->getElementCount(); - else if (EC != VectorArgTy->getElementCount()) - return false; - } else - // Exit when it is supposed to be a vector argument but it isn't. + Intrinsic::ID IID = II->getIntrinsicID(); + for (auto Arg : enumerate(II->args())) { + auto *ArgTy = Arg.value()->getType(); + if (isVectorIntrinsicWithScalarOpAtArg(IID, Arg.index())) { + ScalarArgTypes.push_back(ArgTy); + } else if (auto *VectorArgTy = dyn_cast(ArgTy)) { + ScalarArgTypes.push_back(VectorArgTy->getElementType()); + // When return type is void, set EC to the first vector argument, and + // disallow vector arguments with different ECs. + if (EC.isZero()) + EC = VectorArgTy->getElementCount(); + else if (EC != VectorArgTy->getElementCount()) return false; - } - // Try to reconstruct the name for the scalar version of the instruction, - // using scalar argument types. - ScalarName = Intrinsic::isOverloaded(IID) - ? Intrinsic::getName(IID, ScalarArgTypes, I.getModule()) - : Intrinsic::getName(IID).str(); - } else { - assert(VTy && "Return type must be a vector"); - auto *ScalarTy = VTy->getScalarType(); - LibFunc Func; - if (!TLI.getLibFunc(I.getOpcode(), ScalarTy, Func)) + } else + // Exit when it is supposed to be a vector argument but it isn't. return false; - ScalarName = TLI.getName(Func); - ScalarArgTypes = {ScalarTy, ScalarTy}; } + // Try to reconstruct the name for the scalar version of the instruction, + // using scalar argument types. + std::string ScalarName = + Intrinsic::isOverloaded(IID) + ? Intrinsic::getName(IID, ScalarArgTypes, II->getModule()) + : Intrinsic::getName(IID).str(); + // Try to find the mapping for the scalar version of this intrinsic and the // exact vector width of the call operands in the TargetLibraryInfo. First, // check with a non-masked variant, and if that fails try with a masked one. @@ -162,7 +146,7 @@ static bool replaceWithCallToVeclib(const TargetLibraryInfo &TLI, // Replace the call to the intrinsic with a call to the vector library // function. - Type *ScalarRetTy = I.getType()->getScalarType(); + Type *ScalarRetTy = II->getType()->getScalarType(); FunctionType *ScalarFTy = FunctionType::get(ScalarRetTy, ScalarArgTypes, /*isVarArg*/ false); const std::string MangledName = VD->getVectorFunctionABIVariantString(); @@ -174,22 +158,19 @@ static bool replaceWithCallToVeclib(const TargetLibraryInfo &TLI, // specification when being created, this is why we need to add extra check to // make sure that the operands of the vector function obtained via VFABI match // the operands of the original vector instruction. - if (CI) { - for (auto &VFParam : OptInfo->Shape.Parameters) { - if (VFParam.ParamKind == VFParamKind::GlobalPredicate) - continue; + for (auto &VFParam : OptInfo->Shape.Parameters) { + if (VFParam.ParamKind == VFParamKind::GlobalPredicate) + continue; - // tryDemangleForVFABI must return valid ParamPos, otherwise it could be - // a bug in the VFABI parser. - assert(VFParam.ParamPos < CI->arg_size() && - "ParamPos has invalid range."); - Type *OrigTy = CI->getArgOperand(VFParam.ParamPos)->getType(); - if (OrigTy->isVectorTy() != (VFParam.ParamKind == VFParamKind::Vector)) { - LLVM_DEBUG(dbgs() << DEBUG_TYPE << ": Will not replace: " << ScalarName - << ". Wrong type at index " << VFParam.ParamPos - << ": " << *OrigTy << "\n"); - return false; - } + // tryDemangleForVFABI must return valid ParamPos, otherwise it could be + // a bug in the VFABI parser. + assert(VFParam.ParamPos < II->arg_size() && "ParamPos has invalid range"); + Type *OrigTy = II->getArgOperand(VFParam.ParamPos)->getType(); + if (OrigTy->isVectorTy() != (VFParam.ParamKind == VFParamKind::Vector)) { + LLVM_DEBUG(dbgs() << DEBUG_TYPE << ": Will not replace: " << ScalarName + << ". Wrong type at index " << VFParam.ParamPos << ": " + << *OrigTy << "\n"); + return false; } } @@ -197,45 +178,32 @@ static bool replaceWithCallToVeclib(const TargetLibraryInfo &TLI, if (!VectorFTy) return false; - Function *TLIFunc = getTLIFunction(I.getModule(), VectorFTy, - VD->getVectorFnName(), FuncToReplace); - - replaceWithTLIFunction(I, *OptInfo, TLIFunc); + Function *TLIFunc = + getTLIFunction(II->getModule(), VectorFTy, VD->getVectorFnName(), + II->getCalledFunction()); + replaceWithTLIFunction(II, *OptInfo, TLIFunc); LLVM_DEBUG(dbgs() << DEBUG_TYPE << ": Replaced call to `" << ScalarName << "` with call to `" << TLIFunc->getName() << "`.\n"); ++NumCallsReplaced; return true; } -/// Supported instruction \p I must be a vectorized frem or a call to an -/// intrinsic that returns either void or a vector. -static bool isSupportedInstruction(Instruction *I) { - Type *Ty = I->getType(); - if (auto *CI = dyn_cast(I)) - return (Ty->isVectorTy() || Ty->isVoidTy()) && CI->getCalledFunction() && - CI->getCalledFunction()->getIntrinsicID() != - Intrinsic::not_intrinsic; - if (I->getOpcode() == Instruction::FRem && Ty->isVectorTy()) - return true; - return false; -} - static bool runImpl(const TargetLibraryInfo &TLI, Function &F) { - bool Changed = false; SmallVector ReplacedCalls; for (auto &I : instructions(F)) { - if (!isSupportedInstruction(&I)) - continue; - if (replaceWithCallToVeclib(TLI, I)) { - ReplacedCalls.push_back(&I); - Changed = true; + // Process only intrinsic calls that return void or a vector. + if (auto *II = dyn_cast(&I)) { + if (!II->getType()->isVectorTy() && !II->getType()->isVoidTy()) + continue; + + if (replaceWithCallToVeclib(TLI, II)) + ReplacedCalls.push_back(&I); } } - // Erase the calls to the intrinsics that have been replaced - // with calls to the vector library. - for (auto *CI : ReplacedCalls) - CI->eraseFromParent(); - return Changed; + // Erase any intrinsic calls that were replaced with vector library calls. + for (auto *I : ReplacedCalls) + I->eraseFromParent(); + return !ReplacedCalls.empty(); } //////////////////////////////////////////////////////////////////////////////// @@ -246,7 +214,7 @@ PreservedAnalyses ReplaceWithVeclib::run(Function &F, const TargetLibraryInfo &TLI = AM.getResult(F); auto Changed = runImpl(TLI, F); if (Changed) { - LLVM_DEBUG(dbgs() << "Instructions replaced with vector libraries: " + LLVM_DEBUG(dbgs() << "Intrinsic calls replaced with vector libraries: " << NumCallsReplaced << "\n"); PreservedAnalyses PA; diff --git a/llvm/test/CodeGen/AArch64/replace-with-veclib-armpl.ll b/llvm/test/CodeGen/AArch64/replace-with-veclib-armpl.ll index ce0c5572577dcf6..7d23e870637a8ac 100644 --- a/llvm/test/CodeGen/AArch64/replace-with-veclib-armpl.ll +++ b/llvm/test/CodeGen/AArch64/replace-with-veclib-armpl.ll @@ -15,7 +15,7 @@ declare @llvm.cos.nxv2f64() declare @llvm.cos.nxv4f32() ;. -; CHECK: @llvm.compiler.used = appending global [40 x ptr] [ptr @armpl_vcosq_f64, ptr @armpl_vcosq_f32, ptr @armpl_svcos_f64_x, ptr @armpl_svcos_f32_x, ptr @armpl_vexpq_f64, ptr @armpl_vexpq_f32, ptr @armpl_svexp_f64_x, ptr @armpl_svexp_f32_x, ptr @armpl_vexp10q_f64, ptr @armpl_vexp10q_f32, ptr @armpl_svexp10_f64_x, ptr @armpl_svexp10_f32_x, ptr @armpl_vexp2q_f64, ptr @armpl_vexp2q_f32, ptr @armpl_svexp2_f64_x, ptr @armpl_svexp2_f32_x, ptr @armpl_vlogq_f64, ptr @armpl_vlogq_f32, ptr @armpl_svlog_f64_x, ptr @armpl_svlog_f32_x, ptr @armpl_vlog10q_f64, ptr @armpl_vlog10q_f32, ptr @armpl_svlog10_f64_x, ptr @armpl_svlog10_f32_x, ptr @armpl_vlog2q_f64, ptr @armpl_vlog2q_f32, ptr @armpl_svlog2_f64_x, ptr @armpl_svlog2_f32_x, ptr @armpl_vsinq_f64, ptr @armpl_vsinq_f32, ptr @armpl_svsin_f64_x, ptr @armpl_svsin_f32_x, ptr @armpl_vtanq_f64, ptr @armpl_vtanq_f32, ptr @armpl_svtan_f64_x, ptr @armpl_svtan_f32_x, ptr @armpl_vfmodq_f64, ptr @armpl_vfmodq_f32, ptr @armpl_svfmod_f64_x, ptr @armpl_svfmod_f32_x], section "llvm.metadata" +; CHECK: @llvm.compiler.used = appending global [36 x ptr] [ptr @armpl_vcosq_f64, ptr @armpl_vcosq_f32, ptr @armpl_svcos_f64_x, ptr @armpl_svcos_f32_x, ptr @armpl_vexpq_f64, ptr @armpl_vexpq_f32, ptr @armpl_svexp_f64_x, ptr @armpl_svexp_f32_x, ptr @armpl_vexp10q_f64, ptr @armpl_vexp10q_f32, ptr @armpl_svexp10_f64_x, ptr @armpl_svexp10_f32_x, ptr @armpl_vexp2q_f64, ptr @armpl_vexp2q_f32, ptr @armpl_svexp2_f64_x, ptr @armpl_svexp2_f32_x, ptr @armpl_vlogq_f64, ptr @armpl_vlogq_f32, ptr @armpl_svlog_f64_x, ptr @armpl_svlog_f32_x, ptr @armpl_vlog10q_f64, ptr @armpl_vlog10q_f32, ptr @armpl_svlog10_f64_x, ptr @armpl_svlog10_f32_x, ptr @armpl_vlog2q_f64, ptr @armpl_vlog2q_f32, ptr @armpl_svlog2_f64_x, ptr @armpl_svlog2_f32_x, ptr @armpl_vsinq_f64, ptr @armpl_vsinq_f32, ptr @armpl_svsin_f64_x, ptr @armpl_svsin_f32_x, ptr @armpl_vtanq_f64, ptr @armpl_vtanq_f32, ptr @armpl_svtan_f64_x, ptr @armpl_svtan_f32_x], section "llvm.metadata" ;. define <2 x double> @llvm_cos_f64(<2 x double> %in) { ; CHECK-LABEL: define <2 x double> @llvm_cos_f64 @@ -469,46 +469,6 @@ define @llvm_tan_vscale_f32( %in) #0 { ret %1 } -define <2 x double> @frem_f64(<2 x double> %in) { -; CHECK-LABEL: define <2 x double> @frem_f64 -; CHECK-SAME: (<2 x double> [[IN:%.*]]) { -; CHECK-NEXT: [[TMP1:%.*]] = call <2 x double> @armpl_vfmodq_f64(<2 x double> [[IN]], <2 x double> [[IN]]) -; CHECK-NEXT: ret <2 x double> [[TMP1]] -; - %1= frem <2 x double> %in, %in - ret <2 x double> %1 -} - -define <4 x float> @frem_f32(<4 x float> %in) { -; CHECK-LABEL: define <4 x float> @frem_f32 -; CHECK-SAME: (<4 x float> [[IN:%.*]]) { -; CHECK-NEXT: [[TMP1:%.*]] = call <4 x float> @armpl_vfmodq_f32(<4 x float> [[IN]], <4 x float> [[IN]]) -; CHECK-NEXT: ret <4 x float> [[TMP1]] -; - %1= frem <4 x float> %in, %in - ret <4 x float> %1 -} - -define @frem_vscale_f64( %in) #0 { -; CHECK-LABEL: define @frem_vscale_f64 -; CHECK-SAME: ( [[IN:%.*]]) #[[ATTR1]] { -; CHECK-NEXT: [[TMP1:%.*]] = call @armpl_svfmod_f64_x( [[IN]], [[IN]], shufflevector ( insertelement ( poison, i1 true, i64 0), poison, zeroinitializer)) -; CHECK-NEXT: ret [[TMP1]] -; - %1= frem %in, %in - ret %1 -} - -define @frem_vscale_f32( %in) #0 { -; CHECK-LABEL: define @frem_vscale_f32 -; CHECK-SAME: ( [[IN:%.*]]) #[[ATTR1]] { -; CHECK-NEXT: [[TMP1:%.*]] = call @armpl_svfmod_f32_x( [[IN]], [[IN]], shufflevector ( insertelement ( poison, i1 true, i64 0), poison, zeroinitializer)) -; CHECK-NEXT: ret [[TMP1]] -; - %1= frem %in, %in - ret %1 -} - attributes #0 = { "target-features"="+sve" } ;. ; CHECK: attributes #[[ATTR0:[0-9]+]] = { nocallback nofree nosync nounwind speculatable willreturn memory(none) } diff --git a/llvm/test/CodeGen/AArch64/replace-with-veclib-sleef-scalable.ll b/llvm/test/CodeGen/AArch64/replace-with-veclib-sleef-scalable.ll index c58a9bc8342de36..15d100a518c1509 100644 --- a/llvm/test/CodeGen/AArch64/replace-with-veclib-sleef-scalable.ll +++ b/llvm/test/CodeGen/AArch64/replace-with-veclib-sleef-scalable.ll @@ -4,7 +4,7 @@ target triple = "aarch64-unknown-linux-gnu" ;. -; CHECK: @llvm.compiler.used = appending global [20 x ptr] [ptr @_ZGVsMxv_cos, ptr @_ZGVsMxv_cosf, ptr @_ZGVsMxv_exp, ptr @_ZGVsMxv_expf, ptr @_ZGVsMxv_exp10, ptr @_ZGVsMxv_exp10f, ptr @_ZGVsMxv_exp2, ptr @_ZGVsMxv_exp2f, ptr @_ZGVsMxv_log, ptr @_ZGVsMxv_logf, ptr @_ZGVsMxv_log10, ptr @_ZGVsMxv_log10f, ptr @_ZGVsMxv_log2, ptr @_ZGVsMxv_log2f, ptr @_ZGVsMxv_sin, ptr @_ZGVsMxv_sinf, ptr @_ZGVsMxv_tan, ptr @_ZGVsMxv_tanf, ptr @_ZGVsMxvv_fmod, ptr @_ZGVsMxvv_fmodf], section "llvm.metadata" +; CHECK: @llvm.compiler.used = appending global [18 x ptr] [ptr @_ZGVsMxv_cos, ptr @_ZGVsMxv_cosf, ptr @_ZGVsMxv_exp, ptr @_ZGVsMxv_expf, ptr @_ZGVsMxv_exp10, ptr @_ZGVsMxv_exp10f, ptr @_ZGVsMxv_exp2, ptr @_ZGVsMxv_exp2f, ptr @_ZGVsMxv_log, ptr @_ZGVsMxv_logf, ptr @_ZGVsMxv_log10, ptr @_ZGVsMxv_log10f, ptr @_ZGVsMxv_log2, ptr @_ZGVsMxv_log2f, ptr @_ZGVsMxv_sin, ptr @_ZGVsMxv_sinf, ptr @_ZGVsMxv_tan, ptr @_ZGVsMxv_tanf], section "llvm.metadata" ;. define @llvm_ceil_vscale_f64( %in) { ; CHECK-LABEL: @llvm_ceil_vscale_f64( @@ -403,24 +403,6 @@ define @llvm_trunc_vscale_f32( %in) { ret %1 } -define @frem_f64( %in) { -; CHECK-LABEL: @frem_f64( -; CHECK-NEXT: [[TMP1:%.*]] = call @_ZGVsMxvv_fmod( [[IN:%.*]], [[IN]], shufflevector ( insertelement ( poison, i1 true, i64 0), poison, zeroinitializer)) -; CHECK-NEXT: ret [[TMP1]] -; - %1= frem %in, %in - ret %1 -} - -define @frem_f32( %in) { -; CHECK-LABEL: @frem_f32( -; CHECK-NEXT: [[TMP1:%.*]] = call @_ZGVsMxvv_fmodf( [[IN:%.*]], [[IN]], shufflevector ( insertelement ( poison, i1 true, i64 0), poison, zeroinitializer)) -; CHECK-NEXT: ret [[TMP1]] -; - %1= frem %in, %in - ret %1 -} - declare @llvm.ceil.nxv2f64() declare @llvm.ceil.nxv4f32() declare @llvm.copysign.nxv2f64(, ) diff --git a/llvm/test/CodeGen/AArch64/replace-with-veclib-sleef.ll b/llvm/test/CodeGen/AArch64/replace-with-veclib-sleef.ll index 1df2caed4244e1d..a3da3b8120218c9 100644 --- a/llvm/test/CodeGen/AArch64/replace-with-veclib-sleef.ll +++ b/llvm/test/CodeGen/AArch64/replace-with-veclib-sleef.ll @@ -4,7 +4,7 @@ target triple = "aarch64-unknown-linux-gnu" ;. -; CHECK: @llvm.compiler.used = appending global [20 x ptr] [ptr @_ZGVnN2v_cos, ptr @_ZGVnN4v_cosf, ptr @_ZGVnN2v_exp, ptr @_ZGVnN4v_expf, ptr @_ZGVnN2v_exp10, ptr @_ZGVnN4v_exp10f, ptr @_ZGVnN2v_exp2, ptr @_ZGVnN4v_exp2f, ptr @_ZGVnN2v_log, ptr @_ZGVnN4v_logf, ptr @_ZGVnN2v_log10, ptr @_ZGVnN4v_log10f, ptr @_ZGVnN2v_log2, ptr @_ZGVnN4v_log2f, ptr @_ZGVnN2v_sin, ptr @_ZGVnN4v_sinf, ptr @_ZGVnN2v_tan, ptr @_ZGVnN4v_tanf, ptr @_ZGVnN2vv_fmod, ptr @_ZGVnN4vv_fmodf], section "llvm.metadata" +; CHECK: @llvm.compiler.used = appending global [18 x ptr] [ptr @_ZGVnN2v_cos, ptr @_ZGVnN4v_cosf, ptr @_ZGVnN2v_exp, ptr @_ZGVnN4v_expf, ptr @_ZGVnN2v_exp10, ptr @_ZGVnN4v_exp10f, ptr @_ZGVnN2v_exp2, ptr @_ZGVnN4v_exp2f, ptr @_ZGVnN2v_log, ptr @_ZGVnN4v_logf, ptr @_ZGVnN2v_log10, ptr @_ZGVnN4v_log10f, ptr @_ZGVnN2v_log2, ptr @_ZGVnN4v_log2f, ptr @_ZGVnN2v_sin, ptr @_ZGVnN4v_sinf, ptr @_ZGVnN2v_tan, ptr @_ZGVnN4v_tanf], section "llvm.metadata" ;. define <2 x double> @llvm_ceil_f64(<2 x double> %in) { ; CHECK-LABEL: @llvm_ceil_f64( @@ -402,24 +402,6 @@ define <4 x float> @llvm_trunc_f32(<4 x float> %in) { ret <4 x float> %1 } -define <2 x double> @frem_f64(<2 x double> %in) { -; CHECK-LABEL: @frem_f64( -; CHECK-NEXT: [[TMP1:%.*]] = call <2 x double> @_ZGVnN2vv_fmod(<2 x double> [[IN:%.*]], <2 x double> [[IN]]) -; CHECK-NEXT: ret <2 x double> [[TMP1]] -; - %1= frem <2 x double> %in, %in - ret <2 x double> %1 -} - -define <4 x float> @frem_f32(<4 x float> %in) { -; CHECK-LABEL: @frem_f32( -; CHECK-NEXT: [[TMP1:%.*]] = call <4 x float> @_ZGVnN4vv_fmodf(<4 x float> [[IN:%.*]], <4 x float> [[IN]]) -; CHECK-NEXT: ret <4 x float> [[TMP1]] -; - %1= frem <4 x float> %in, %in - ret <4 x float> %1 -} - declare <2 x double> @llvm.ceil.v2f64(<2 x double>) declare <4 x float> @llvm.ceil.v4f32(<4 x float>) declare <2 x double> @llvm.copysign.v2f64(<2 x double>, <2 x double>) diff --git a/llvm/unittests/Analysis/ReplaceWithVecLibTest.cpp b/llvm/unittests/Analysis/ReplaceWithVecLibTest.cpp index 95f17f93bf4c0b8..d97549cb69d556d 100644 --- a/llvm/unittests/Analysis/ReplaceWithVecLibTest.cpp +++ b/llvm/unittests/Analysis/ReplaceWithVecLibTest.cpp @@ -99,7 +99,7 @@ TEST_F(ReplaceWithVecLibTest, TestValidMapping) { ElementCount::getScalable(4), /*Masked*/ true, "_ZGVsMxvu"}; EXPECT_EQ(run(CorrectVD, IR), - "Instructions replaced with vector libraries: 1"); + "Intrinsic calls replaced with vector libraries: 1"); } // The VFABI prefix in TLI describes signature which is not matching the powi From 832b91fcb9fd42d92a741129c6e085ef15ab385b Mon Sep 17 00:00:00 2001 From: Orlando Cazalet-Hyams Date: Mon, 10 Jun 2024 08:47:39 +0100 Subject: [PATCH 15/36] [dexter] Correctly identify stop-reason while driving VisualStudio (#94754) Prior to this patch VisualStudio._get_step_info incorrectly identifies the reason the debugger has stopped. e.g., stepping through a program would be reported as a StopReason.Breakpoint rather than StopReason.Step. Fix. No test added as there are no VisualStudio tests (tested locally). --- .../dex/debugger/visualstudio/VisualStudio.py | 42 ++++++++++++++++--- 1 file changed, 37 insertions(+), 5 deletions(-) diff --git a/cross-project-tests/debuginfo-tests/dexter/dex/debugger/visualstudio/VisualStudio.py b/cross-project-tests/debuginfo-tests/dexter/dex/debugger/visualstudio/VisualStudio.py index 0e20cfbbd264b23..17587b3f3e18d68 100644 --- a/cross-project-tests/debuginfo-tests/dexter/dex/debugger/visualstudio/VisualStudio.py +++ b/cross-project-tests/debuginfo-tests/dexter/dex/debugger/visualstudio/VisualStudio.py @@ -10,6 +10,7 @@ import imp import os import sys +from enum import IntEnum from pathlib import PurePath, Path from collections import defaultdict, namedtuple @@ -37,6 +38,26 @@ def _load_com_module(): VSBreakpoint = namedtuple("VSBreakpoint", "path, line, col, cond") +# Visual Studio events. +# https://learn.microsoft.com/en-us/dotnet/api/envdte.dbgeventreason?view=visualstudiosdk-2022 +class DbgEvent(IntEnum): + dbgEventReasonNone = 1 + dbgEventReasonGo = 2 + dbgEventReasonAttachProgram = 3 + dbgEventReasonDetachProgram = 4 + dbgEventReasonLaunchProgram = 5 + dbgEventReasonEndProgram = 6 + dbgEventReasonStopDebugging = 7 + dbgEventReasonStep = 8 + dbgEventReasonBreakpoint = 9 + dbgEventReasonExceptionThrown = 10 + dbgEventReasonExceptionNotHandled = 11 + dbgEventReasonUserBreak = 12 + dbgEventReasonContextSwitch = 13 + + first = dbgEventReasonNone + last = dbgEventReasonContextSwitch + class VisualStudio( DebuggerBase, metaclass=abc.ABCMeta ): # pylint: disable=abstract-method @@ -307,6 +328,20 @@ def set_current_stack_frame(self, idx: int = 0): ) ) + def _translate_stop_reason(self, reason): + if reason == DbgEvent.dbgEventReasonNone: + return None + if reason == DbgEvent.dbgEventReasonBreakpoint: + return StopReason.BREAKPOINT + if reason == DbgEvent.dbgEventReasonStep: + return StopReason.STEP + if reason == DbgEvent.dbgEventReasonEndProgram: + return StopReason.PROGRAM_EXIT + if reason == DbgEvent.dbgEventReasonExceptionNotHandled: + return StopReason.ERROR + assert reason <= DbgEvent.last and reason >= DbgEvent.first + return StopReason.OTHER + def _get_step_info(self, watches, step_index): thread = self._debugger.CurrentThread stackframes = thread.StackFrames @@ -347,16 +382,13 @@ def _get_step_info(self, watches, step_index): frames[0].loc = loc state_frames[0].location = SourceLocation(**self._location) - reason = StopReason.BREAKPOINT - if loc.path is None: # pylint: disable=no-member - reason = StopReason.STEP - + stop_reason = self._translate_stop_reason(self._debugger.LastBreakReason) program_state = ProgramState(frames=state_frames) return StepIR( step_index=step_index, frames=frames, - stop_reason=reason, + stop_reason=stop_reason, program_state=program_state, ) From e58f830a1de20af79a6ce41c85b787f4a64b9c74 Mon Sep 17 00:00:00 2001 From: jeanPerier Date: Mon, 10 Jun 2024 10:11:06 +0200 Subject: [PATCH 16/36] [flang] add source to SHAPE API (#94781) --- flang/include/flang/Runtime/inquiry.h | 3 ++- flang/runtime/inquiry.cpp | 5 +++-- flang/unittests/Runtime/Inquiry.cpp | 9 ++++++--- 3 files changed, 11 insertions(+), 6 deletions(-) diff --git a/flang/include/flang/Runtime/inquiry.h b/flang/include/flang/Runtime/inquiry.h index 7161d1e41c4bb8a..0a3cd51236fa39b 100644 --- a/flang/include/flang/Runtime/inquiry.h +++ b/flang/include/flang/Runtime/inquiry.h @@ -24,7 +24,8 @@ extern "C" { std::int64_t RTDECL(LboundDim)(const Descriptor &array, int dim, const char *sourceFile = nullptr, int line = 0); -void RTDECL(Shape)(void *result, const Descriptor &array, int kind); +void RTDECL(Shape)(void *result, const Descriptor &array, int kind, + const char *sourceFile = nullptr, int line = 0); std::int64_t RTDECL(Size)( const Descriptor &array, const char *sourceFile = nullptr, int line = 0); diff --git a/flang/runtime/inquiry.cpp b/flang/runtime/inquiry.cpp index ea114174de7fd25..443e6291e5e2374 100644 --- a/flang/runtime/inquiry.cpp +++ b/flang/runtime/inquiry.cpp @@ -85,8 +85,9 @@ std::int64_t RTDEF(SizeDim)( return static_cast(dimension.Extent()); } -void RTDEF(Shape)(void *result, const Descriptor &array, int kind) { - Terminator terminator{__FILE__, __LINE__}; +void RTDEF(Shape)(void *result, const Descriptor &array, int kind, + const char *sourceFile, int line) { + Terminator terminator{sourceFile, line}; INTERNAL_CHECK(array.rank() <= common::maxRank); for (SubscriptValue i{0}; i < array.rank(); ++i) { const Dimension &dimension{array.GetDimension(i)}; diff --git a/flang/unittests/Runtime/Inquiry.cpp b/flang/unittests/Runtime/Inquiry.cpp index 665a930ee4ff9b2..220ebb765f8c721 100644 --- a/flang/unittests/Runtime/Inquiry.cpp +++ b/flang/unittests/Runtime/Inquiry.cpp @@ -87,7 +87,8 @@ TEST(Inquiry, Shape) { auto int8Result{ MakeArray(std::vector{array->rank()}, std::vector(array->rank(), 0))}; - RTNAME(Shape)(int8Result->raw().base_addr, *array, /*KIND=*/1); + RTNAME(Shape) + (int8Result->raw().base_addr, *array, /*KIND=*/1, __FILE__, __LINE__); EXPECT_EQ(*int8Result->ZeroBasedIndexedElement(0), 2); EXPECT_EQ(*int8Result->ZeroBasedIndexedElement(1), 3); @@ -95,7 +96,8 @@ TEST(Inquiry, Shape) { auto int32Result{ MakeArray(std::vector{array->rank()}, std::vector(array->rank(), 0))}; - RTNAME(Shape)(int32Result->raw().base_addr, *array, /*KIND=*/4); + RTNAME(Shape) + (int32Result->raw().base_addr, *array, /*KIND=*/4, __FILE__, __LINE__); EXPECT_EQ(*int32Result->ZeroBasedIndexedElement(0), 2); EXPECT_EQ(*int32Result->ZeroBasedIndexedElement(1), 3); @@ -103,7 +105,8 @@ TEST(Inquiry, Shape) { auto int64Result{ MakeArray(std::vector{array->rank()}, std::vector(array->rank(), 0))}; - RTNAME(Shape)(int64Result->raw().base_addr, *array, /*KIND=*/8); + RTNAME(Shape) + (int64Result->raw().base_addr, *array, /*KIND=*/8, __FILE__, __LINE__); EXPECT_EQ(*int64Result->ZeroBasedIndexedElement(0), 2); EXPECT_EQ(*int64Result->ZeroBasedIndexedElement(1), 3); } From cc158d4c0bb50554dadcad44c10cafb052376cea Mon Sep 17 00:00:00 2001 From: Nikita Popov Date: Fri, 31 May 2024 09:25:38 +0200 Subject: [PATCH 17/36] Reapply [ConstantFold] Remove non-trivial gep-of-gep fold (#93823) Reapply after https://github.com/llvm/llvm-project/pull/93956, which changed clang array initialization codegen to avoid size regressions for unoptimized builds. ----- This fold is subtly incorrect, because DL-unaware constant folding does not know the correct index type to use, and just performs the addition in the type that happens to already be there. This is incorrect, since sext(X)+sext(Y) is generally not the same as sext(X+Y). See the `@constexpr_gep_of_gep_with_narrow_type()` for a miscompile with the current implementation. One could try to restrict the fold to cases where no overflow occurs, but I'm not bothering with that here, because the DL-aware constant folding will take care of this anyway. I've only kept the straightforward zero-index case, where we just concatenate two GEPs. --- llvm/lib/IR/ConstantFold.cpp | 56 ++----------------- llvm/test/Other/constant-fold-gep.ll | 2 +- llvm/test/Transforms/InstCombine/gepgep.ll | 2 +- .../Transforms/InstCombine/getelementptr.ll | 3 +- .../InstCombine/ptrtoint-nullgep.ll | 16 +++--- 5 files changed, 15 insertions(+), 64 deletions(-) diff --git a/llvm/lib/IR/ConstantFold.cpp b/llvm/lib/IR/ConstantFold.cpp index 7b8b9a7b652daec..77a833610a3a911 100644 --- a/llvm/lib/IR/ConstantFold.cpp +++ b/llvm/lib/IR/ConstantFold.cpp @@ -1416,64 +1416,16 @@ static Constant *foldGEPOfGEP(GEPOperator *GEP, Type *PointeeTy, bool InBounds, if (GEP->getInRange()) return nullptr; + // Only handle simple case with leading zero index. We cannot perform an + // actual addition as we don't know the correct index type size to use. Constant *Idx0 = cast(Idxs[0]); - if (Idx0->isNullValue()) { - // Handle the simple case of a zero index. - SmallVector NewIndices; - NewIndices.reserve(Idxs.size() + GEP->getNumIndices()); - NewIndices.append(GEP->idx_begin(), GEP->idx_end()); - NewIndices.append(Idxs.begin() + 1, Idxs.end()); - return ConstantExpr::getGetElementPtr( - GEP->getSourceElementType(), cast(GEP->getPointerOperand()), - NewIndices, InBounds && GEP->isInBounds()); - } - - gep_type_iterator LastI = gep_type_end(GEP); - for (gep_type_iterator I = gep_type_begin(GEP), E = gep_type_end(GEP); - I != E; ++I) - LastI = I; - - // We can't combine GEPs if the last index is a struct type. - if (!LastI.isSequential()) - return nullptr; - // We could perform the transform with non-constant index, but prefer leaving - // it as GEP of GEP rather than GEP of add for now. - ConstantInt *CI = dyn_cast(Idx0); - if (!CI) - return nullptr; - - // TODO: This code may be extended to handle vectors as well. - auto *LastIdx = cast(GEP->getOperand(GEP->getNumOperands()-1)); - Type *LastIdxTy = LastIdx->getType(); - if (LastIdxTy->isVectorTy()) + if (!Idx0->isNullValue()) return nullptr; SmallVector NewIndices; NewIndices.reserve(Idxs.size() + GEP->getNumIndices()); - NewIndices.append(GEP->idx_begin(), GEP->idx_end() - 1); - - // Add the last index of the source with the first index of the new GEP. - // Make sure to handle the case when they are actually different types. - if (LastIdxTy != Idx0->getType()) { - unsigned CommonExtendedWidth = - std::max(LastIdxTy->getIntegerBitWidth(), - Idx0->getType()->getIntegerBitWidth()); - CommonExtendedWidth = std::max(CommonExtendedWidth, 64U); - - Type *CommonTy = - Type::getIntNTy(LastIdxTy->getContext(), CommonExtendedWidth); - if (Idx0->getType() != CommonTy) - Idx0 = ConstantFoldCastInstruction(Instruction::SExt, Idx0, CommonTy); - if (LastIdx->getType() != CommonTy) - LastIdx = - ConstantFoldCastInstruction(Instruction::SExt, LastIdx, CommonTy); - if (!Idx0 || !LastIdx) - return nullptr; - } - - NewIndices.push_back(ConstantExpr::get(Instruction::Add, Idx0, LastIdx)); + NewIndices.append(GEP->idx_begin(), GEP->idx_end()); NewIndices.append(Idxs.begin() + 1, Idxs.end()); - return ConstantExpr::getGetElementPtr( GEP->getSourceElementType(), cast(GEP->getPointerOperand()), NewIndices, InBounds && GEP->isInBounds()); diff --git a/llvm/test/Other/constant-fold-gep.ll b/llvm/test/Other/constant-fold-gep.ll index d4bd24df215e85b..e0d535e69527309 100644 --- a/llvm/test/Other/constant-fold-gep.ll +++ b/llvm/test/Other/constant-fold-gep.ll @@ -104,7 +104,7 @@ ; Fold GEP of a GEP. Very simple cases are folded without targetdata. -; PLAIN: @Y = global ptr getelementptr inbounds ([3 x { i32, i32 }], ptr @ext, i64 2) +; PLAIN: @Y = global ptr getelementptr inbounds ([3 x { i32, i32 }], ptr getelementptr inbounds ([3 x { i32, i32 }], ptr @ext, i64 1), i64 1) ; PLAIN: @Z = global ptr getelementptr inbounds (i32, ptr getelementptr inbounds ([3 x { i32, i32 }], ptr @ext, i64 0, i64 1, i32 0), i64 1) ; OPT: @Y = local_unnamed_addr global ptr getelementptr inbounds (i8, ptr @ext, i64 48) ; OPT: @Z = local_unnamed_addr global ptr getelementptr inbounds (i8, ptr @ext, i64 12) diff --git a/llvm/test/Transforms/InstCombine/gepgep.ll b/llvm/test/Transforms/InstCombine/gepgep.ll index d2a0e1db392ba0f..6b6a1597d560bea 100644 --- a/llvm/test/Transforms/InstCombine/gepgep.ll +++ b/llvm/test/Transforms/InstCombine/gepgep.ll @@ -10,7 +10,7 @@ declare void @use(ptr) define void @f() { ; CHECK-LABEL: define void @f() { -; CHECK-NEXT: call void @use(ptr getelementptr (i8, ptr @buffer, i64 add (i64 sub (i64 0, i64 ptrtoint (ptr @buffer to i64)), i64 127))) +; CHECK-NEXT: call void @use(ptr getelementptr (i8, ptr getelementptr (i8, ptr @buffer, i64 add (i64 sub (i64 0, i64 ptrtoint (ptr @buffer to i64)), i64 63)), i64 64)) ; CHECK-NEXT: ret void ; call void @use(ptr getelementptr (i8, ptr getelementptr (i8, ptr @buffer, i64 add (i64 sub (i64 0, i64 ptrtoint (ptr @buffer to i64)), i64 63)), i64 64)) diff --git a/llvm/test/Transforms/InstCombine/getelementptr.ll b/llvm/test/Transforms/InstCombine/getelementptr.ll index d5cfb74a4443013..871913610429b30 100644 --- a/llvm/test/Transforms/InstCombine/getelementptr.ll +++ b/llvm/test/Transforms/InstCombine/getelementptr.ll @@ -1716,10 +1716,9 @@ if.else: @g = external global i8 -; FIXME: This is a miscompile define ptr @constexpr_gep_of_gep_with_narrow_type() { ; CHECK-LABEL: @constexpr_gep_of_gep_with_narrow_type( -; CHECK-NEXT: ret ptr getelementptr (i8, ptr @g, i64 -2) +; CHECK-NEXT: ret ptr getelementptr (i8, ptr @g, i64 254) ; ret ptr getelementptr (i8, ptr getelementptr (i8, ptr @g, i8 127), i8 127) } diff --git a/llvm/test/Transforms/InstCombine/ptrtoint-nullgep.ll b/llvm/test/Transforms/InstCombine/ptrtoint-nullgep.ll index e6e63778d82f005..9d6f0abce84c7c5 100644 --- a/llvm/test/Transforms/InstCombine/ptrtoint-nullgep.ll +++ b/llvm/test/Transforms/InstCombine/ptrtoint-nullgep.ll @@ -68,10 +68,10 @@ define void @constant_fold_ptrtoint_of_gep_of_nullgep() { ; LLPARSER-NEXT: call void @use_i64(i64 ptrtoint (ptr addrspace(1) getelementptr inbounds (i8, ptr addrspace(1) null, i64 1234) to i64)) ; LLPARSER-NEXT: call void @use_i64(i64 ptrtoint (ptr addrspace(1) getelementptr (i8, ptr addrspace(1) null, i64 1234) to i64)) ; LLPARSER-NEXT: call void @use_i64(i64 ptrtoint (ptr addrspace(1) getelementptr (i8, ptr addrspace(1) null, i64 1234) to i64)) -; LLPARSER-NEXT: call void @use_i64(i64 0) -; LLPARSER-NEXT: call void @use_i64(i64 0) -; LLPARSER-NEXT: call void @use_i64(i64 0) -; LLPARSER-NEXT: call void @use_i64(i64 0) +; LLPARSER-NEXT: call void @use_i64(i64 ptrtoint (ptr addrspace(1) getelementptr inbounds (i8, ptr addrspace(1) getelementptr inbounds (i8, ptr addrspace(1) null, i64 -1), i64 1) to i64)) +; LLPARSER-NEXT: call void @use_i64(i64 ptrtoint (ptr addrspace(1) getelementptr (i8, ptr addrspace(1) getelementptr inbounds (i8, ptr addrspace(1) null, i64 -1), i64 1) to i64)) +; LLPARSER-NEXT: call void @use_i64(i64 ptrtoint (ptr addrspace(1) getelementptr inbounds (i8, ptr addrspace(1) getelementptr (i8, ptr addrspace(1) null, i64 -1), i64 1) to i64)) +; LLPARSER-NEXT: call void @use_i64(i64 ptrtoint (ptr addrspace(1) getelementptr (i8, ptr addrspace(1) getelementptr (i8, ptr addrspace(1) null, i64 -1), i64 1) to i64)) ; LLPARSER-NEXT: ret void ; ; INSTSIMPLIFY-LABEL: define {{[^@]+}}@constant_fold_ptrtoint_of_gep_of_nullgep() { @@ -83,10 +83,10 @@ define void @constant_fold_ptrtoint_of_gep_of_nullgep() { ; INSTSIMPLIFY-NEXT: call void @use_i64(i64 ptrtoint (ptr addrspace(1) getelementptr inbounds (i8, ptr addrspace(1) null, i64 1234) to i64)) ; INSTSIMPLIFY-NEXT: call void @use_i64(i64 ptrtoint (ptr addrspace(1) getelementptr (i8, ptr addrspace(1) null, i64 1234) to i64)) ; INSTSIMPLIFY-NEXT: call void @use_i64(i64 ptrtoint (ptr addrspace(1) getelementptr (i8, ptr addrspace(1) null, i64 1234) to i64)) -; INSTSIMPLIFY-NEXT: call void @use_i64(i64 0) -; INSTSIMPLIFY-NEXT: call void @use_i64(i64 0) -; INSTSIMPLIFY-NEXT: call void @use_i64(i64 0) -; INSTSIMPLIFY-NEXT: call void @use_i64(i64 0) +; INSTSIMPLIFY-NEXT: call void @use_i64(i64 ptrtoint (ptr addrspace(1) getelementptr inbounds (i8, ptr addrspace(1) getelementptr inbounds (i8, ptr addrspace(1) null, i64 -1), i64 1) to i64)) +; INSTSIMPLIFY-NEXT: call void @use_i64(i64 ptrtoint (ptr addrspace(1) getelementptr (i8, ptr addrspace(1) getelementptr inbounds (i8, ptr addrspace(1) null, i64 -1), i64 1) to i64)) +; INSTSIMPLIFY-NEXT: call void @use_i64(i64 ptrtoint (ptr addrspace(1) getelementptr inbounds (i8, ptr addrspace(1) getelementptr (i8, ptr addrspace(1) null, i64 -1), i64 1) to i64)) +; INSTSIMPLIFY-NEXT: call void @use_i64(i64 ptrtoint (ptr addrspace(1) getelementptr (i8, ptr addrspace(1) getelementptr (i8, ptr addrspace(1) null, i64 -1), i64 1) to i64)) ; INSTSIMPLIFY-NEXT: ret void ; ; INSTCOMBINE-LABEL: define {{[^@]+}}@constant_fold_ptrtoint_of_gep_of_nullgep() { From c0b65a2491be44f0b162d55563c305c514c01d90 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Timm=20B=C3=A4der?= Date: Mon, 10 Jun 2024 10:10:43 +0200 Subject: [PATCH 18/36] [clang][Interp] Diagnose casts from void pointers --- clang/lib/AST/Interp/ByteCodeExprGen.cpp | 5 ++++- clang/lib/AST/Interp/Interp.h | 7 +++++++ clang/lib/AST/Interp/Opcodes.td | 1 + clang/test/AST/Interp/c.c | 6 ++---- clang/test/Sema/constexpr-void-cast.c | 4 ++++ 5 files changed, 18 insertions(+), 5 deletions(-) diff --git a/clang/lib/AST/Interp/ByteCodeExprGen.cpp b/clang/lib/AST/Interp/ByteCodeExprGen.cpp index ac245e7a0f24cee..6654a27c9216893 100644 --- a/clang/lib/AST/Interp/ByteCodeExprGen.cpp +++ b/clang/lib/AST/Interp/ByteCodeExprGen.cpp @@ -325,8 +325,11 @@ bool ByteCodeExprGen::VisitCastExpr(const CastExpr *CE) { assert(isPtrType(*FromT)); assert(isPtrType(*ToT)); - if (FromT == ToT) + if (FromT == ToT) { + if (SubExpr->getType()->isVoidPointerType()) + return this->visit(SubExpr) && this->emitVoidPtrCast(CE); return this->delegate(SubExpr); + } if (!this->visit(SubExpr)) return false; diff --git a/clang/lib/AST/Interp/Interp.h b/clang/lib/AST/Interp/Interp.h index 4bb9e7f0d22ea63..0ad710c5ec1afc9 100644 --- a/clang/lib/AST/Interp/Interp.h +++ b/clang/lib/AST/Interp/Interp.h @@ -1980,6 +1980,13 @@ static inline bool CastPointerIntegralAPS(InterpState &S, CodePtr OpPC, return true; } +static inline bool VoidPtrCast(InterpState &S, CodePtr OpPC) { + const SourceInfo &E = S.Current->getSource(OpPC); + S.CCEDiag(E, diag::note_constexpr_invalid_cast) + << 2 << S.getLangOpts().CPlusPlus << S.Current->getRange(OpPC); + return true; +} + //===----------------------------------------------------------------------===// // Zero, Nullptr //===----------------------------------------------------------------------===// diff --git a/clang/lib/AST/Interp/Opcodes.td b/clang/lib/AST/Interp/Opcodes.td index a5ac8206104c840..ac5426c87c2123e 100644 --- a/clang/lib/AST/Interp/Opcodes.td +++ b/clang/lib/AST/Interp/Opcodes.td @@ -665,6 +665,7 @@ def CastPointerIntegralAPS : Opcode { let HasGroup = 0; let Args = [ArgUint32]; } +def VoidPtrCast : Opcode; def DecayPtr : Opcode { let Types = [PtrTypeClass, PtrTypeClass]; diff --git a/clang/test/AST/Interp/c.c b/clang/test/AST/Interp/c.c index f4c7bf16f2f95ee..1cc450e48def06f 100644 --- a/clang/test/AST/Interp/c.c +++ b/clang/test/AST/Interp/c.c @@ -66,12 +66,10 @@ _Static_assert((&a - 100) != 0, ""); // pedantic-ref-warning {{is a GNU extensio // pedantic-ref-note {{-100 of non-array}} \ // pedantic-expected-note {{-100 of non-array}} /// extern variable of a composite type. -/// FIXME: The 'this conversion is not allowed' note is missing in the new interpreter. extern struct Test50S Test50; _Static_assert(&Test50 != (void*)0, ""); // all-warning {{always true}} \ - // pedantic-ref-warning {{is a GNU extension}} \ - // pedantic-ref-note {{this conversion is not allowed in a constant expression}} \ - // pedantic-expected-warning {{is a GNU extension}} + // pedantic-warning {{is a GNU extension}} \ + // pedantic-note {{this conversion is not allowed in a constant expression}} struct y {int x,y;}; int a2[(intptr_t)&((struct y*)0)->y]; // all-warning {{folded to constant array}} diff --git a/clang/test/Sema/constexpr-void-cast.c b/clang/test/Sema/constexpr-void-cast.c index 91e4027f67fe387..2ffc59f509b4b72 100644 --- a/clang/test/Sema/constexpr-void-cast.c +++ b/clang/test/Sema/constexpr-void-cast.c @@ -1,8 +1,12 @@ // RUN: %clang_cc1 -x c -fsyntax-only %s -verify=c -std=c11 // RUN: %clang_cc1 -x c -fsyntax-only %s -pedantic -verify=c-pedantic -std=c11 +// RUN: %clang_cc1 -x c -fsyntax-only %s -verify=c -std=c11 -fexperimental-new-constant-interpreter +// RUN: %clang_cc1 -x c -fsyntax-only %s -pedantic -verify=c-pedantic -std=c11 -fexperimental-new-constant-interpreter // // RUN: %clang_cc1 -x c++ -fsyntax-only %s -verify=cxx // RUN: %clang_cc1 -x c++ -fsyntax-only %s -pedantic -verify=cxx-pedantic +// RUN: %clang_cc1 -x c++ -fsyntax-only %s -verify=cxx -fexperimental-new-constant-interpreter +// RUN: %clang_cc1 -x c++ -fsyntax-only %s -pedantic -verify=cxx-pedantic -fexperimental-new-constant-interpreter // c-no-diagnostics // cxx-no-diagnostics From 26224ca2de193dcfd21ede7a846c28ea892b77e3 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Bal=C3=A1zs=20K=C3=A9ri?= Date: Mon, 10 Jun 2024 10:21:14 +0200 Subject: [PATCH 19/36] [clang][analyzer] Improved PointerSubChecker (#93676) The checker is made more exact (only pointer into array is allowed, check array index) and more tests are added. --- .../Checkers/PointerSubChecker.cpp | 111 +++++++++++++--- clang/test/Analysis/casts.c | 4 + clang/test/Analysis/pointer-sub.c | 120 ++++++++++++++++++ clang/test/Analysis/ptr-arith.c | 14 +- 4 files changed, 222 insertions(+), 27 deletions(-) create mode 100644 clang/test/Analysis/pointer-sub.c diff --git a/clang/lib/StaticAnalyzer/Checkers/PointerSubChecker.cpp b/clang/lib/StaticAnalyzer/Checkers/PointerSubChecker.cpp index 2438cf30b39b594..b73534136fdf020 100644 --- a/clang/lib/StaticAnalyzer/Checkers/PointerSubChecker.cpp +++ b/clang/lib/StaticAnalyzer/Checkers/PointerSubChecker.cpp @@ -17,6 +17,7 @@ #include "clang/StaticAnalyzer/Core/Checker.h" #include "clang/StaticAnalyzer/Core/CheckerManager.h" #include "clang/StaticAnalyzer/Core/PathSensitive/CheckerContext.h" +#include "clang/StaticAnalyzer/Core/PathSensitive/DynamicExtent.h" #include "llvm/ADT/StringRef.h" using namespace clang; @@ -26,16 +27,88 @@ namespace { class PointerSubChecker : public Checker< check::PreStmt > { const BugType BT{this, "Pointer subtraction"}; + const llvm::StringLiteral Msg_MemRegionDifferent = + "Subtraction of two pointers that do not point into the same array " + "is undefined behavior."; + const llvm::StringLiteral Msg_LargeArrayIndex = + "Using an array index greater than the array size at pointer subtraction " + "is undefined behavior."; + const llvm::StringLiteral Msg_NegativeArrayIndex = + "Using a negative array index at pointer subtraction " + "is undefined behavior."; + const llvm::StringLiteral Msg_BadVarIndex = + "Indexing the address of a variable with other than 1 at this place " + "is undefined behavior."; + + bool checkArrayBounds(CheckerContext &C, const Expr *E, + const ElementRegion *ElemReg, + const MemRegion *Reg) const; + void reportBug(CheckerContext &C, const Expr *E, + const llvm::StringLiteral &Msg) const; public: void checkPreStmt(const BinaryOperator *B, CheckerContext &C) const; }; } +bool PointerSubChecker::checkArrayBounds(CheckerContext &C, const Expr *E, + const ElementRegion *ElemReg, + const MemRegion *Reg) const { + if (!ElemReg) + return true; + + ProgramStateRef State = C.getState(); + const MemRegion *SuperReg = ElemReg->getSuperRegion(); + SValBuilder &SVB = C.getSValBuilder(); + + if (SuperReg == Reg) { + if (const llvm::APSInt *I = SVB.getKnownValue(State, ElemReg->getIndex()); + I && (!I->isOne() && !I->isZero())) + reportBug(C, E, Msg_BadVarIndex); + return false; + } + + DefinedOrUnknownSVal ElemCount = + getDynamicElementCount(State, SuperReg, SVB, ElemReg->getElementType()); + auto IndexTooLarge = SVB.evalBinOp(C.getState(), BO_GT, ElemReg->getIndex(), + ElemCount, SVB.getConditionType()) + .getAs(); + if (IndexTooLarge) { + ProgramStateRef S1, S2; + std::tie(S1, S2) = C.getState()->assume(*IndexTooLarge); + if (S1 && !S2) { + reportBug(C, E, Msg_LargeArrayIndex); + return false; + } + } + auto IndexTooSmall = SVB.evalBinOp(State, BO_LT, ElemReg->getIndex(), + SVB.makeZeroVal(SVB.getArrayIndexType()), + SVB.getConditionType()) + .getAs(); + if (IndexTooSmall) { + ProgramStateRef S1, S2; + std::tie(S1, S2) = State->assume(*IndexTooSmall); + if (S1 && !S2) { + reportBug(C, E, Msg_NegativeArrayIndex); + return false; + } + } + return true; +} + +void PointerSubChecker::reportBug(CheckerContext &C, const Expr *E, + const llvm::StringLiteral &Msg) const { + if (ExplodedNode *N = C.generateNonFatalErrorNode()) { + auto R = std::make_unique(BT, Msg, N); + R->addRange(E->getSourceRange()); + C.emitReport(std::move(R)); + } +} + void PointerSubChecker::checkPreStmt(const BinaryOperator *B, CheckerContext &C) const { // When doing pointer subtraction, if the two pointers do not point to the - // same memory chunk, emit a warning. + // same array, emit a warning. if (B->getOpcode() != BO_Sub) return; @@ -44,28 +117,36 @@ void PointerSubChecker::checkPreStmt(const BinaryOperator *B, const MemRegion *LR = LV.getAsRegion(); const MemRegion *RR = RV.getAsRegion(); - - if (!(LR && RR)) + if (!LR || !RR) return; - const MemRegion *BaseLR = LR->getBaseRegion(); - const MemRegion *BaseRR = RR->getBaseRegion(); + // Allow subtraction of identical pointers. + if (LR == RR) + return; - if (BaseLR == BaseRR) + // No warning if one operand is unknown. + if (isa(LR) || isa(RR)) return; - // Allow arithmetic on different symbolic regions. - if (isa(BaseLR) || isa(BaseRR)) + const auto *ElemLR = dyn_cast(LR); + const auto *ElemRR = dyn_cast(RR); + + if (!checkArrayBounds(C, B->getLHS(), ElemLR, RR)) + return; + if (!checkArrayBounds(C, B->getRHS(), ElemRR, LR)) return; - if (ExplodedNode *N = C.generateNonFatalErrorNode()) { - constexpr llvm::StringLiteral Msg = - "Subtraction of two pointers that do not point to the same memory " - "chunk may cause incorrect result."; - auto R = std::make_unique(BT, Msg, N); - R->addRange(B->getSourceRange()); - C.emitReport(std::move(R)); + if (ElemLR && ElemRR) { + const MemRegion *SuperLR = ElemLR->getSuperRegion(); + const MemRegion *SuperRR = ElemRR->getSuperRegion(); + if (SuperLR == SuperRR) + return; + // Allow arithmetic on different symbolic regions. + if (isa(SuperLR) || isa(SuperRR)) + return; } + + reportBug(C, B, Msg_MemRegionDifferent); } void ento::registerPointerSubChecker(CheckerManager &mgr) { diff --git a/clang/test/Analysis/casts.c b/clang/test/Analysis/casts.c index 30cd74be564fdae..7dad4edfa89b99a 100644 --- a/clang/test/Analysis/casts.c +++ b/clang/test/Analysis/casts.c @@ -138,7 +138,9 @@ void multiDimensionalArrayPointerCasts(void) { clang_analyzer_eval(y1 == y2); // expected-warning{{TRUE}} // FIXME: should be FALSE (i.e. equal pointers). + // FIXME: pointer subtraction warning might be incorrect clang_analyzer_eval(y1 - y2); // expected-warning{{UNKNOWN}} + // expected-warning@-1{{Subtraction of two pointers that do not point into the same array is undefined behavior}} // FIXME: should be TRUE (i.e. same symbol). clang_analyzer_eval(*y1 == *y2); // expected-warning{{UNKNOWN}} @@ -147,7 +149,9 @@ void multiDimensionalArrayPointerCasts(void) { clang_analyzer_eval(y1 == y3); // expected-warning{{TRUE}} // FIXME: should be FALSE (i.e. equal pointers). + // FIXME: pointer subtraction warning might be incorrect clang_analyzer_eval(y1 - y3); // expected-warning{{UNKNOWN}} + // expected-warning@-1{{Subtraction of two pointers that do not point into the same array is undefined behavior}} // FIXME: should be TRUE (i.e. same symbol). clang_analyzer_eval(*y1 == *y3); // expected-warning{{UNKNOWN}} diff --git a/clang/test/Analysis/pointer-sub.c b/clang/test/Analysis/pointer-sub.c new file mode 100644 index 000000000000000..9a446547e2868fc --- /dev/null +++ b/clang/test/Analysis/pointer-sub.c @@ -0,0 +1,120 @@ +// RUN: %clang_analyze_cc1 -analyzer-checker=alpha.core.PointerSub -verify %s + +void f1(void) { + int x, y, z[10]; + int d = &y - &x; // expected-warning{{Subtraction of two pointers that do not point into the same array is undefined behavior}} + d = z - &y; // expected-warning{{Subtraction of two pointers that do not point into the same array is undefined behavior}} + d = &x - &x; // no-warning (subtraction of any two identical pointers is allowed) + d = (long *)&x - (long *)&x; + d = (&x + 1) - &x; // no-warning ('&x' is like a single-element array) + d = &x - (&x + 1); // no-warning + d = (&x + 0) - &x; // no-warning + d = (&x - 1) - &x; // expected-warning{{Indexing the address of a variable with other than 1 at this place is undefined behavior}} + d = (&x + 2) - &x; // expected-warning{{Indexing the address of a variable with other than 1 at this place is undefined behavior}} + + d = (z + 9) - z; // no-warning (pointers to same array) + d = (z + 10) - z; // no-warning (pointer to "one after the end") + d = (z + 11) - z; // expected-warning{{Using an array index greater than the array size at pointer subtraction is undefined behavior}} + d = (z - 1) - z; // expected-warning{{Using a negative array index at pointer subtraction is undefined behavior}} +} + +void f2(void) { + int a[10], b[10], c; + int *p = &a[2]; + int *q = &a[8]; + int d = q - p; // no-warning (pointers into the same array) + + q = &b[3]; + d = q - p; // expected-warning{{Subtraction of two pointers that}} + + q = a + 10; + d = q - p; // no warning (use of pointer to one after the end is allowed) + q = a + 11; + d = q - a; // expected-warning{{Using an array index greater than the array size at pointer subtraction is undefined behavior}} + + d = &a[4] - a; // no-warning + d = &a[2] - p; // no-warning + d = &c - p; // expected-warning{{Subtraction of two pointers that}} + + d = (int *)((char *)(&a[4]) + sizeof(int)) - &a[4]; // no-warning (pointers into the same array data) + d = (int *)((char *)(&a[4]) + 1) - &a[4]; // expected-warning{{Subtraction of two pointers that}} +} + +void f3(void) { + int a[3][4]; + int d; + + d = &(a[2]) - &(a[1]); + d = a[2] - a[1]; // expected-warning{{Subtraction of two pointers that}} + d = a[1] - a[1]; + d = &(a[1][2]) - &(a[1][0]); + d = &(a[1][2]) - &(a[0][0]); // expected-warning{{Subtraction of two pointers that}} + + d = (int *)((char *)(&a[2][2]) + sizeof(int)) - &a[2][2]; // expected-warning{{Subtraction of two pointers that}} + d = (int *)((char *)(&a[2][2]) + 1) - &a[2][2]; // expected-warning{{Subtraction of two pointers that}} + d = (int (*)[4])((char *)&a[2] + sizeof(int (*)[4])) - &a[2]; // expected-warning{{Subtraction of two pointers that}} + d = (int (*)[4])((char *)&a[2] + 1) - &a[2]; // expected-warning{{Subtraction of two pointers that}} +} + +void f4(void) { + int n = 4, m = 3; + int a[n][m]; + int (*p)[m] = a; // p == &a[0] + p += 1; // p == &a[1] + + // FIXME: This is a known problem with -Wpointer-arith (https://github.com/llvm/llvm-project/issues/28328) + int d = p - a; // d == 1 // expected-warning{{subtraction of pointers to type 'int[m]' of zero size has undefined behavior}} + + // FIXME: This is a known problem with -Wpointer-arith (https://github.com/llvm/llvm-project/issues/28328) + d = &(a[2]) - &(a[1]); // expected-warning{{subtraction of pointers to type 'int[m]' of zero size has undefined behavior}} + + d = a[2] - a[1]; // expected-warning{{Subtraction of two pointers that}} +} + +typedef struct { + int a; + int b; + int c[10]; + int d[10]; +} S; + +void f5(void) { + S s; + int y; + int d; + + d = &s.b - &s.a; // expected-warning{{Subtraction of two pointers that}} + d = &s.c[0] - &s.a; // expected-warning{{Subtraction of two pointers that}} + d = &s.b - &y; // expected-warning{{Subtraction of two pointers that}} + d = &s.c[3] - &s.c[2]; + d = &s.d[3] - &s.c[2]; // expected-warning{{Subtraction of two pointers that}} + d = s.d - s.c; // expected-warning{{Subtraction of two pointers that}} + + S sa[10]; + d = &sa[2] - &sa[1]; + d = &sa[2].a - &sa[1].b; // expected-warning{{Subtraction of two pointers that}} +} + +void f6(void) { + long long l; + char *a1 = (char *)&l; + int d = a1[3] - l; + + long long la1[3]; + long long la2[3]; + char *pla1 = (char *)la1; + char *pla2 = (char *)la2; + d = pla1[1] - pla1[0]; + d = (long long *)&pla1[1] - &l; // expected-warning{{Subtraction of two pointers that}} + d = &pla2[3] - &pla1[3]; // expected-warning{{Subtraction of two pointers that}} +} + +void f7(int *p) { + int a[10]; + int d = &a[10] - p; // no-warning ('p' is unknown, even if it cannot point into 'a') +} + +void f8(int n) { + int a[10]; + int d = a[n] - a[0]; +} diff --git a/clang/test/Analysis/ptr-arith.c b/clang/test/Analysis/ptr-arith.c index 40c8188704e811b..f99dfabb0736666 100644 --- a/clang/test/Analysis/ptr-arith.c +++ b/clang/test/Analysis/ptr-arith.c @@ -1,5 +1,5 @@ -// RUN: %clang_analyze_cc1 -analyzer-checker=alpha.core.FixedAddr,alpha.core.PointerArithm,alpha.core.PointerSub,debug.ExprInspection -Wno-pointer-to-int-cast -verify -triple x86_64-apple-darwin9 -Wno-tautological-pointer-compare -analyzer-config eagerly-assume=false %s -// RUN: %clang_analyze_cc1 -analyzer-checker=alpha.core.FixedAddr,alpha.core.PointerArithm,alpha.core.PointerSub,debug.ExprInspection -Wno-pointer-to-int-cast -verify -triple i686-apple-darwin9 -Wno-tautological-pointer-compare -analyzer-config eagerly-assume=false %s +// RUN: %clang_analyze_cc1 -analyzer-checker=alpha.core.FixedAddr,alpha.core.PointerArithm,debug.ExprInspection -Wno-pointer-to-int-cast -verify -triple x86_64-apple-darwin9 -Wno-tautological-pointer-compare -analyzer-config eagerly-assume=false %s +// RUN: %clang_analyze_cc1 -analyzer-checker=alpha.core.FixedAddr,alpha.core.PointerArithm,debug.ExprInspection -Wno-pointer-to-int-cast -verify -triple i686-apple-darwin9 -Wno-tautological-pointer-compare -analyzer-config eagerly-assume=false %s void clang_analyzer_eval(int); void clang_analyzer_dump(int); @@ -35,16 +35,6 @@ domain_port (const char *domain_b, const char *domain_e, return port; } -void f3(void) { - int x, y; - int d = &y - &x; // expected-warning{{Subtraction of two pointers that do not point to the same memory chunk may cause incorrect result}} - - int a[10]; - int *p = &a[2]; - int *q = &a[8]; - d = q-p; // no-warning -} - void f4(void) { int *p; p = (int*) 0x10000; // expected-warning{{Using a fixed address is not portable because that address will probably not be valid in all environments or platforms}} From d732a3298a70d83e5571c594de9be63df85668a7 Mon Sep 17 00:00:00 2001 From: Orlando Cazalet-Hyams Date: Mon, 10 Jun 2024 09:21:40 +0100 Subject: [PATCH 20/36] [RemoveDIs] C API: Add before-dbg-record versions of IRBuilder position funcs (#92417) Add `LLVMPositionBuilderBeforeDbgRecords` and `LLVMPositionBuilderBeforeInstrAndDbgRecords` to `llvm/include/llvm-c/Core.h` which behave the same as `LLVMPositionBuilder` and `LVMPositionBuilderBefore` except that the position is set before debug records attached to the target instruction (the existing functions set the insertion point to after any attached debug records). More info on debug records and the migration towards using them can be found here: https://llvm.org/docs/RemoveDIsDebugInfo.html The distinction is important in some situations. An important example is when inserting a phi before another instruction which has debug records attached to it (these come "before" the instruction). Inserting before the instruction but after the debug records would result in having debug records before a phi, which is illegal. That results in an assertion failure: `llvm/lib/IR/Instruction.cpp:166: Assertion '!isa(this) && "Inserting PHI after debug-records!"' failed.` In llvm (C++) we've added bit to instruction iterators that carries around the extra information. Adding dedicated functions seemed like the least invasive and least suprising way to update the C API. Update llvm/tools/llvm-c-test/debuginfo.c to test this functionality. Update the OCaml bindings, the migration docs and release notes. --- llvm/bindings/ocaml/llvm/llvm.ml | 5 ++++ llvm/bindings/ocaml/llvm/llvm.mli | 12 +++++++++ llvm/bindings/ocaml/llvm/llvm_ocaml.c | 12 +++++++++ llvm/docs/ReleaseNotes.rst | 2 ++ llvm/docs/RemoveDIsDebugInfo.md | 14 +++++++++- llvm/include/llvm-c/Core.h | 19 +++++++++++++ llvm/lib/IR/Core.cpp | 27 ++++++++++++++++--- llvm/test/Bindings/OCaml/core.ml | 2 +- llvm/test/Bindings/llvm-c/debug_info.ll | 3 +++ .../Bindings/llvm-c/debug_info_new_format.ll | 3 +++ llvm/tools/llvm-c-test/debuginfo.c | 19 ++++++++++++- 11 files changed, 111 insertions(+), 7 deletions(-) diff --git a/llvm/bindings/ocaml/llvm/llvm.ml b/llvm/bindings/ocaml/llvm/llvm.ml index 7eeaae49e105904..7bfaf8656ab3f2a 100644 --- a/llvm/bindings/ocaml/llvm/llvm.ml +++ b/llvm/bindings/ocaml/llvm/llvm.ml @@ -1135,6 +1135,9 @@ external delete_instruction : llvalue -> unit = "llvm_delete_instruction" external builder : llcontext -> llbuilder = "llvm_builder" external position_builder : (llbasicblock, llvalue) llpos -> llbuilder -> unit = "llvm_position_builder" +external position_builder_before_dbg_records : (llbasicblock, llvalue) llpos -> + llbuilder -> unit + = "llvm_position_builder_before_dbg_records" external insertion_block : llbuilder -> llbasicblock = "llvm_insertion_block" external insert_into_builder : llvalue -> string -> llbuilder -> unit = "llvm_insert_into_builder" @@ -1148,6 +1151,8 @@ let builder_before context i = builder_at context (Before i) let builder_at_end context bb = builder_at context (At_end bb) let position_before i = position_builder (Before i) +let position_before_dbg_records i = + position_builder_before_dbg_records (Before i) let position_at_end bb = position_builder (At_end bb) diff --git a/llvm/bindings/ocaml/llvm/llvm.mli b/llvm/bindings/ocaml/llvm/llvm.mli index 36cc095dc66e3de..89b894b35849da8 100644 --- a/llvm/bindings/ocaml/llvm/llvm.mli +++ b/llvm/bindings/ocaml/llvm/llvm.mli @@ -1874,10 +1874,22 @@ val builder_at_end : llcontext -> llbasicblock -> llbuilder See the constructor for [llvm::LLVMBuilder]. *) val position_builder : (llbasicblock, llvalue) llpos -> llbuilder -> unit +(** [position_builder_before_dbg_records ip bb before_dbg_records] moves the + instruction builder [bb] to the position [ip], before any debug records + there. + See the constructor for [llvm::LLVMBuilder]. *) +val position_builder_before_dbg_records : (llbasicblock, llvalue) llpos -> + llbuilder -> unit + (** [position_before ins b] moves the instruction builder [b] to before the instruction [isn]. See the method [llvm::LLVMBuilder::SetInsertPoint]. *) val position_before : llvalue -> llbuilder -> unit +(** [position_before_dbg_records ins b] moves the instruction builder [b] + to before the instruction [isn] and any debug records attached to it. + See the method [llvm::LLVMBuilder::SetInsertPoint]. *) +val position_before_dbg_records : llvalue -> llbuilder -> unit + (** [position_at_end bb b] moves the instruction builder [b] to the end of the basic block [bb]. See the method [llvm::LLVMBuilder::SetInsertPoint]. *) val position_at_end : llbasicblock -> llbuilder -> unit diff --git a/llvm/bindings/ocaml/llvm/llvm_ocaml.c b/llvm/bindings/ocaml/llvm/llvm_ocaml.c index 26a3ac21501d236..3976a96a6da9b43 100644 --- a/llvm/bindings/ocaml/llvm/llvm_ocaml.c +++ b/llvm/bindings/ocaml/llvm/llvm_ocaml.c @@ -2004,6 +2004,18 @@ value llvm_builder(value C) { return alloc_builder(LLVMCreateBuilderInContext(Context_val(C))); } +/* (llbasicblock, llvalue) llpos -> llbuilder -> unit */ +value llvm_position_builder_before_dbg_records(value Pos, value B) { + if (Tag_val(Pos) == 0) { + LLVMBasicBlockRef BB = BasicBlock_val(Field(Pos, 0)); + LLVMPositionBuilderAtEnd(Builder_val(B), BB); + } else { + LLVMValueRef I = Value_val(Field(Pos, 0)); + LLVMPositionBuilderBeforeInstrAndDbgRecords(Builder_val(B), I); + } + return Val_unit; +} + /* (llbasicblock, llvalue) llpos -> llbuilder -> unit */ value llvm_position_builder(value Pos, value B) { if (Tag_val(Pos) == 0) { diff --git a/llvm/docs/ReleaseNotes.rst b/llvm/docs/ReleaseNotes.rst index e1e1652d1e0a127..f86f9430201f8aa 100644 --- a/llvm/docs/ReleaseNotes.rst +++ b/llvm/docs/ReleaseNotes.rst @@ -214,6 +214,8 @@ Changes to the C API * ``LLVMConstICmp`` * ``LLVMConstFCmp`` +* Added ``LLVMPositionBuilderBeforeDbgRecords`` and ``LLVMPositionBuilderBeforeInstrAndDbgRecords``. Same as ``LLVMPositionBuilder`` and ``LLVMPositionBuilderBefore`` except the insertion position is set to before the debug records that precede the target instruction. See the `debug info migration guide `_ for more info. ``LLVMPositionBuilder`` and ``LLVMPositionBuilderBefore`` are unchanged; they insert before the indicated instruction but after any attached debug records. + Changes to the CodeGen infrastructure ------------------------------------- diff --git a/llvm/docs/RemoveDIsDebugInfo.md b/llvm/docs/RemoveDIsDebugInfo.md index 3cdf63c63b5a0bb..ef11c8ed8b43359 100644 --- a/llvm/docs/RemoveDIsDebugInfo.md +++ b/llvm/docs/RemoveDIsDebugInfo.md @@ -36,7 +36,7 @@ For a more in-depth overview of how to update existing code to support debug rec # C-API changes -All the functions that have been added are temporary and will be deprecated in the future. The intention is that they'll help downstream projects adapt during the transition period. +Some new functions that have been added are temporary and will be deprecated in the future. The intention is that they'll help downstream projects adapt during the transition period. ``` New functions (all to be deprecated) @@ -60,8 +60,20 @@ LLVMDIBuilderInsertDeclareBefore # Insert a debug record (new debug info forma LLVMDIBuilderInsertDeclareAtEnd # Same as above. LLVMDIBuilderInsertDbgValueBefore # Same as above. LLVMDIBuilderInsertDbgValueAtEnd # Same as above. + +New functions (no plans to deprecate) +---------------------------------- +LLVMPositionBuilderBeforeDbgRecords # See info below. +LLVMPositionBuilderBeforeInstrAndDbgRecords # See info below. ``` +`LLVMPositionBuilderBeforeDbgRecords` and `LLVMPositionBuilderBeforeInstrAndDbgRecords` behave the same as `LLVMPositionBuilder` and `LLVMPositionBuilderBefore` except the insertion position is set before the debug records that precede the target instruction. Note that this doesn't mean that debug intrinsics before the chosen instruction are skipped, only debug records (which unlike debug records are not themselves instructions). + +If you don't know which function to call then follow this rule: +If you are trying to insert at the start of a block, or purposfully skip debug intrinsics to determine the insertion point for any other reason, then call the new functions. + +`LLVMPositionBuilder` and `LLVMPositionBuilderBefore` are unchanged. They insert before the indicated instruction but after any attached debug records. + # The new "Debug Record" model Below is a brief overview of the new representation that replaces debug intrinsics; for an instructive guide on updating old code, see [here](#how-to-update-existing-code). diff --git a/llvm/include/llvm-c/Core.h b/llvm/include/llvm-c/Core.h index 7d6b8c76079749e..cd9d7e0f843d050 100644 --- a/llvm/include/llvm-c/Core.h +++ b/llvm/include/llvm-c/Core.h @@ -3951,9 +3951,28 @@ const unsigned *LLVMGetIndices(LLVMValueRef Inst); LLVMBuilderRef LLVMCreateBuilderInContext(LLVMContextRef C); LLVMBuilderRef LLVMCreateBuilder(void); +/** + * Set the builder poisiton before Instr but after any attached debug records, + * or if Instr is null set the position to the end of Block. + */ void LLVMPositionBuilder(LLVMBuilderRef Builder, LLVMBasicBlockRef Block, LLVMValueRef Instr); +/** + * Set the builder poisiton before Instr and any attached debug records, + * or if Instr is null set the position to the end of Block. + */ +void LLVMPositionBuilderBeforeDbgRecords(LLVMBuilderRef Builder, + LLVMBasicBlockRef Block, + LLVMValueRef Inst); +/** + * Set the builder poisiton before Instr but after any attached debug records. + */ void LLVMPositionBuilderBefore(LLVMBuilderRef Builder, LLVMValueRef Instr); +/** + * Set the builder poisiton before Instr and any attached debug records. + */ +void LLVMPositionBuilderBeforeInstrAndDbgRecords(LLVMBuilderRef Builder, + LLVMValueRef Instr); void LLVMPositionBuilderAtEnd(LLVMBuilderRef Builder, LLVMBasicBlockRef Block); LLVMBasicBlockRef LLVMGetInsertBlock(LLVMBuilderRef Builder); void LLVMClearInsertionPosition(LLVMBuilderRef Builder); diff --git a/llvm/lib/IR/Core.cpp b/llvm/lib/IR/Core.cpp index 4a508a022cf7eec..3a91a4eb6e0d106 100644 --- a/llvm/lib/IR/Core.cpp +++ b/llvm/lib/IR/Core.cpp @@ -3137,16 +3137,35 @@ LLVMBuilderRef LLVMCreateBuilder(void) { return LLVMCreateBuilderInContext(LLVMGetGlobalContext()); } +static void LLVMPositionBuilderImpl(IRBuilder<> *Builder, BasicBlock *Block, + Instruction *Instr, bool BeforeDbgRecords) { + BasicBlock::iterator I = Instr ? Instr->getIterator() : Block->end(); + I.setHeadBit(BeforeDbgRecords); + Builder->SetInsertPoint(Block, I); +} + void LLVMPositionBuilder(LLVMBuilderRef Builder, LLVMBasicBlockRef Block, LLVMValueRef Instr) { - BasicBlock *BB = unwrap(Block); - auto I = Instr ? unwrap(Instr)->getIterator() : BB->end(); - unwrap(Builder)->SetInsertPoint(BB, I); + return LLVMPositionBuilderImpl(unwrap(Builder), unwrap(Block), + unwrap(Instr), false); +} + +void LLVMPositionBuilderBeforeDbgRecords(LLVMBuilderRef Builder, + LLVMBasicBlockRef Block, + LLVMValueRef Instr) { + return LLVMPositionBuilderImpl(unwrap(Builder), unwrap(Block), + unwrap(Instr), true); } void LLVMPositionBuilderBefore(LLVMBuilderRef Builder, LLVMValueRef Instr) { Instruction *I = unwrap(Instr); - unwrap(Builder)->SetInsertPoint(I->getParent(), I->getIterator()); + return LLVMPositionBuilderImpl(unwrap(Builder), I->getParent(), I, false); +} + +void LLVMPositionBuilderBeforeInstrAndDbgRecords(LLVMBuilderRef Builder, + LLVMValueRef Instr) { + Instruction *I = unwrap(Instr); + return LLVMPositionBuilderImpl(unwrap(Builder), I->getParent(), I, true); } void LLVMPositionBuilderAtEnd(LLVMBuilderRef Builder, LLVMBasicBlockRef Block) { diff --git a/llvm/test/Bindings/OCaml/core.ml b/llvm/test/Bindings/OCaml/core.ml index 36f4c468d99f48d..923a3541e239c39 100644 --- a/llvm/test/Bindings/OCaml/core.ml +++ b/llvm/test/Bindings/OCaml/core.ml @@ -1150,7 +1150,7 @@ let test_builder () = (* CHECK: ret{{.*}}P1 *) let ret = build_ret p1 atentry in - position_before ret atentry + position_before_dbg_records ret atentry end; (* see test/Feature/exception.ll *) diff --git a/llvm/test/Bindings/llvm-c/debug_info.ll b/llvm/test/Bindings/llvm-c/debug_info.ll index 9358bac59bd21e6..71986fbb2348ff4 100644 --- a/llvm/test/Bindings/llvm-c/debug_info.ll +++ b/llvm/test/Bindings/llvm-c/debug_info.ll @@ -10,7 +10,10 @@ ; CHECK-NEXT: call void @llvm.dbg.declare(metadata i64 0, metadata !40, metadata !DIExpression()), !dbg !43 ; CHECK-NEXT: br label %vars ; CHECK: vars: +; CHECK-NEXT: %p1 = phi i64 [ 0, %entry ] +; CHECK-NEXT: %p2 = phi i64 [ 0, %entry ] ; CHECK-NEXT: call void @llvm.dbg.value(metadata i64 0, metadata !41, metadata !DIExpression(DW_OP_constu, 0, DW_OP_stack_value)), !dbg !44 +; CHECK-NEXT: %a = add i64 %p1, %p2 ; CHECK-NEXT: ret i64 0 ; CHECK-NEXT: } diff --git a/llvm/test/Bindings/llvm-c/debug_info_new_format.ll b/llvm/test/Bindings/llvm-c/debug_info_new_format.ll index 05b6ef4de9adbb4..1b6f2c4ea403aab 100644 --- a/llvm/test/Bindings/llvm-c/debug_info_new_format.ll +++ b/llvm/test/Bindings/llvm-c/debug_info_new_format.ll @@ -11,7 +11,10 @@ ; CHECK-NEXT: #dbg_declare(i64 0, !40, !DIExpression(), !43) ; CHECK-NEXT: br label %vars ; CHECK: vars: +; CHECK-NEXT: %p1 = phi i64 [ 0, %entry ] +; CHECK-NEXT: %p2 = phi i64 [ 0, %entry ] ; CHECK-NEXT: #dbg_value(i64 0, !41, !DIExpression(DW_OP_constu, 0, DW_OP_stack_value), !44) +; CHECK-NEXT: %a = add i64 %p1, %p2 ; CHECK-NEXT: ret i64 0 ; CHECK-NEXT: } diff --git a/llvm/tools/llvm-c-test/debuginfo.c b/llvm/tools/llvm-c-test/debuginfo.c index 35c65f885af329e..032657203b5f2ee 100644 --- a/llvm/tools/llvm-c-test/debuginfo.c +++ b/llvm/tools/llvm-c-test/debuginfo.c @@ -228,7 +228,24 @@ int llvm_test_dibuilder(bool NewDebugInfoFormat) { LLVMPositionBuilderAtEnd(Builder, FooVarBlock); LLVMTypeRef I64 = LLVMInt64TypeInContext(Ctx); LLVMValueRef Zero = LLVMConstInt(I64, 0, false); - LLVMBuildRet(Builder, Zero); + LLVMValueRef Ret = LLVMBuildRet(Builder, Zero); + + // Insert a `phi` before the `ret`. In the new debug info mode we need to + // be careful to insert before debug records too, else the debug records + // will come before the `phi` (and be absorbed onto it) which is an invalid + // state. + LLVMValueRef InsertPos = LLVMGetFirstInstruction(FooVarBlock); + LLVMPositionBuilderBeforeInstrAndDbgRecords(Builder, InsertPos); + LLVMValueRef Phi1 = LLVMBuildPhi(Builder, I64, "p1"); + LLVMAddIncoming(Phi1, &Zero, &FooEntryBlock, 1); + // Do the same again using the other position-setting function. + LLVMPositionBuilderBeforeDbgRecords(Builder, FooVarBlock, InsertPos); + LLVMValueRef Phi2 = LLVMBuildPhi(Builder, I64, "p2"); + LLVMAddIncoming(Phi2, &Zero, &FooEntryBlock, 1); + // Insert a non-phi before the `ret` but not before the debug records to + // test that works as expected. + LLVMPositionBuilder(Builder, FooVarBlock, Ret); + LLVMBuildAdd(Builder, Phi1, Phi2, "a"); char *MStr = LLVMPrintModuleToString(M); puts(MStr); From 0257f9cde540f265da38ef14fd395f76d3e08bb1 Mon Sep 17 00:00:00 2001 From: jeanPerier Date: Mon, 10 Jun 2024 10:22:53 +0200 Subject: [PATCH 21/36] [flang] lower SHAPE with assumed-rank arguments (#94812) Allocate result statically on the stack (using max rank) and use the runtime to fill it in correctly. --- .../flang/Optimizer/Builder/Runtime/Inquiry.h | 6 ++ flang/lib/Optimizer/Builder/IntrinsicCall.cpp | 34 ++++++++++- .../lib/Optimizer/Builder/Runtime/Inquiry.cpp | 14 +++++ .../Lower/HLFIR/assumed-rank-inquiries-3.f90 | 56 +++++++++++++++++++ 4 files changed, 108 insertions(+), 2 deletions(-) create mode 100644 flang/test/Lower/HLFIR/assumed-rank-inquiries-3.f90 diff --git a/flang/include/flang/Optimizer/Builder/Runtime/Inquiry.h b/flang/include/flang/Optimizer/Builder/Runtime/Inquiry.h index 132592a0197f8fd..5f14d7781004b3c 100644 --- a/flang/include/flang/Optimizer/Builder/Runtime/Inquiry.h +++ b/flang/include/flang/Optimizer/Builder/Runtime/Inquiry.h @@ -32,6 +32,12 @@ mlir::Value genLboundDim(fir::FirOpBuilder &builder, mlir::Location loc, void genUbound(fir::FirOpBuilder &builder, mlir::Location loc, mlir::Value resultBox, mlir::Value array, mlir::Value kind); +/// Generate call to `Shape` runtime routine. +/// First argument is a raw pointer to the result array storage that +/// must be allocated by the caller. +void genShape(fir::FirOpBuilder &builder, mlir::Location loc, + mlir::Value resultAddr, mlir::Value arrayt, mlir::Value kind); + /// Generate call to `Size` runtime routine. This routine is a specialized /// version when the DIM argument is not specified by the user. mlir::Value genSize(fir::FirOpBuilder &builder, mlir::Location loc, diff --git a/flang/lib/Optimizer/Builder/IntrinsicCall.cpp b/flang/lib/Optimizer/Builder/IntrinsicCall.cpp index 861b26de06370ef..b3e1ee3da3a7786 100644 --- a/flang/lib/Optimizer/Builder/IntrinsicCall.cpp +++ b/flang/lib/Optimizer/Builder/IntrinsicCall.cpp @@ -5992,15 +5992,45 @@ mlir::Value IntrinsicLibrary::genSetExponent(mlir::Type resultType, fir::getBase(args[1]))); } +/// Generate runtime call to inquire about all the bounds/extents of an +/// assumed-rank array. +template +static fir::ExtendedValue genAssumedRankBoundInquiry( + fir::FirOpBuilder &builder, mlir::Location loc, mlir::Type resultType, + llvm::ArrayRef args, int kindPos, Func genRtCall) { + const fir::ExtendedValue &array = args[0]; + // Allocate an array with the maximum rank, that is big enough to hold the + // result but still "small" (15 elements). Static size alloca make stack + // analysis/manipulation easier. + mlir::Type resultElementType = fir::unwrapSequenceType(resultType); + mlir::Type allocSeqType = + fir::SequenceType::get({Fortran::common::maxRank}, resultElementType); + mlir::Value resultStorage = builder.createTemporary(loc, allocSeqType); + mlir::Value arrayBox = builder.createBox(loc, array); + mlir::Value kind = isStaticallyAbsent(args, kindPos) + ? builder.createIntegerConstant( + loc, builder.getI32Type(), + builder.getKindMap().defaultIntegerKind()) + : fir::getBase(args[kindPos]); + genRtCall(builder, loc, resultStorage, arrayBox, kind); + mlir::Type baseType = + fir::ReferenceType::get(builder.getVarLenSeqTy(resultElementType)); + mlir::Value resultBase = builder.createConvert(loc, baseType, resultStorage); + mlir::Value rank = + builder.create(loc, builder.getIndexType(), arrayBox); + return fir::ArrayBoxValue{resultBase, {rank}}; +} + // SHAPE fir::ExtendedValue IntrinsicLibrary::genShape(mlir::Type resultType, llvm::ArrayRef args) { assert(args.size() >= 1); const fir::ExtendedValue &array = args[0]; + if (array.hasAssumedRank()) + return genAssumedRankBoundInquiry(builder, loc, resultType, args, + /*kindPos=*/1, fir::runtime::genShape); int rank = array.rank(); - if (rank == 0) - TODO(loc, "shape intrinsic lowering with assumed-rank source"); mlir::Type indexType = builder.getIndexType(); mlir::Type extentType = fir::unwrapSequenceType(resultType); mlir::Type seqType = fir::SequenceType::get( diff --git a/flang/lib/Optimizer/Builder/Runtime/Inquiry.cpp b/flang/lib/Optimizer/Builder/Runtime/Inquiry.cpp index 16f63bea4617a6b..34c4020b5907c9b 100644 --- a/flang/lib/Optimizer/Builder/Runtime/Inquiry.cpp +++ b/flang/lib/Optimizer/Builder/Runtime/Inquiry.cpp @@ -87,3 +87,17 @@ mlir::Value fir::runtime::genIsContiguous(fir::FirOpBuilder &builder, auto args = fir::runtime::createArguments(builder, loc, fTy, array); return builder.create(loc, isContiguousFunc, args).getResult(0); } + +void fir::runtime::genShape(fir::FirOpBuilder &builder, mlir::Location loc, + mlir::Value resultAddr, mlir::Value array, + mlir::Value kind) { + mlir::func::FuncOp func = + fir::runtime::getRuntimeFunc(loc, builder); + auto fTy = func.getFunctionType(); + auto sourceFile = fir::factory::locationToFilename(builder, loc); + auto sourceLine = + fir::factory::locationToLineNo(builder, loc, fTy.getInput(4)); + auto args = fir::runtime::createArguments( + builder, loc, fTy, resultAddr, array, kind, sourceFile, sourceLine); + builder.create(loc, func, args).getResult(0); +} diff --git a/flang/test/Lower/HLFIR/assumed-rank-inquiries-3.f90 b/flang/test/Lower/HLFIR/assumed-rank-inquiries-3.f90 new file mode 100644 index 000000000000000..bbeff5ff051915f --- /dev/null +++ b/flang/test/Lower/HLFIR/assumed-rank-inquiries-3.f90 @@ -0,0 +1,56 @@ +! Test shape lowering for assumed-rank +! RUN: bbc -emit-hlfir -o - %s -allow-assumed-rank | FileCheck %s + +subroutine test_shape(x) + real :: x(..) + call takes_integer_array(shape(x)) +end subroutine +! CHECK-LABEL: func.func @_QPtest_shape( +! CHECK: %[[VAL_1:.*]] = fir.alloca !fir.array<15xi32> +! CHECK: %[[VAL_4:.*]] = arith.constant 4 : i32 +! CHECK: %[[VAL_7:.*]] = fir.convert %[[VAL_1]] : (!fir.ref>) -> !fir.llvm_ptr +! CHECK: %[[VAL_8:.*]] = fir.convert %[[VAL_3:.*]] : (!fir.box>) -> !fir.box +! CHECK: %[[VAL_10:.*]] = fir.call @_FortranAShape(%[[VAL_7]], %[[VAL_8]], %[[VAL_4]], %{{.*}}, %{{.*}}) +! CHECK: %[[VAL_11:.*]] = fir.convert %[[VAL_1]] : (!fir.ref>) -> !fir.ref> +! CHECK: %[[VAL_12:.*]] = fir.box_rank %[[VAL_3]] : (!fir.box>) -> index +! CHECK: %[[VAL_13:.*]] = fir.shape %[[VAL_12]] : (index) -> !fir.shape<1> +! CHECK: %[[VAL_14:.*]]:2 = hlfir.declare %[[VAL_11]](%[[VAL_13]]) {uniq_name = ".tmp.intrinsic_result"} : (!fir.ref>, !fir.shape<1>) -> (!fir.box>, !fir.ref>) +! CHECK: %[[VAL_15:.*]] = arith.constant false +! CHECK: %[[VAL_16:.*]] = hlfir.as_expr %[[VAL_14]]#0 move %[[VAL_15]] : (!fir.box>, i1) -> !hlfir.expr +! CHECK: %[[VAL_17:.*]]:3 = hlfir.associate %[[VAL_16]](%[[VAL_13]]) {adapt.valuebyref} : (!hlfir.expr, !fir.shape<1>) -> (!fir.box>, !fir.ref>, i1) +! CHECK: fir.call @_QPtakes_integer_array(%[[VAL_17]]#1) fastmath : (!fir.ref>) -> () +! CHECK: hlfir.end_associate %[[VAL_17]]#1, %[[VAL_17]]#2 : !fir.ref>, i1 +! CHECK: hlfir.destroy %[[VAL_16]] : !hlfir.expr +! CHECK: return +! CHECK: } + +subroutine test_shape_kind(x) + real :: x(..) + call takes_integer8_array(shape(x, kind=8)) +end subroutine +! CHECK-LABEL: func.func @_QPtest_shape_kind( +! CHECK: %[[VAL_1:.*]] = fir.alloca !fir.array<15xi64> +! CHECK: %[[VAL_4:.*]] = arith.constant 8 : i32 +! CHECK: %[[VAL_7:.*]] = fir.convert %[[VAL_1]] : (!fir.ref>) -> !fir.llvm_ptr +! CHECK: %[[VAL_8:.*]] = fir.convert %[[VAL_3:.*]] : (!fir.box>) -> !fir.box +! CHECK: %[[VAL_10:.*]] = fir.call @_FortranAShape(%[[VAL_7]], %[[VAL_8]], %[[VAL_4]], %{{.*}}, %{{.*}}) +! CHECK: %[[VAL_11:.*]] = fir.convert %[[VAL_1]] : (!fir.ref>) -> !fir.ref> +! CHECK: %[[VAL_12:.*]] = fir.box_rank %[[VAL_3]] : (!fir.box>) -> index +! CHECK: %[[VAL_13:.*]] = fir.shape %[[VAL_12]] : (index) -> !fir.shape<1> +! CHECK: %[[VAL_14:.*]]:2 = hlfir.declare %[[VAL_11]](%[[VAL_13]]) {uniq_name = ".tmp.intrinsic_result"} : (!fir.ref>, !fir.shape<1>) -> (!fir.box>, !fir.ref>) + +subroutine test_shape_2(x) + real, pointer :: x(..) + call takes_integer_array(shape(x)) +end subroutine +! CHECK-LABEL: func.func @_QPtest_shape_2( +! CHECK: %[[VAL_1:.*]] = fir.alloca !fir.array<15xi32> +! CHECK: %[[VAL_4:.*]] = fir.load %[[VAL_3:.*]] : !fir.ref>>> +! CHECK: %[[VAL_5:.*]] = arith.constant 4 : i32 +! CHECK: %[[VAL_8:.*]] = fir.convert %[[VAL_1]] : (!fir.ref>) -> !fir.llvm_ptr +! CHECK: %[[VAL_9:.*]] = fir.convert %[[VAL_4]] : (!fir.box>>) -> !fir.box +! CHECK: %[[VAL_11:.*]] = fir.call @_FortranAShape(%[[VAL_8]], %[[VAL_9]], %[[VAL_5]], %{{.*}}, %{{.*}}) +! CHECK: %[[VAL_12:.*]] = fir.convert %[[VAL_1]] : (!fir.ref>) -> !fir.ref> +! CHECK: %[[VAL_13:.*]] = fir.box_rank %[[VAL_4]] : (!fir.box>>) -> index +! CHECK: %[[VAL_14:.*]] = fir.shape %[[VAL_13]] : (index) -> !fir.shape<1> +! CHECK: %[[VAL_15:.*]]:2 = hlfir.declare %[[VAL_12]](%[[VAL_14]]) {uniq_name = ".tmp.intrinsic_result"} : (!fir.ref>, !fir.shape<1>) -> (!fir.box>, !fir.ref>) From 0af2e75f8c688cc6fb7123c969c8edfa98a003fa Mon Sep 17 00:00:00 2001 From: Shivam Gupta Date: Mon, 10 Jun 2024 13:53:39 +0530 Subject: [PATCH 22/36] [lldb] Fix redundant condition in compression type check (NFC) (#94841) The `else if` condition for checking `m_compression_type` is redundant as it matches with a previous `if` condition, making the expression always false. Reported by cppcheck as a possible cut-and-paste error. Caught by cppcheck - lldb/source/Plugins/Process/gdb-remote/GDBRemoteCommunication.cpp:543:35: style: Expression is always false because 'else if' condition matches previous condition at line 535. [multiCondition] Fix #91222 --- .../Plugins/Process/gdb-remote/GDBRemoteCommunication.cpp | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/lldb/source/Plugins/Process/gdb-remote/GDBRemoteCommunication.cpp b/lldb/source/Plugins/Process/gdb-remote/GDBRemoteCommunication.cpp index 8a47eed3d7cbe1f..187370eb36cae46 100644 --- a/lldb/source/Plugins/Process/gdb-remote/GDBRemoteCommunication.cpp +++ b/lldb/source/Plugins/Process/gdb-remote/GDBRemoteCommunication.cpp @@ -539,9 +539,8 @@ bool GDBRemoteCommunication::DecompressPacket() { else if (m_compression_type == CompressionType::ZlibDeflate) scratchbuf_size = compression_decode_scratch_buffer_size (COMPRESSION_ZLIB); else if (m_compression_type == CompressionType::LZMA) - scratchbuf_size = compression_decode_scratch_buffer_size (COMPRESSION_LZMA); - else if (m_compression_type == CompressionType::LZFSE) - scratchbuf_size = compression_decode_scratch_buffer_size (COMPRESSION_LZFSE); + scratchbuf_size = + compression_decode_scratch_buffer_size(COMPRESSION_LZMA); if (scratchbuf_size > 0) { m_decompression_scratch = (void*) malloc (scratchbuf_size); m_decompression_scratch_type = m_compression_type; From 30bfab3694075cfec9409821be0cfe79ae5d0770 Mon Sep 17 00:00:00 2001 From: Shivam Gupta Date: Mon, 10 Jun 2024 13:53:55 +0530 Subject: [PATCH 23/36] [lldb] Remove redundant condition in watch mask check (NFC) (#94842) This issue is reported by cppcheck as a pointless test in the watch mask check. The `else if` condition is opposite to the previous `if` condition, making the expression always true. Caught by cppcheck - lldb/source/Plugins/Process/Linux/NativeRegisterContextLinux_arm.cpp:509:25: style: Expression is always true because 'else if' condition is opposite to previous condition at line 505. [multiCondition] Fix #91223 --- .../Plugins/Process/Linux/NativeRegisterContextLinux_arm.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/lldb/source/Plugins/Process/Linux/NativeRegisterContextLinux_arm.cpp b/lldb/source/Plugins/Process/Linux/NativeRegisterContextLinux_arm.cpp index 5ad2f7a8e9455b1..4668c25eab08365 100644 --- a/lldb/source/Plugins/Process/Linux/NativeRegisterContextLinux_arm.cpp +++ b/lldb/source/Plugins/Process/Linux/NativeRegisterContextLinux_arm.cpp @@ -506,7 +506,7 @@ uint32_t NativeRegisterContextLinux_arm::SetHardwareWatchpoint( return LLDB_INVALID_INDEX32; else if (watch_mask <= 0x02) size = 2; - else if (watch_mask <= 0x04) + else size = 4; addr = addr & (~0x03); From 38c01c3d91f8815468dedf67db9eaa0bd0c6ea7a Mon Sep 17 00:00:00 2001 From: Orlando Cazalet-Hyams Date: Mon, 10 Jun 2024 09:25:26 +0100 Subject: [PATCH 24/36] NFC fix typos from #92417 --- llvm/include/llvm-c/Core.h | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/llvm/include/llvm-c/Core.h b/llvm/include/llvm-c/Core.h index cd9d7e0f843d050..d4a10e927760521 100644 --- a/llvm/include/llvm-c/Core.h +++ b/llvm/include/llvm-c/Core.h @@ -3952,24 +3952,24 @@ const unsigned *LLVMGetIndices(LLVMValueRef Inst); LLVMBuilderRef LLVMCreateBuilderInContext(LLVMContextRef C); LLVMBuilderRef LLVMCreateBuilder(void); /** - * Set the builder poisiton before Instr but after any attached debug records, + * Set the builder position before Instr but after any attached debug records, * or if Instr is null set the position to the end of Block. */ void LLVMPositionBuilder(LLVMBuilderRef Builder, LLVMBasicBlockRef Block, LLVMValueRef Instr); /** - * Set the builder poisiton before Instr and any attached debug records, + * Set the builder position before Instr and any attached debug records, * or if Instr is null set the position to the end of Block. */ void LLVMPositionBuilderBeforeDbgRecords(LLVMBuilderRef Builder, LLVMBasicBlockRef Block, LLVMValueRef Inst); /** - * Set the builder poisiton before Instr but after any attached debug records. + * Set the builder position before Instr but after any attached debug records. */ void LLVMPositionBuilderBefore(LLVMBuilderRef Builder, LLVMValueRef Instr); /** - * Set the builder poisiton before Instr and any attached debug records. + * Set the builder position before Instr and any attached debug records. */ void LLVMPositionBuilderBeforeInstrAndDbgRecords(LLVMBuilderRef Builder, LLVMValueRef Instr); From 760d880ea602117aa2e6bba4cf31069f09225b4b Mon Sep 17 00:00:00 2001 From: Fabio D'Urso Date: Mon, 10 Jun 2024 10:39:17 +0200 Subject: [PATCH 25/36] [scudo] Apply filling when realloc shrinks and re-grows a block in-place (#93212) --- compiler-rt/lib/scudo/standalone/combined.h | 14 +++++++++++ .../scudo/standalone/tests/combined_test.cpp | 23 +++++++++++++++---- 2 files changed, 32 insertions(+), 5 deletions(-) diff --git a/compiler-rt/lib/scudo/standalone/combined.h b/compiler-rt/lib/scudo/standalone/combined.h index f9ed36581f8d3a8..73da686287747d8 100644 --- a/compiler-rt/lib/scudo/standalone/combined.h +++ b/compiler-rt/lib/scudo/standalone/combined.h @@ -565,6 +565,20 @@ class Allocator { storeSecondaryAllocationStackMaybe(Options, OldPtr, NewSize); } } + + // If we have reduced the size, set the extra bytes to the fill value + // so that we are ready to grow it again in the future. + if (NewSize < OldSize) { + const FillContentsMode FillContents = + TSDRegistry.getDisableMemInit() ? NoFill + : Options.getFillContentsMode(); + if (FillContents != NoFill) { + memset(reinterpret_cast(OldTaggedPtr) + NewSize, + FillContents == ZeroFill ? 0 : PatternFillByte, + OldSize - NewSize); + } + } + return OldTaggedPtr; } } diff --git a/compiler-rt/lib/scudo/standalone/tests/combined_test.cpp b/compiler-rt/lib/scudo/standalone/tests/combined_test.cpp index 1a36155bcd423a0..655dc87cbac64ef 100644 --- a/compiler-rt/lib/scudo/standalone/tests/combined_test.cpp +++ b/compiler-rt/lib/scudo/standalone/tests/combined_test.cpp @@ -447,19 +447,32 @@ SCUDO_TYPED_TEST(ScudoCombinedDeathTest, ReallocateSame) { // returns the same chunk. This requires that all the sizes we iterate on use // the same block size, but that should be the case for MaxSize - 64 with our // default class size maps. - constexpr scudo::uptr ReallocSize = + constexpr scudo::uptr InitialSize = TypeParam::Primary::SizeClassMap::MaxSize - 64; - void *P = Allocator->allocate(ReallocSize, Origin); const char Marker = 'A'; - memset(P, Marker, ReallocSize); + Allocator->setFillContents(scudo::PatternOrZeroFill); + + void *P = Allocator->allocate(InitialSize, Origin); + scudo::uptr CurrentSize = InitialSize; for (scudo::sptr Delta = -32; Delta < 32; Delta += 8) { + memset(P, Marker, CurrentSize); const scudo::uptr NewSize = - static_cast(static_cast(ReallocSize) + Delta); + static_cast(static_cast(InitialSize) + Delta); void *NewP = Allocator->reallocate(P, NewSize); EXPECT_EQ(NewP, P); - for (scudo::uptr I = 0; I < ReallocSize - 32; I++) + + // Verify that existing contents have been preserved. + for (scudo::uptr I = 0; I < scudo::Min(CurrentSize, NewSize); I++) EXPECT_EQ((reinterpret_cast(NewP))[I], Marker); + + // Verify that new bytes are set according to FillContentsMode. + for (scudo::uptr I = CurrentSize; I < NewSize; I++) { + EXPECT_EQ((reinterpret_cast(NewP))[I], + scudo::PatternFillByte); + } + checkMemoryTaggingMaybe(Allocator, NewP, NewSize, 0); + CurrentSize = NewSize; } Allocator->deallocate(P, Origin); } From a0faf79b164c98476a4e399bd96766013f826a11 Mon Sep 17 00:00:00 2001 From: jeanPerier Date: Mon, 10 Jun 2024 10:57:50 +0200 Subject: [PATCH 26/36] [flang][runtime] add LBOUND API for assumed-rank arrays (#94808) --- flang/include/flang/Runtime/inquiry.h | 4 ++- flang/runtime/inquiry.cpp | 11 ++++++++ flang/unittests/Runtime/Inquiry.cpp | 38 ++++++++++++++++++++++++++- 3 files changed, 51 insertions(+), 2 deletions(-) diff --git a/flang/include/flang/Runtime/inquiry.h b/flang/include/flang/Runtime/inquiry.h index 0a3cd51236fa39b..dde6e722ad6e955 100644 --- a/flang/include/flang/Runtime/inquiry.h +++ b/flang/include/flang/Runtime/inquiry.h @@ -24,9 +24,11 @@ extern "C" { std::int64_t RTDECL(LboundDim)(const Descriptor &array, int dim, const char *sourceFile = nullptr, int line = 0); -void RTDECL(Shape)(void *result, const Descriptor &array, int kind, +void RTDECL(Lbound)(void *result, const Descriptor &array, int kind, const char *sourceFile = nullptr, int line = 0); +void RTDECL(Shape)(void *result, const Descriptor &array, int kind, + const char *sourceFile = nullptr, int line = 0); std::int64_t RTDECL(Size)( const Descriptor &array, const char *sourceFile = nullptr, int line = 0); diff --git a/flang/runtime/inquiry.cpp b/flang/runtime/inquiry.cpp index 443e6291e5e2374..5ffd9755fd0e52d 100644 --- a/flang/runtime/inquiry.cpp +++ b/flang/runtime/inquiry.cpp @@ -96,5 +96,16 @@ void RTDEF(Shape)(void *result, const Descriptor &array, int kind, } } +void RTDEF(Lbound)(void *result, const Descriptor &array, int kind, + const char *sourceFile, int line) { + Terminator terminator{sourceFile, line}; + INTERNAL_CHECK(array.rank() <= common::maxRank); + for (SubscriptValue i{0}; i < array.rank(); ++i) { + const Dimension &dimension{array.GetDimension(i)}; + Fortran::runtime::ApplyIntegerKind( + kind, terminator, result, i, dimension.LowerBound()); + } +} + } // extern "C" } // namespace Fortran::runtime diff --git a/flang/unittests/Runtime/Inquiry.cpp b/flang/unittests/Runtime/Inquiry.cpp index 220ebb765f8c721..98a350d3ad98a1f 100644 --- a/flang/unittests/Runtime/Inquiry.cpp +++ b/flang/unittests/Runtime/Inquiry.cpp @@ -14,7 +14,7 @@ using namespace Fortran::runtime; using Fortran::common::TypeCategory; -TEST(Inquiry, Lbound) { +TEST(Inquiry, LboundDim) { // ARRAY 1 3 5 // 2 4 6 auto array{MakeArray( @@ -26,6 +26,42 @@ TEST(Inquiry, Lbound) { EXPECT_EQ(RTNAME(LboundDim)(*array, 2, __FILE__, __LINE__), std::int64_t{-1}); } +TEST(Inquiry, Lbound) { + // ARRAY 1 3 5 + // 2 4 6 + auto array{MakeArray( + std::vector{2, 3}, std::vector{1, 2, 3, 4, 5, 6})}; + array->GetDimension(0).SetLowerBound(0); + array->GetDimension(1).SetLowerBound(-1); + + // LBOUND(ARRAY, KIND=1) + auto int8Result{ + MakeArray(std::vector{array->rank()}, + std::vector(array->rank(), 0))}; + RTNAME(Lbound) + (int8Result->raw().base_addr, *array, /*KIND=*/1, __FILE__, __LINE__); + EXPECT_EQ(*int8Result->ZeroBasedIndexedElement(0), 0); + EXPECT_EQ(*int8Result->ZeroBasedIndexedElement(1), -1); + + // LBOUND(ARRAY, KIND=4) + auto int32Result{ + MakeArray(std::vector{array->rank()}, + std::vector(array->rank(), 0))}; + RTNAME(Lbound) + (int32Result->raw().base_addr, *array, /*KIND=*/4, __FILE__, __LINE__); + EXPECT_EQ(*int32Result->ZeroBasedIndexedElement(0), 0); + EXPECT_EQ(*int32Result->ZeroBasedIndexedElement(1), -1); + + // LBOUND(ARRAY, KIND=8) + auto int64Result{ + MakeArray(std::vector{array->rank()}, + std::vector(array->rank(), 0))}; + RTNAME(Lbound) + (int64Result->raw().base_addr, *array, /*KIND=*/8, __FILE__, __LINE__); + EXPECT_EQ(*int64Result->ZeroBasedIndexedElement(0), 0); + EXPECT_EQ(*int64Result->ZeroBasedIndexedElement(1), -1); +} + TEST(Inquiry, Ubound) { // ARRAY 1 3 5 // 2 4 6 From 23b8f59f0916982cb0674b6175b7c812b86ffb99 Mon Sep 17 00:00:00 2001 From: Dmitry Vasilyev Date: Mon, 10 Jun 2024 13:02:55 +0400 Subject: [PATCH 27/36] [lldb] Gracefully down TestCoroutineHandle test in case the 'coroutine' feature is missing (#94903) Do not let the compiler gets failed in case the target platform does not support the 'coroutine' C++ features. Just compile without it and let lldb know about missed/unsupported feature. --- .../generic/coroutine_handle/Makefile | 8 ++++++ .../generic/coroutine_handle/main.cpp | 27 ++++++++++++++++--- 2 files changed, 32 insertions(+), 3 deletions(-) diff --git a/lldb/test/API/functionalities/data-formatter/data-formatter-stl/generic/coroutine_handle/Makefile b/lldb/test/API/functionalities/data-formatter/data-formatter-stl/generic/coroutine_handle/Makefile index 6914024392cfd0c..3e5da0ecae6697e 100644 --- a/lldb/test/API/functionalities/data-formatter/data-formatter-stl/generic/coroutine_handle/Makefile +++ b/lldb/test/API/functionalities/data-formatter/data-formatter-stl/generic/coroutine_handle/Makefile @@ -1,4 +1,12 @@ CXX_SOURCES := main.cpp CFLAGS_EXTRAS := -std=c++20 +ifeq "1" "$(USE_LIBSTDCPP)" + CFLAGS_EXTRAS += -DUSE_LIBSTDCPP +endif + +ifeq "1" "$(USE_LIBCPP)" + CFLAGS_EXTRAS += -DUSE_LIBCPP +endif + include Makefile.rules diff --git a/lldb/test/API/functionalities/data-formatter/data-formatter-stl/generic/coroutine_handle/main.cpp b/lldb/test/API/functionalities/data-formatter/data-formatter-stl/generic/coroutine_handle/main.cpp index 8cb81c3bc9f4c4e..aaf76e0d8d99a15 100644 --- a/lldb/test/API/functionalities/data-formatter/data-formatter-stl/generic/coroutine_handle/main.cpp +++ b/lldb/test/API/functionalities/data-formatter/data-formatter-stl/generic/coroutine_handle/main.cpp @@ -1,13 +1,27 @@ +#if defined(USE_LIBSTDCPP) +#include +// glibc++ >= 11 and c++20 +#if defined(_GLIBCXX_RELEASE) && _GLIBCXX_RELEASE >= 11 #include +#define HAS_CPP_COROUTINES 1 +#endif +#endif + +// libc++ always has 'coroutine' feature. +#if defined(USE_LIBCPP) +#include +#define HAS_CPP_COROUTINES 1 +#endif bool is_implementation_supported() { -#ifdef _GLIBCXX_RELEASE - return _GLIBCXX_RELEASE >= 11; -#else +#ifdef HAS_CPP_COROUTINES return true; +#else + return false; #endif } +#ifdef HAS_CPP_COROUTINES // `int_generator` is a stripped down, minimal coroutine generator // type. struct int_generator { @@ -39,8 +53,11 @@ int_generator my_generator_func() { co_yield 42; } // a place to reliably set a breakpoint on. void empty_function_so_we_can_set_a_breakpoint() {} +#endif // HAS_CPP_COROUTINES + int main() { bool is_supported = is_implementation_supported(); +#ifdef HAS_CPP_COROUTINES int_generator gen = my_generator_func(); std::coroutine_handle<> type_erased_hdl = gen.hdl; std::coroutine_handle incorrectly_typed_hdl = @@ -48,4 +65,8 @@ int main() { gen.hdl.resume(); // Break at initial_suspend gen.hdl.resume(); // Break after co_yield empty_function_so_we_can_set_a_breakpoint(); // Break at final_suspend + return 0; +#else + return 0; // Break at initial_suspend +#endif // HAS_CPP_COROUTINES } From f97bcdbd95c075134b3d928299592c54c96ebaad Mon Sep 17 00:00:00 2001 From: Jay Foad Date: Mon, 10 Jun 2024 10:10:33 +0100 Subject: [PATCH 28/36] [KnownBits] Speed up ForeachKnownBits in unit test. NFC. (#94939) Use fast unsigned arithmetic before constructing an APInt. This gives me a ~2x speed up when running this in my Release+Asserts build: $ unittests/Support/SupportTests --gtest_filter=KnownBitsTest.*Exhaustive --- llvm/unittests/Support/KnownBitsTest.h | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/llvm/unittests/Support/KnownBitsTest.h b/llvm/unittests/Support/KnownBitsTest.h index 556da2b9ecdad0e..974051891d48e7f 100644 --- a/llvm/unittests/Support/KnownBitsTest.h +++ b/llvm/unittests/Support/KnownBitsTest.h @@ -34,13 +34,13 @@ template void ForeachKnownBits(unsigned Bits, FnTy Fn) { template void ForeachNumInKnownBits(const KnownBits &Known, FnTy Fn) { unsigned Bits = Known.getBitWidth(); - unsigned Max = 1 << Bits; + assert(Bits < 32); + unsigned Max = 1u << Bits; + unsigned Zero = Known.Zero.getZExtValue(); + unsigned One = Known.One.getZExtValue(); for (unsigned N = 0; N < Max; ++N) { - APInt Num(Bits, N); - if ((Num & Known.Zero) != 0 || (~Num & Known.One) != 0) - continue; - - Fn(Num); + if ((N & Zero) == 0 && (~N & One) == 0) + Fn(APInt(Bits, N)); } } From c9fd7b1a541af66d11a847564d31460a80c66d47 Mon Sep 17 00:00:00 2001 From: Jay Foad Date: Mon, 10 Jun 2024 10:11:57 +0100 Subject: [PATCH 29/36] [AMDGPU] Remove unused checks left over from X86 --- .../AMDGPU/fold-int-pow2-with-fmul-or-fdiv.ll | 1512 ----------------- 1 file changed, 1512 deletions(-) diff --git a/llvm/test/CodeGen/AMDGPU/fold-int-pow2-with-fmul-or-fdiv.ll b/llvm/test/CodeGen/AMDGPU/fold-int-pow2-with-fmul-or-fdiv.ll index 7c5f6d5e33efe73..301299daaa61f4b 100644 --- a/llvm/test/CodeGen/AMDGPU/fold-int-pow2-with-fmul-or-fdiv.ll +++ b/llvm/test/CodeGen/AMDGPU/fold-int-pow2-with-fmul-or-fdiv.ll @@ -372,52 +372,6 @@ define <8 x half> @fdiv_pow2_8xhalf(<8 x i16> %i) { } define double @fmul_pow_shl_cnt(i64 %cnt) nounwind { -; CHECK-SSE-LABEL: fmul_pow_shl_cnt: -; CHECK-SSE: # %bb.0: -; CHECK-SSE-NEXT: movq %rdi, %rcx -; CHECK-SSE-NEXT: movl $1, %eax -; CHECK-SSE-NEXT: # kill: def $cl killed $cl killed $rcx -; CHECK-SSE-NEXT: shlq %cl, %rax -; CHECK-SSE-NEXT: movq %rax, %xmm1 -; CHECK-SSE-NEXT: punpckldq {{.*#+}} xmm1 = xmm1[0],mem[0],xmm1[1],mem[1] -; CHECK-SSE-NEXT: subpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1 -; CHECK-SSE-NEXT: movapd %xmm1, %xmm0 -; CHECK-SSE-NEXT: unpckhpd {{.*#+}} xmm0 = xmm0[1],xmm1[1] -; CHECK-SSE-NEXT: addsd %xmm1, %xmm0 -; CHECK-SSE-NEXT: mulsd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0 -; CHECK-SSE-NEXT: retq -; -; CHECK-AVX2-LABEL: fmul_pow_shl_cnt: -; CHECK-AVX2: # %bb.0: -; CHECK-AVX2-NEXT: movq %rdi, %rcx -; CHECK-AVX2-NEXT: movl $1, %eax -; CHECK-AVX2-NEXT: # kill: def $cl killed $cl killed $rcx -; CHECK-AVX2-NEXT: shlq %cl, %rax -; CHECK-AVX2-NEXT: vmovq %rax, %xmm0 -; CHECK-AVX2-NEXT: vpunpckldq {{.*#+}} xmm0 = xmm0[0],mem[0],xmm0[1],mem[1] -; CHECK-AVX2-NEXT: vsubpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-AVX2-NEXT: vshufpd {{.*#+}} xmm1 = xmm0[1,0] -; CHECK-AVX2-NEXT: vaddsd %xmm0, %xmm1, %xmm0 -; CHECK-AVX2-NEXT: vmulsd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-AVX2-NEXT: retq -; -; CHECK-NO-FASTFMA-LABEL: fmul_pow_shl_cnt: -; CHECK-NO-FASTFMA: # %bb.0: -; CHECK-NO-FASTFMA-NEXT: movq %rdi, %rcx -; CHECK-NO-FASTFMA-NEXT: movl $1, %eax -; CHECK-NO-FASTFMA-NEXT: # kill: def $cl killed $cl killed $rcx -; CHECK-NO-FASTFMA-NEXT: shlq %cl, %rax -; CHECK-NO-FASTFMA-NEXT: vcvtusi2sd %rax, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vmulsd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: retq -; -; CHECK-FMA-LABEL: fmul_pow_shl_cnt: -; CHECK-FMA: # %bb.0: -; CHECK-FMA-NEXT: movl $1, %eax -; CHECK-FMA-NEXT: shlxq %rdi, %rax, %rax -; CHECK-FMA-NEXT: vcvtusi2sd %rax, %xmm0, %xmm0 -; CHECK-FMA-NEXT: vmulsd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-FMA-NEXT: retq ; VI-LABEL: fmul_pow_shl_cnt: ; VI: ; %bb.0: ; VI-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) @@ -462,52 +416,6 @@ define double @fmul_pow_shl_cnt(i64 %cnt) nounwind { } define double @fmul_pow_shl_cnt2(i64 %cnt) nounwind { -; CHECK-SSE-LABEL: fmul_pow_shl_cnt2: -; CHECK-SSE: # %bb.0: -; CHECK-SSE-NEXT: movq %rdi, %rcx -; CHECK-SSE-NEXT: movl $2, %eax -; CHECK-SSE-NEXT: # kill: def $cl killed $cl killed $rcx -; CHECK-SSE-NEXT: shlq %cl, %rax -; CHECK-SSE-NEXT: movq %rax, %xmm1 -; CHECK-SSE-NEXT: punpckldq {{.*#+}} xmm1 = xmm1[0],mem[0],xmm1[1],mem[1] -; CHECK-SSE-NEXT: subpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1 -; CHECK-SSE-NEXT: movapd %xmm1, %xmm0 -; CHECK-SSE-NEXT: unpckhpd {{.*#+}} xmm0 = xmm0[1],xmm1[1] -; CHECK-SSE-NEXT: addsd %xmm1, %xmm0 -; CHECK-SSE-NEXT: mulsd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0 -; CHECK-SSE-NEXT: retq -; -; CHECK-AVX2-LABEL: fmul_pow_shl_cnt2: -; CHECK-AVX2: # %bb.0: -; CHECK-AVX2-NEXT: movq %rdi, %rcx -; CHECK-AVX2-NEXT: movl $2, %eax -; CHECK-AVX2-NEXT: # kill: def $cl killed $cl killed $rcx -; CHECK-AVX2-NEXT: shlq %cl, %rax -; CHECK-AVX2-NEXT: vmovq %rax, %xmm0 -; CHECK-AVX2-NEXT: vpunpckldq {{.*#+}} xmm0 = xmm0[0],mem[0],xmm0[1],mem[1] -; CHECK-AVX2-NEXT: vsubpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-AVX2-NEXT: vshufpd {{.*#+}} xmm1 = xmm0[1,0] -; CHECK-AVX2-NEXT: vaddsd %xmm0, %xmm1, %xmm0 -; CHECK-AVX2-NEXT: vmulsd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-AVX2-NEXT: retq -; -; CHECK-NO-FASTFMA-LABEL: fmul_pow_shl_cnt2: -; CHECK-NO-FASTFMA: # %bb.0: -; CHECK-NO-FASTFMA-NEXT: movq %rdi, %rcx -; CHECK-NO-FASTFMA-NEXT: movl $2, %eax -; CHECK-NO-FASTFMA-NEXT: # kill: def $cl killed $cl killed $rcx -; CHECK-NO-FASTFMA-NEXT: shlq %cl, %rax -; CHECK-NO-FASTFMA-NEXT: vcvtusi2sd %rax, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vmulsd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: retq -; -; CHECK-FMA-LABEL: fmul_pow_shl_cnt2: -; CHECK-FMA: # %bb.0: -; CHECK-FMA-NEXT: movl $2, %eax -; CHECK-FMA-NEXT: shlxq %rdi, %rax, %rax -; CHECK-FMA-NEXT: vcvtusi2sd %rax, %xmm0, %xmm0 -; CHECK-FMA-NEXT: vmulsd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-FMA-NEXT: retq ; VI-LABEL: fmul_pow_shl_cnt2: ; VI: ; %bb.0: ; VI-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) @@ -552,51 +460,6 @@ define double @fmul_pow_shl_cnt2(i64 %cnt) nounwind { } define float @fmul_pow_select(i32 %cnt, i1 %c) nounwind { -; CHECK-SSE-LABEL: fmul_pow_select: -; CHECK-SSE: # %bb.0: -; CHECK-SSE-NEXT: movl %edi, %ecx -; CHECK-SSE-NEXT: andl $1, %esi -; CHECK-SSE-NEXT: movl $2, %eax -; CHECK-SSE-NEXT: subl %esi, %eax -; CHECK-SSE-NEXT: # kill: def $cl killed $cl killed $ecx -; CHECK-SSE-NEXT: shll %cl, %eax -; CHECK-SSE-NEXT: cvtsi2ss %rax, %xmm0 -; CHECK-SSE-NEXT: mulss {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0 -; CHECK-SSE-NEXT: retq -; -; CHECK-AVX2-LABEL: fmul_pow_select: -; CHECK-AVX2: # %bb.0: -; CHECK-AVX2-NEXT: movl %edi, %ecx -; CHECK-AVX2-NEXT: andl $1, %esi -; CHECK-AVX2-NEXT: movl $2, %eax -; CHECK-AVX2-NEXT: subl %esi, %eax -; CHECK-AVX2-NEXT: # kill: def $cl killed $cl killed $ecx -; CHECK-AVX2-NEXT: shll %cl, %eax -; CHECK-AVX2-NEXT: vcvtsi2ss %rax, %xmm0, %xmm0 -; CHECK-AVX2-NEXT: vmulss {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-AVX2-NEXT: retq -; -; CHECK-NO-FASTFMA-LABEL: fmul_pow_select: -; CHECK-NO-FASTFMA: # %bb.0: -; CHECK-NO-FASTFMA-NEXT: movl %edi, %ecx -; CHECK-NO-FASTFMA-NEXT: andl $1, %esi -; CHECK-NO-FASTFMA-NEXT: movl $2, %eax -; CHECK-NO-FASTFMA-NEXT: subl %esi, %eax -; CHECK-NO-FASTFMA-NEXT: # kill: def $cl killed $cl killed $ecx -; CHECK-NO-FASTFMA-NEXT: shll %cl, %eax -; CHECK-NO-FASTFMA-NEXT: vcvtusi2ss %eax, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vmulss {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: retq -; -; CHECK-FMA-LABEL: fmul_pow_select: -; CHECK-FMA: # %bb.0: -; CHECK-FMA-NEXT: andl $1, %esi -; CHECK-FMA-NEXT: movl $2, %eax -; CHECK-FMA-NEXT: subl %esi, %eax -; CHECK-FMA-NEXT: shlxl %edi, %eax, %eax -; CHECK-FMA-NEXT: vcvtusi2ss %eax, %xmm0, %xmm0 -; CHECK-FMA-NEXT: vmulss {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-FMA-NEXT: retq ; VI-LABEL: fmul_pow_select: ; VI: ; %bb.0: ; VI-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) @@ -640,55 +503,6 @@ define float @fmul_pow_select(i32 %cnt, i1 %c) nounwind { } define float @fmul_fly_pow_mul_min_pow2(i64 %cnt) nounwind { -; CHECK-SSE-LABEL: fmul_fly_pow_mul_min_pow2: -; CHECK-SSE: # %bb.0: -; CHECK-SSE-NEXT: movq %rdi, %rcx -; CHECK-SSE-NEXT: movl $8, %eax -; CHECK-SSE-NEXT: # kill: def $cl killed $cl killed $rcx -; CHECK-SSE-NEXT: shlq %cl, %rax -; CHECK-SSE-NEXT: cmpq $8192, %rax # imm = 0x2000 -; CHECK-SSE-NEXT: movl $8192, %ecx # imm = 0x2000 -; CHECK-SSE-NEXT: cmovbq %rax, %rcx -; CHECK-SSE-NEXT: cvtsi2ss %rcx, %xmm0 -; CHECK-SSE-NEXT: mulss {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0 -; CHECK-SSE-NEXT: retq -; -; CHECK-AVX2-LABEL: fmul_fly_pow_mul_min_pow2: -; CHECK-AVX2: # %bb.0: -; CHECK-AVX2-NEXT: movq %rdi, %rcx -; CHECK-AVX2-NEXT: movl $8, %eax -; CHECK-AVX2-NEXT: # kill: def $cl killed $cl killed $rcx -; CHECK-AVX2-NEXT: shlq %cl, %rax -; CHECK-AVX2-NEXT: cmpq $8192, %rax # imm = 0x2000 -; CHECK-AVX2-NEXT: movl $8192, %ecx # imm = 0x2000 -; CHECK-AVX2-NEXT: cmovbq %rax, %rcx -; CHECK-AVX2-NEXT: vcvtsi2ss %rcx, %xmm0, %xmm0 -; CHECK-AVX2-NEXT: vmulss {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-AVX2-NEXT: retq -; -; CHECK-NO-FASTFMA-LABEL: fmul_fly_pow_mul_min_pow2: -; CHECK-NO-FASTFMA: # %bb.0: -; CHECK-NO-FASTFMA-NEXT: movq %rdi, %rcx -; CHECK-NO-FASTFMA-NEXT: movl $8, %eax -; CHECK-NO-FASTFMA-NEXT: # kill: def $cl killed $cl killed $rcx -; CHECK-NO-FASTFMA-NEXT: shlq %cl, %rax -; CHECK-NO-FASTFMA-NEXT: cmpq $8192, %rax # imm = 0x2000 -; CHECK-NO-FASTFMA-NEXT: movl $8192, %ecx # imm = 0x2000 -; CHECK-NO-FASTFMA-NEXT: cmovbq %rax, %rcx -; CHECK-NO-FASTFMA-NEXT: vcvtsi2ss %rcx, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vmulss {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: retq -; -; CHECK-FMA-LABEL: fmul_fly_pow_mul_min_pow2: -; CHECK-FMA: # %bb.0: -; CHECK-FMA-NEXT: movl $8, %eax -; CHECK-FMA-NEXT: shlxq %rdi, %rax, %rax -; CHECK-FMA-NEXT: cmpq $8192, %rax # imm = 0x2000 -; CHECK-FMA-NEXT: movl $8192, %ecx # imm = 0x2000 -; CHECK-FMA-NEXT: cmovbq %rax, %rcx -; CHECK-FMA-NEXT: vcvtsi2ss %rcx, %xmm0, %xmm0 -; CHECK-FMA-NEXT: vmulss {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-FMA-NEXT: retq ; VI-LABEL: fmul_fly_pow_mul_min_pow2: ; VI: ; %bb.0: ; VI-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) @@ -757,63 +571,6 @@ define float @fmul_fly_pow_mul_min_pow2(i64 %cnt) nounwind { } define double @fmul_pow_mul_max_pow2(i16 %cnt) nounwind { -; CHECK-SSE-LABEL: fmul_pow_mul_max_pow2: -; CHECK-SSE: # %bb.0: -; CHECK-SSE-NEXT: movl %edi, %ecx -; CHECK-SSE-NEXT: movl $2, %eax -; CHECK-SSE-NEXT: shll %cl, %eax -; CHECK-SSE-NEXT: movl $1, %edx -; CHECK-SSE-NEXT: # kill: def $cl killed $cl killed $ecx -; CHECK-SSE-NEXT: shll %cl, %edx -; CHECK-SSE-NEXT: cmpw %ax, %dx -; CHECK-SSE-NEXT: cmovbel %eax, %edx -; CHECK-SSE-NEXT: movzwl %dx, %eax -; CHECK-SSE-NEXT: cvtsi2sd %eax, %xmm0 -; CHECK-SSE-NEXT: mulsd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0 -; CHECK-SSE-NEXT: retq -; -; CHECK-AVX2-LABEL: fmul_pow_mul_max_pow2: -; CHECK-AVX2: # %bb.0: -; CHECK-AVX2-NEXT: movl %edi, %ecx -; CHECK-AVX2-NEXT: movl $2, %eax -; CHECK-AVX2-NEXT: shll %cl, %eax -; CHECK-AVX2-NEXT: movl $1, %edx -; CHECK-AVX2-NEXT: # kill: def $cl killed $cl killed $ecx -; CHECK-AVX2-NEXT: shll %cl, %edx -; CHECK-AVX2-NEXT: cmpw %ax, %dx -; CHECK-AVX2-NEXT: cmovbel %eax, %edx -; CHECK-AVX2-NEXT: movzwl %dx, %eax -; CHECK-AVX2-NEXT: vcvtsi2sd %eax, %xmm0, %xmm0 -; CHECK-AVX2-NEXT: vmulsd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-AVX2-NEXT: retq -; -; CHECK-NO-FASTFMA-LABEL: fmul_pow_mul_max_pow2: -; CHECK-NO-FASTFMA: # %bb.0: -; CHECK-NO-FASTFMA-NEXT: movl %edi, %ecx -; CHECK-NO-FASTFMA-NEXT: movl $2, %eax -; CHECK-NO-FASTFMA-NEXT: shll %cl, %eax -; CHECK-NO-FASTFMA-NEXT: movl $1, %edx -; CHECK-NO-FASTFMA-NEXT: # kill: def $cl killed $cl killed $ecx -; CHECK-NO-FASTFMA-NEXT: shll %cl, %edx -; CHECK-NO-FASTFMA-NEXT: cmpw %ax, %dx -; CHECK-NO-FASTFMA-NEXT: cmovbel %eax, %edx -; CHECK-NO-FASTFMA-NEXT: movzwl %dx, %eax -; CHECK-NO-FASTFMA-NEXT: vcvtsi2sd %eax, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vmulsd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: retq -; -; CHECK-FMA-LABEL: fmul_pow_mul_max_pow2: -; CHECK-FMA: # %bb.0: -; CHECK-FMA-NEXT: movl $2, %eax -; CHECK-FMA-NEXT: shlxl %edi, %eax, %eax -; CHECK-FMA-NEXT: movl $1, %ecx -; CHECK-FMA-NEXT: shlxl %edi, %ecx, %ecx -; CHECK-FMA-NEXT: cmpw %ax, %cx -; CHECK-FMA-NEXT: cmoval %ecx, %eax -; CHECK-FMA-NEXT: movzwl %ax, %eax -; CHECK-FMA-NEXT: vcvtsi2sd %eax, %xmm0, %xmm0 -; CHECK-FMA-NEXT: vmulsd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-FMA-NEXT: retq ; VI-LABEL: fmul_pow_mul_max_pow2: ; VI: ; %bb.0: ; VI-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) @@ -852,48 +609,6 @@ define double @fmul_pow_mul_max_pow2(i16 %cnt) nounwind { } define double @fmul_pow_shl_cnt_fail_maybe_non_pow2(i64 %v, i64 %cnt) nounwind { -; CHECK-SSE-LABEL: fmul_pow_shl_cnt_fail_maybe_non_pow2: -; CHECK-SSE: # %bb.0: -; CHECK-SSE-NEXT: movq %rsi, %rcx -; CHECK-SSE-NEXT: # kill: def $cl killed $cl killed $rcx -; CHECK-SSE-NEXT: shlq %cl, %rdi -; CHECK-SSE-NEXT: movq %rdi, %xmm1 -; CHECK-SSE-NEXT: punpckldq {{.*#+}} xmm1 = xmm1[0],mem[0],xmm1[1],mem[1] -; CHECK-SSE-NEXT: subpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1 -; CHECK-SSE-NEXT: movapd %xmm1, %xmm0 -; CHECK-SSE-NEXT: unpckhpd {{.*#+}} xmm0 = xmm0[1],xmm1[1] -; CHECK-SSE-NEXT: addsd %xmm1, %xmm0 -; CHECK-SSE-NEXT: mulsd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0 -; CHECK-SSE-NEXT: retq -; -; CHECK-AVX2-LABEL: fmul_pow_shl_cnt_fail_maybe_non_pow2: -; CHECK-AVX2: # %bb.0: -; CHECK-AVX2-NEXT: movq %rsi, %rcx -; CHECK-AVX2-NEXT: # kill: def $cl killed $cl killed $rcx -; CHECK-AVX2-NEXT: shlq %cl, %rdi -; CHECK-AVX2-NEXT: vmovq %rdi, %xmm0 -; CHECK-AVX2-NEXT: vpunpckldq {{.*#+}} xmm0 = xmm0[0],mem[0],xmm0[1],mem[1] -; CHECK-AVX2-NEXT: vsubpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-AVX2-NEXT: vshufpd {{.*#+}} xmm1 = xmm0[1,0] -; CHECK-AVX2-NEXT: vaddsd %xmm0, %xmm1, %xmm0 -; CHECK-AVX2-NEXT: vmulsd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-AVX2-NEXT: retq -; -; CHECK-NO-FASTFMA-LABEL: fmul_pow_shl_cnt_fail_maybe_non_pow2: -; CHECK-NO-FASTFMA: # %bb.0: -; CHECK-NO-FASTFMA-NEXT: movq %rsi, %rcx -; CHECK-NO-FASTFMA-NEXT: # kill: def $cl killed $cl killed $rcx -; CHECK-NO-FASTFMA-NEXT: shlq %cl, %rdi -; CHECK-NO-FASTFMA-NEXT: vcvtusi2sd %rdi, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vmulsd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: retq -; -; CHECK-FMA-LABEL: fmul_pow_shl_cnt_fail_maybe_non_pow2: -; CHECK-FMA: # %bb.0: -; CHECK-FMA-NEXT: shlxq %rsi, %rdi, %rax -; CHECK-FMA-NEXT: vcvtusi2sd %rax, %xmm0, %xmm0 -; CHECK-FMA-NEXT: vmulsd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-FMA-NEXT: retq ; VI-LABEL: fmul_pow_shl_cnt_fail_maybe_non_pow2: ; VI: ; %bb.0: ; VI-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) @@ -938,90 +653,6 @@ define double @fmul_pow_shl_cnt_fail_maybe_non_pow2(i64 %v, i64 %cnt) nounwind { } define <2 x float> @fmul_pow_shl_cnt_vec_fail_expensive_cast(<2 x i64> %cnt) nounwind { -; CHECK-SSE-LABEL: fmul_pow_shl_cnt_vec_fail_expensive_cast: -; CHECK-SSE: # %bb.0: -; CHECK-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm0[2,3,2,3] -; CHECK-SSE-NEXT: movdqa {{.*#+}} xmm3 = [2,2] -; CHECK-SSE-NEXT: movdqa %xmm3, %xmm1 -; CHECK-SSE-NEXT: psllq %xmm2, %xmm1 -; CHECK-SSE-NEXT: psllq %xmm0, %xmm3 -; CHECK-SSE-NEXT: movq %xmm3, %rax -; CHECK-SSE-NEXT: testq %rax, %rax -; CHECK-SSE-NEXT: js .LBB6_1 -; CHECK-SSE-NEXT: # %bb.2: -; CHECK-SSE-NEXT: xorps %xmm0, %xmm0 -; CHECK-SSE-NEXT: cvtsi2ss %rax, %xmm0 -; CHECK-SSE-NEXT: jmp .LBB6_3 -; CHECK-SSE-NEXT: .LBB6_1: -; CHECK-SSE-NEXT: movq %rax, %rcx -; CHECK-SSE-NEXT: shrq %rcx -; CHECK-SSE-NEXT: andl $1, %eax -; CHECK-SSE-NEXT: orq %rcx, %rax -; CHECK-SSE-NEXT: xorps %xmm0, %xmm0 -; CHECK-SSE-NEXT: cvtsi2ss %rax, %xmm0 -; CHECK-SSE-NEXT: addss %xmm0, %xmm0 -; CHECK-SSE-NEXT: .LBB6_3: -; CHECK-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[2,3,2,3] -; CHECK-SSE-NEXT: movq %xmm1, %rax -; CHECK-SSE-NEXT: testq %rax, %rax -; CHECK-SSE-NEXT: js .LBB6_4 -; CHECK-SSE-NEXT: # %bb.5: -; CHECK-SSE-NEXT: xorps %xmm1, %xmm1 -; CHECK-SSE-NEXT: cvtsi2ss %rax, %xmm1 -; CHECK-SSE-NEXT: jmp .LBB6_6 -; CHECK-SSE-NEXT: .LBB6_4: -; CHECK-SSE-NEXT: movq %rax, %rcx -; CHECK-SSE-NEXT: shrq %rcx -; CHECK-SSE-NEXT: andl $1, %eax -; CHECK-SSE-NEXT: orq %rcx, %rax -; CHECK-SSE-NEXT: xorps %xmm1, %xmm1 -; CHECK-SSE-NEXT: cvtsi2ss %rax, %xmm1 -; CHECK-SSE-NEXT: addss %xmm1, %xmm1 -; CHECK-SSE-NEXT: .LBB6_6: -; CHECK-SSE-NEXT: unpcklps {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1] -; CHECK-SSE-NEXT: mulps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0 -; CHECK-SSE-NEXT: retq -; -; CHECK-AVX2-LABEL: fmul_pow_shl_cnt_vec_fail_expensive_cast: -; CHECK-AVX2: # %bb.0: -; CHECK-AVX2-NEXT: vpbroadcastq {{.*#+}} xmm1 = [2,2] -; CHECK-AVX2-NEXT: vpsllvq %xmm0, %xmm1, %xmm0 -; CHECK-AVX2-NEXT: vpsrlq $1, %xmm0, %xmm1 -; CHECK-AVX2-NEXT: vblendvpd %xmm0, %xmm1, %xmm0, %xmm1 -; CHECK-AVX2-NEXT: vpextrq $1, %xmm1, %rax -; CHECK-AVX2-NEXT: vcvtsi2ss %rax, %xmm2, %xmm2 -; CHECK-AVX2-NEXT: vmovq %xmm1, %rax -; CHECK-AVX2-NEXT: vcvtsi2ss %rax, %xmm3, %xmm1 -; CHECK-AVX2-NEXT: vinsertps {{.*#+}} xmm1 = xmm1[0],xmm2[0],zero,zero -; CHECK-AVX2-NEXT: vaddps %xmm1, %xmm1, %xmm2 -; CHECK-AVX2-NEXT: vpxor %xmm3, %xmm3, %xmm3 -; CHECK-AVX2-NEXT: vpcmpgtq %xmm0, %xmm3, %xmm0 -; CHECK-AVX2-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[1,3,2,3] -; CHECK-AVX2-NEXT: vblendvps %xmm0, %xmm2, %xmm1, %xmm0 -; CHECK-AVX2-NEXT: vbroadcastss {{.*#+}} xmm1 = [1.5E+1,1.5E+1,1.5E+1,1.5E+1] -; CHECK-AVX2-NEXT: vmulps %xmm1, %xmm0, %xmm0 -; CHECK-AVX2-NEXT: retq -; -; CHECK-NO-FASTFMA-LABEL: fmul_pow_shl_cnt_vec_fail_expensive_cast: -; CHECK-NO-FASTFMA: # %bb.0: -; CHECK-NO-FASTFMA-NEXT: vpbroadcastq {{.*#+}} xmm1 = [2,2] -; CHECK-NO-FASTFMA-NEXT: vpsllvq %xmm0, %xmm1, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vpextrq $1, %xmm0, %rax -; CHECK-NO-FASTFMA-NEXT: vcvtusi2ss %rax, %xmm2, %xmm1 -; CHECK-NO-FASTFMA-NEXT: vmovq %xmm0, %rax -; CHECK-NO-FASTFMA-NEXT: vcvtusi2ss %rax, %xmm2, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vinsertps {{.*#+}} xmm0 = xmm0[0],xmm1[0],zero,zero -; CHECK-NO-FASTFMA-NEXT: vbroadcastss {{.*#+}} xmm1 = [1.5E+1,1.5E+1,1.5E+1,1.5E+1] -; CHECK-NO-FASTFMA-NEXT: vmulps %xmm1, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: retq -; -; CHECK-FMA-LABEL: fmul_pow_shl_cnt_vec_fail_expensive_cast: -; CHECK-FMA: # %bb.0: -; CHECK-FMA-NEXT: vpbroadcastq {{.*#+}} xmm1 = [2,2] -; CHECK-FMA-NEXT: vpsllvq %xmm0, %xmm1, %xmm0 -; CHECK-FMA-NEXT: vcvtuqq2ps %xmm0, %xmm0 -; CHECK-FMA-NEXT: vmulps {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %xmm0, %xmm0 -; CHECK-FMA-NEXT: retq ; VI-LABEL: fmul_pow_shl_cnt_vec_fail_expensive_cast: ; VI: ; %bb.0: ; VI-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) @@ -1110,60 +741,6 @@ define <2 x float> @fmul_pow_shl_cnt_vec_fail_expensive_cast(<2 x i64> %cnt) nou } define <2 x double> @fmul_pow_shl_cnt_vec(<2 x i64> %cnt) nounwind { -; CHECK-SSE-LABEL: fmul_pow_shl_cnt_vec: -; CHECK-SSE: # %bb.0: -; CHECK-SSE-NEXT: movdqa {{.*#+}} xmm1 = [2,2] -; CHECK-SSE-NEXT: movdqa %xmm1, %xmm2 -; CHECK-SSE-NEXT: psllq %xmm0, %xmm2 -; CHECK-SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,3,2,3] -; CHECK-SSE-NEXT: psllq %xmm0, %xmm1 -; CHECK-SSE-NEXT: movsd {{.*#+}} xmm1 = xmm2[0],xmm1[1] -; CHECK-SSE-NEXT: movapd {{.*#+}} xmm0 = [4294967295,4294967295] -; CHECK-SSE-NEXT: andpd %xmm1, %xmm0 -; CHECK-SSE-NEXT: orpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0 -; CHECK-SSE-NEXT: psrlq $32, %xmm1 -; CHECK-SSE-NEXT: por {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1 -; CHECK-SSE-NEXT: subpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1 -; CHECK-SSE-NEXT: addpd %xmm0, %xmm1 -; CHECK-SSE-NEXT: mulpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1 -; CHECK-SSE-NEXT: movapd %xmm1, %xmm0 -; CHECK-SSE-NEXT: retq -; -; CHECK-AVX2-LABEL: fmul_pow_shl_cnt_vec: -; CHECK-AVX2: # %bb.0: -; CHECK-AVX2-NEXT: vpbroadcastq {{.*#+}} xmm1 = [2,2] -; CHECK-AVX2-NEXT: vpsllvq %xmm0, %xmm1, %xmm0 -; CHECK-AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1 -; CHECK-AVX2-NEXT: vpblendd {{.*#+}} xmm1 = xmm0[0],xmm1[1],xmm0[2],xmm1[3] -; CHECK-AVX2-NEXT: vpor {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1, %xmm1 -; CHECK-AVX2-NEXT: vpsrlq $32, %xmm0, %xmm0 -; CHECK-AVX2-NEXT: vpor {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-AVX2-NEXT: vsubpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-AVX2-NEXT: vaddpd %xmm0, %xmm1, %xmm0 -; CHECK-AVX2-NEXT: vmulpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-AVX2-NEXT: retq -; -; CHECK-NO-FASTFMA-LABEL: fmul_pow_shl_cnt_vec: -; CHECK-NO-FASTFMA: # %bb.0: -; CHECK-NO-FASTFMA-NEXT: vpbroadcastq {{.*#+}} xmm1 = [2,2] -; CHECK-NO-FASTFMA-NEXT: vpsllvq %xmm0, %xmm1, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vpxor %xmm1, %xmm1, %xmm1 -; CHECK-NO-FASTFMA-NEXT: vpblendd {{.*#+}} xmm1 = xmm0[0],xmm1[1],xmm0[2],xmm1[3] -; CHECK-NO-FASTFMA-NEXT: vpor {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1, %xmm1 -; CHECK-NO-FASTFMA-NEXT: vpsrlq $32, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vpor {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vsubpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vaddpd %xmm0, %xmm1, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vmulpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: retq -; -; CHECK-FMA-LABEL: fmul_pow_shl_cnt_vec: -; CHECK-FMA: # %bb.0: -; CHECK-FMA-NEXT: vpbroadcastq {{.*#+}} xmm1 = [2,2] -; CHECK-FMA-NEXT: vpsllvq %xmm0, %xmm1, %xmm0 -; CHECK-FMA-NEXT: vcvtuqq2pd %xmm0, %xmm0 -; CHECK-FMA-NEXT: vmulpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to2}, %xmm0, %xmm0 -; CHECK-FMA-NEXT: retq ; VI-LABEL: fmul_pow_shl_cnt_vec: ; VI: ; %bb.0: ; VI-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) @@ -1228,64 +805,6 @@ define <2 x double> @fmul_pow_shl_cnt_vec(<2 x i64> %cnt) nounwind { } define <4 x float> @fmul_pow_shl_cnt_vec_preserve_fma(<4 x i32> %cnt, <4 x float> %add) nounwind { -; CHECK-SSE-LABEL: fmul_pow_shl_cnt_vec_preserve_fma: -; CHECK-SSE: # %bb.0: -; CHECK-SSE-NEXT: pslld $23, %xmm0 -; CHECK-SSE-NEXT: paddd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0 -; CHECK-SSE-NEXT: cvttps2dq %xmm0, %xmm0 -; CHECK-SSE-NEXT: movdqa {{.*#+}} xmm2 = [2,2,2,2] -; CHECK-SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,1,3,3] -; CHECK-SSE-NEXT: pmuludq %xmm2, %xmm0 -; CHECK-SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3] -; CHECK-SSE-NEXT: pmuludq %xmm2, %xmm3 -; CHECK-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm3[0,2,2,3] -; CHECK-SSE-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1] -; CHECK-SSE-NEXT: movdqa {{.*#+}} xmm2 = [65535,65535,65535,65535] -; CHECK-SSE-NEXT: pand %xmm0, %xmm2 -; CHECK-SSE-NEXT: por {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm2 -; CHECK-SSE-NEXT: psrld $16, %xmm0 -; CHECK-SSE-NEXT: por {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0 -; CHECK-SSE-NEXT: subps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0 -; CHECK-SSE-NEXT: addps %xmm2, %xmm0 -; CHECK-SSE-NEXT: mulps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0 -; CHECK-SSE-NEXT: addps %xmm1, %xmm0 -; CHECK-SSE-NEXT: retq -; -; CHECK-AVX2-LABEL: fmul_pow_shl_cnt_vec_preserve_fma: -; CHECK-AVX2: # %bb.0: -; CHECK-AVX2-NEXT: vpbroadcastd {{.*#+}} xmm2 = [2,2,2,2] -; CHECK-AVX2-NEXT: vpsllvd %xmm0, %xmm2, %xmm0 -; CHECK-AVX2-NEXT: vpbroadcastd {{.*#+}} xmm2 = [1258291200,1258291200,1258291200,1258291200] -; CHECK-AVX2-NEXT: vpblendw {{.*#+}} xmm2 = xmm0[0],xmm2[1],xmm0[2],xmm2[3],xmm0[4],xmm2[5],xmm0[6],xmm2[7] -; CHECK-AVX2-NEXT: vpsrld $16, %xmm0, %xmm0 -; CHECK-AVX2-NEXT: vpbroadcastd {{.*#+}} xmm3 = [1392508928,1392508928,1392508928,1392508928] -; CHECK-AVX2-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0],xmm3[1],xmm0[2],xmm3[3],xmm0[4],xmm3[5],xmm0[6],xmm3[7] -; CHECK-AVX2-NEXT: vbroadcastss {{.*#+}} xmm3 = [5.49764202E+11,5.49764202E+11,5.49764202E+11,5.49764202E+11] -; CHECK-AVX2-NEXT: vsubps %xmm3, %xmm0, %xmm0 -; CHECK-AVX2-NEXT: vaddps %xmm0, %xmm2, %xmm0 -; CHECK-AVX2-NEXT: vbroadcastss {{.*#+}} xmm2 = [5.0E+0,5.0E+0,5.0E+0,5.0E+0] -; CHECK-AVX2-NEXT: vmulps %xmm2, %xmm0, %xmm0 -; CHECK-AVX2-NEXT: vaddps %xmm1, %xmm0, %xmm0 -; CHECK-AVX2-NEXT: retq -; -; CHECK-NO-FASTFMA-LABEL: fmul_pow_shl_cnt_vec_preserve_fma: -; CHECK-NO-FASTFMA: # %bb.0: -; CHECK-NO-FASTFMA-NEXT: vpbroadcastd {{.*#+}} xmm2 = [2,2,2,2] -; CHECK-NO-FASTFMA-NEXT: vpsllvd %xmm0, %xmm2, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vcvtudq2ps %zmm0, %zmm0 -; CHECK-NO-FASTFMA-NEXT: vbroadcastss {{.*#+}} xmm2 = [5.0E+0,5.0E+0,5.0E+0,5.0E+0] -; CHECK-NO-FASTFMA-NEXT: vmulps %xmm2, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vaddps %xmm1, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vzeroupper -; CHECK-NO-FASTFMA-NEXT: retq -; -; CHECK-FMA-LABEL: fmul_pow_shl_cnt_vec_preserve_fma: -; CHECK-FMA: # %bb.0: -; CHECK-FMA-NEXT: vpbroadcastd {{.*#+}} xmm2 = [2,2,2,2] -; CHECK-FMA-NEXT: vpsllvd %xmm0, %xmm2, %xmm0 -; CHECK-FMA-NEXT: vcvtudq2ps %xmm0, %xmm0 -; CHECK-FMA-NEXT: vfmadd132ps {{.*#+}} xmm0 = (xmm0 * mem) + xmm1 -; CHECK-FMA-NEXT: retq ; VI-LABEL: fmul_pow_shl_cnt_vec_preserve_fma: ; VI: ; %bb.0: ; VI-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) @@ -1356,60 +875,6 @@ define <4 x float> @fmul_pow_shl_cnt_vec_preserve_fma(<4 x i32> %cnt, <4 x float } define <2 x double> @fmul_pow_shl_cnt_vec_non_splat_todo(<2 x i64> %cnt) nounwind { -; CHECK-SSE-LABEL: fmul_pow_shl_cnt_vec_non_splat_todo: -; CHECK-SSE: # %bb.0: -; CHECK-SSE-NEXT: movdqa {{.*#+}} xmm1 = [2,2] -; CHECK-SSE-NEXT: movdqa %xmm1, %xmm2 -; CHECK-SSE-NEXT: psllq %xmm0, %xmm2 -; CHECK-SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,3,2,3] -; CHECK-SSE-NEXT: psllq %xmm0, %xmm1 -; CHECK-SSE-NEXT: movsd {{.*#+}} xmm1 = xmm2[0],xmm1[1] -; CHECK-SSE-NEXT: movapd {{.*#+}} xmm0 = [4294967295,4294967295] -; CHECK-SSE-NEXT: andpd %xmm1, %xmm0 -; CHECK-SSE-NEXT: orpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0 -; CHECK-SSE-NEXT: psrlq $32, %xmm1 -; CHECK-SSE-NEXT: por {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1 -; CHECK-SSE-NEXT: subpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1 -; CHECK-SSE-NEXT: addpd %xmm0, %xmm1 -; CHECK-SSE-NEXT: mulpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1 -; CHECK-SSE-NEXT: movapd %xmm1, %xmm0 -; CHECK-SSE-NEXT: retq -; -; CHECK-AVX2-LABEL: fmul_pow_shl_cnt_vec_non_splat_todo: -; CHECK-AVX2: # %bb.0: -; CHECK-AVX2-NEXT: vpbroadcastq {{.*#+}} xmm1 = [2,2] -; CHECK-AVX2-NEXT: vpsllvq %xmm0, %xmm1, %xmm0 -; CHECK-AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1 -; CHECK-AVX2-NEXT: vpblendd {{.*#+}} xmm1 = xmm0[0],xmm1[1],xmm0[2],xmm1[3] -; CHECK-AVX2-NEXT: vpor {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1, %xmm1 -; CHECK-AVX2-NEXT: vpsrlq $32, %xmm0, %xmm0 -; CHECK-AVX2-NEXT: vpor {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-AVX2-NEXT: vsubpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-AVX2-NEXT: vaddpd %xmm0, %xmm1, %xmm0 -; CHECK-AVX2-NEXT: vmulpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-AVX2-NEXT: retq -; -; CHECK-NO-FASTFMA-LABEL: fmul_pow_shl_cnt_vec_non_splat_todo: -; CHECK-NO-FASTFMA: # %bb.0: -; CHECK-NO-FASTFMA-NEXT: vpbroadcastq {{.*#+}} xmm1 = [2,2] -; CHECK-NO-FASTFMA-NEXT: vpsllvq %xmm0, %xmm1, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vpxor %xmm1, %xmm1, %xmm1 -; CHECK-NO-FASTFMA-NEXT: vpblendd {{.*#+}} xmm1 = xmm0[0],xmm1[1],xmm0[2],xmm1[3] -; CHECK-NO-FASTFMA-NEXT: vpor {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1, %xmm1 -; CHECK-NO-FASTFMA-NEXT: vpsrlq $32, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vpor {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vsubpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vaddpd %xmm0, %xmm1, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vmulpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: retq -; -; CHECK-FMA-LABEL: fmul_pow_shl_cnt_vec_non_splat_todo: -; CHECK-FMA: # %bb.0: -; CHECK-FMA-NEXT: vpbroadcastq {{.*#+}} xmm1 = [2,2] -; CHECK-FMA-NEXT: vpsllvq %xmm0, %xmm1, %xmm0 -; CHECK-FMA-NEXT: vcvtuqq2pd %xmm0, %xmm0 -; CHECK-FMA-NEXT: vmulpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-FMA-NEXT: retq ; VI-LABEL: fmul_pow_shl_cnt_vec_non_splat_todo: ; VI: ; %bb.0: ; VI-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) @@ -1476,60 +941,6 @@ define <2 x double> @fmul_pow_shl_cnt_vec_non_splat_todo(<2 x i64> %cnt) nounwin } define <2 x double> @fmul_pow_shl_cnt_vec_non_splat2_todo(<2 x i64> %cnt) nounwind { -; CHECK-SSE-LABEL: fmul_pow_shl_cnt_vec_non_splat2_todo: -; CHECK-SSE: # %bb.0: -; CHECK-SSE-NEXT: movdqa {{.*#+}} xmm1 = [2,1] -; CHECK-SSE-NEXT: movdqa %xmm1, %xmm2 -; CHECK-SSE-NEXT: psllq %xmm0, %xmm2 -; CHECK-SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,3,2,3] -; CHECK-SSE-NEXT: psllq %xmm0, %xmm1 -; CHECK-SSE-NEXT: movsd {{.*#+}} xmm1 = xmm2[0],xmm1[1] -; CHECK-SSE-NEXT: movapd {{.*#+}} xmm0 = [4294967295,4294967295] -; CHECK-SSE-NEXT: andpd %xmm1, %xmm0 -; CHECK-SSE-NEXT: orpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0 -; CHECK-SSE-NEXT: psrlq $32, %xmm1 -; CHECK-SSE-NEXT: por {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1 -; CHECK-SSE-NEXT: subpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1 -; CHECK-SSE-NEXT: addpd %xmm0, %xmm1 -; CHECK-SSE-NEXT: mulpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1 -; CHECK-SSE-NEXT: movapd %xmm1, %xmm0 -; CHECK-SSE-NEXT: retq -; -; CHECK-AVX2-LABEL: fmul_pow_shl_cnt_vec_non_splat2_todo: -; CHECK-AVX2: # %bb.0: -; CHECK-AVX2-NEXT: vmovdqa {{.*#+}} xmm1 = [2,1] -; CHECK-AVX2-NEXT: vpsllvq %xmm0, %xmm1, %xmm0 -; CHECK-AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1 -; CHECK-AVX2-NEXT: vpblendd {{.*#+}} xmm1 = xmm0[0],xmm1[1],xmm0[2],xmm1[3] -; CHECK-AVX2-NEXT: vpor {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1, %xmm1 -; CHECK-AVX2-NEXT: vpsrlq $32, %xmm0, %xmm0 -; CHECK-AVX2-NEXT: vpor {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-AVX2-NEXT: vsubpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-AVX2-NEXT: vaddpd %xmm0, %xmm1, %xmm0 -; CHECK-AVX2-NEXT: vmulpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-AVX2-NEXT: retq -; -; CHECK-NO-FASTFMA-LABEL: fmul_pow_shl_cnt_vec_non_splat2_todo: -; CHECK-NO-FASTFMA: # %bb.0: -; CHECK-NO-FASTFMA-NEXT: vmovdqa {{.*#+}} xmm1 = [2,1] -; CHECK-NO-FASTFMA-NEXT: vpsllvq %xmm0, %xmm1, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vpxor %xmm1, %xmm1, %xmm1 -; CHECK-NO-FASTFMA-NEXT: vpblendd {{.*#+}} xmm1 = xmm0[0],xmm1[1],xmm0[2],xmm1[3] -; CHECK-NO-FASTFMA-NEXT: vpor {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1, %xmm1 -; CHECK-NO-FASTFMA-NEXT: vpsrlq $32, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vpor {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vsubpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vaddpd %xmm0, %xmm1, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vmulpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: retq -; -; CHECK-FMA-LABEL: fmul_pow_shl_cnt_vec_non_splat2_todo: -; CHECK-FMA: # %bb.0: -; CHECK-FMA-NEXT: vmovdqa {{.*#+}} xmm1 = [2,1] -; CHECK-FMA-NEXT: vpsllvq %xmm0, %xmm1, %xmm0 -; CHECK-FMA-NEXT: vcvtuqq2pd %xmm0, %xmm0 -; CHECK-FMA-NEXT: vmulpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to2}, %xmm0, %xmm0 -; CHECK-FMA-NEXT: retq ; VI-LABEL: fmul_pow_shl_cnt_vec_non_splat2_todo: ; VI: ; %bb.0: ; VI-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) @@ -1594,158 +1005,6 @@ define <2 x double> @fmul_pow_shl_cnt_vec_non_splat2_todo(<2 x i64> %cnt) nounwi } define <2 x half> @fmul_pow_shl_cnt_vec_fail_to_large(<2 x i16> %cnt) nounwind { -; CHECK-SSE-LABEL: fmul_pow_shl_cnt_vec_fail_to_large: -; CHECK-SSE: # %bb.0: -; CHECK-SSE-NEXT: subq $40, %rsp -; CHECK-SSE-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3] -; CHECK-SSE-NEXT: pslld $23, %xmm0 -; CHECK-SSE-NEXT: paddd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0 -; CHECK-SSE-NEXT: cvttps2dq %xmm0, %xmm0 -; CHECK-SSE-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7] -; CHECK-SSE-NEXT: pmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0 -; CHECK-SSE-NEXT: movdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill -; CHECK-SSE-NEXT: pextrw $1, %xmm0, %eax -; CHECK-SSE-NEXT: xorps %xmm0, %xmm0 -; CHECK-SSE-NEXT: cvtsi2ss %eax, %xmm0 -; CHECK-SSE-NEXT: callq __truncsfhf2@PLT -; CHECK-SSE-NEXT: movss %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 4-byte Spill -; CHECK-SSE-NEXT: movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload -; CHECK-SSE-NEXT: pextrw $0, %xmm0, %eax -; CHECK-SSE-NEXT: xorps %xmm0, %xmm0 -; CHECK-SSE-NEXT: cvtsi2ss %eax, %xmm0 -; CHECK-SSE-NEXT: callq __truncsfhf2@PLT -; CHECK-SSE-NEXT: callq __extendhfsf2@PLT -; CHECK-SSE-NEXT: mulss {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0 -; CHECK-SSE-NEXT: callq __truncsfhf2@PLT -; CHECK-SSE-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill -; CHECK-SSE-NEXT: movss {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 4-byte Reload -; CHECK-SSE-NEXT: # xmm0 = mem[0],zero,zero,zero -; CHECK-SSE-NEXT: callq __extendhfsf2@PLT -; CHECK-SSE-NEXT: mulss {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0 -; CHECK-SSE-NEXT: callq __truncsfhf2@PLT -; CHECK-SSE-NEXT: movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 # 16-byte Reload -; CHECK-SSE-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3] -; CHECK-SSE-NEXT: movdqa %xmm1, %xmm0 -; CHECK-SSE-NEXT: addq $40, %rsp -; CHECK-SSE-NEXT: retq -; -; CHECK-AVX2-LABEL: fmul_pow_shl_cnt_vec_fail_to_large: -; CHECK-AVX2: # %bb.0: -; CHECK-AVX2-NEXT: subq $40, %rsp -; CHECK-AVX2-NEXT: vpmovzxwd {{.*#+}} xmm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero -; CHECK-AVX2-NEXT: vpbroadcastd {{.*#+}} xmm1 = [2,2,2,2] -; CHECK-AVX2-NEXT: vpsllvd %xmm0, %xmm1, %xmm0 -; CHECK-AVX2-NEXT: vmovdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill -; CHECK-AVX2-NEXT: vpextrw $2, %xmm0, %eax -; CHECK-AVX2-NEXT: vcvtsi2ss %eax, %xmm2, %xmm0 -; CHECK-AVX2-NEXT: callq __truncsfhf2@PLT -; CHECK-AVX2-NEXT: vmovss %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 4-byte Spill -; CHECK-AVX2-NEXT: vmovdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload -; CHECK-AVX2-NEXT: vpextrw $0, %xmm0, %eax -; CHECK-AVX2-NEXT: vcvtsi2ss %eax, %xmm2, %xmm0 -; CHECK-AVX2-NEXT: callq __truncsfhf2@PLT -; CHECK-AVX2-NEXT: callq __extendhfsf2@PLT -; CHECK-AVX2-NEXT: vmulss {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-AVX2-NEXT: callq __truncsfhf2@PLT -; CHECK-AVX2-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill -; CHECK-AVX2-NEXT: vmovss {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 4-byte Reload -; CHECK-AVX2-NEXT: # xmm0 = mem[0],zero,zero,zero -; CHECK-AVX2-NEXT: callq __extendhfsf2@PLT -; CHECK-AVX2-NEXT: vmulss {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-AVX2-NEXT: callq __truncsfhf2@PLT -; CHECK-AVX2-NEXT: vmovdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 # 16-byte Reload -; CHECK-AVX2-NEXT: vpunpcklwd {{.*#+}} xmm0 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3] -; CHECK-AVX2-NEXT: addq $40, %rsp -; CHECK-AVX2-NEXT: retq -; -; CHECK-NO-FASTFMA-LABEL: fmul_pow_shl_cnt_vec_fail_to_large: -; CHECK-NO-FASTFMA: # %bb.0: -; CHECK-NO-FASTFMA-NEXT: vpmovzxwd {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero -; CHECK-NO-FASTFMA-NEXT: vmovdqa {{.*#+}} ymm1 = [2,2,0,0,0,0,0,0] -; CHECK-NO-FASTFMA-NEXT: vpsllvd %ymm0, %ymm1, %ymm0 -; CHECK-NO-FASTFMA-NEXT: vpmovdw %zmm0, %ymm1 -; CHECK-NO-FASTFMA-NEXT: vpextrw $0, %xmm0, %eax -; CHECK-NO-FASTFMA-NEXT: vcvtsi2ss %eax, %xmm2, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vcvtps2ph $4, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vmovd %xmm0, %eax -; CHECK-NO-FASTFMA-NEXT: vpinsrw $0, %eax, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vpextrw $1, %xmm1, %eax -; CHECK-NO-FASTFMA-NEXT: vcvtsi2ss %eax, %xmm2, %xmm1 -; CHECK-NO-FASTFMA-NEXT: vcvtps2ph $4, %xmm1, %xmm1 -; CHECK-NO-FASTFMA-NEXT: vmovd %xmm1, %eax -; CHECK-NO-FASTFMA-NEXT: vpinsrw $0, %eax, %xmm0, %xmm1 -; CHECK-NO-FASTFMA-NEXT: vpunpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3] -; CHECK-NO-FASTFMA-NEXT: vmovaps {{.*#+}} xmm1 = [16,0,0,0] -; CHECK-NO-FASTFMA-NEXT: xorl %eax, %eax -; CHECK-NO-FASTFMA-NEXT: vcvtsi2ss %eax, %xmm2, %xmm2 -; CHECK-NO-FASTFMA-NEXT: vcvtps2ph $4, %xmm2, %xmm2 -; CHECK-NO-FASTFMA-NEXT: vmovd %xmm2, %eax -; CHECK-NO-FASTFMA-NEXT: vmovd %eax, %xmm2 -; CHECK-NO-FASTFMA-NEXT: vpbroadcastw %xmm2, %xmm2 -; CHECK-NO-FASTFMA-NEXT: vpermt2ps %zmm0, %zmm1, %zmm2 -; CHECK-NO-FASTFMA-NEXT: vcvtph2ps %xmm2, %ymm0 -; CHECK-NO-FASTFMA-NEXT: vbroadcastss {{.*#+}} ymm1 = [1.5E+1,1.5E+1,1.5E+1,1.5E+1,1.5E+1,1.5E+1,1.5E+1,1.5E+1] -; CHECK-NO-FASTFMA-NEXT: vmulps %ymm1, %ymm0, %ymm0 -; CHECK-NO-FASTFMA-NEXT: vcvtps2ph $4, %ymm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vzeroupper -; CHECK-NO-FASTFMA-NEXT: retq -; -; CHECK-FMA-LABEL: fmul_pow_shl_cnt_vec_fail_to_large: -; CHECK-FMA: # %bb.0: -; CHECK-FMA-NEXT: vpbroadcastw {{.*#+}} xmm1 = [2,2,2,2,2,2,2,2] -; CHECK-FMA-NEXT: vpsllvw %xmm0, %xmm1, %xmm0 -; CHECK-FMA-NEXT: vpextrw $7, %xmm0, %eax -; CHECK-FMA-NEXT: vcvtsi2ss %eax, %xmm2, %xmm1 -; CHECK-FMA-NEXT: vcvtps2ph $4, %xmm1, %xmm1 -; CHECK-FMA-NEXT: vmovd %xmm1, %eax -; CHECK-FMA-NEXT: vpinsrw $0, %eax, %xmm0, %xmm1 -; CHECK-FMA-NEXT: vpextrw $6, %xmm0, %eax -; CHECK-FMA-NEXT: vcvtsi2ss %eax, %xmm2, %xmm2 -; CHECK-FMA-NEXT: vcvtps2ph $4, %xmm2, %xmm2 -; CHECK-FMA-NEXT: vmovd %xmm2, %eax -; CHECK-FMA-NEXT: vpinsrw $0, %eax, %xmm0, %xmm2 -; CHECK-FMA-NEXT: vpextrw $5, %xmm0, %eax -; CHECK-FMA-NEXT: vcvtsi2ss %eax, %xmm3, %xmm3 -; CHECK-FMA-NEXT: vcvtps2ph $4, %xmm3, %xmm3 -; CHECK-FMA-NEXT: vmovd %xmm3, %eax -; CHECK-FMA-NEXT: vpinsrw $0, %eax, %xmm0, %xmm3 -; CHECK-FMA-NEXT: vpextrw $4, %xmm0, %eax -; CHECK-FMA-NEXT: vcvtsi2ss %eax, %xmm4, %xmm4 -; CHECK-FMA-NEXT: vpunpcklwd {{.*#+}} xmm1 = xmm2[0],xmm1[0],xmm2[1],xmm1[1],xmm2[2],xmm1[2],xmm2[3],xmm1[3] -; CHECK-FMA-NEXT: vcvtps2ph $4, %xmm4, %xmm2 -; CHECK-FMA-NEXT: vmovd %xmm2, %eax -; CHECK-FMA-NEXT: vpinsrw $0, %eax, %xmm0, %xmm2 -; CHECK-FMA-NEXT: vpextrw $3, %xmm0, %eax -; CHECK-FMA-NEXT: vcvtsi2ss %eax, %xmm5, %xmm4 -; CHECK-FMA-NEXT: vcvtps2ph $4, %xmm4, %xmm4 -; CHECK-FMA-NEXT: vpunpcklwd {{.*#+}} xmm2 = xmm2[0],xmm3[0],xmm2[1],xmm3[1],xmm2[2],xmm3[2],xmm2[3],xmm3[3] -; CHECK-FMA-NEXT: vmovd %xmm4, %eax -; CHECK-FMA-NEXT: vpinsrw $0, %eax, %xmm0, %xmm3 -; CHECK-FMA-NEXT: vpextrw $2, %xmm0, %eax -; CHECK-FMA-NEXT: vpunpckldq {{.*#+}} xmm1 = xmm2[0],xmm1[0],xmm2[1],xmm1[1] -; CHECK-FMA-NEXT: vcvtsi2ss %eax, %xmm5, %xmm2 -; CHECK-FMA-NEXT: vcvtps2ph $4, %xmm2, %xmm2 -; CHECK-FMA-NEXT: vmovd %xmm2, %eax -; CHECK-FMA-NEXT: vpinsrw $0, %eax, %xmm0, %xmm2 -; CHECK-FMA-NEXT: vpextrw $1, %xmm0, %eax -; CHECK-FMA-NEXT: vcvtsi2ss %eax, %xmm5, %xmm4 -; CHECK-FMA-NEXT: vpunpcklwd {{.*#+}} xmm2 = xmm2[0],xmm3[0],xmm2[1],xmm3[1],xmm2[2],xmm3[2],xmm2[3],xmm3[3] -; CHECK-FMA-NEXT: vcvtps2ph $4, %xmm4, %xmm3 -; CHECK-FMA-NEXT: vmovd %xmm3, %eax -; CHECK-FMA-NEXT: vpinsrw $0, %eax, %xmm0, %xmm3 -; CHECK-FMA-NEXT: vpextrw $0, %xmm0, %eax -; CHECK-FMA-NEXT: vcvtsi2ss %eax, %xmm5, %xmm0 -; CHECK-FMA-NEXT: vcvtps2ph $4, %xmm0, %xmm0 -; CHECK-FMA-NEXT: vmovd %xmm0, %eax -; CHECK-FMA-NEXT: vpinsrw $0, %eax, %xmm0, %xmm0 -; CHECK-FMA-NEXT: vpunpcklwd {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1],xmm0[2],xmm3[2],xmm0[3],xmm3[3] -; CHECK-FMA-NEXT: vpunpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1] -; CHECK-FMA-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0] -; CHECK-FMA-NEXT: vcvtph2ps %xmm0, %ymm0 -; CHECK-FMA-NEXT: vmulps {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to8}, %ymm0, %ymm0 -; CHECK-FMA-NEXT: vcvtps2ph $4, %ymm0, %xmm0 -; CHECK-FMA-NEXT: vzeroupper -; CHECK-FMA-NEXT: retq ; VI-LABEL: fmul_pow_shl_cnt_vec_fail_to_large: ; VI: ; %bb.0: ; VI-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) @@ -1789,52 +1048,6 @@ define <2 x half> @fmul_pow_shl_cnt_vec_fail_to_large(<2 x i16> %cnt) nounwind { } define double @fmul_pow_shl_cnt_fail_maybe_bad_exp(i64 %cnt) nounwind { -; CHECK-SSE-LABEL: fmul_pow_shl_cnt_fail_maybe_bad_exp: -; CHECK-SSE: # %bb.0: -; CHECK-SSE-NEXT: movq %rdi, %rcx -; CHECK-SSE-NEXT: movl $1, %eax -; CHECK-SSE-NEXT: # kill: def $cl killed $cl killed $rcx -; CHECK-SSE-NEXT: shlq %cl, %rax -; CHECK-SSE-NEXT: movq %rax, %xmm1 -; CHECK-SSE-NEXT: punpckldq {{.*#+}} xmm1 = xmm1[0],mem[0],xmm1[1],mem[1] -; CHECK-SSE-NEXT: subpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1 -; CHECK-SSE-NEXT: movapd %xmm1, %xmm0 -; CHECK-SSE-NEXT: unpckhpd {{.*#+}} xmm0 = xmm0[1],xmm1[1] -; CHECK-SSE-NEXT: addsd %xmm1, %xmm0 -; CHECK-SSE-NEXT: mulsd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0 -; CHECK-SSE-NEXT: retq -; -; CHECK-AVX2-LABEL: fmul_pow_shl_cnt_fail_maybe_bad_exp: -; CHECK-AVX2: # %bb.0: -; CHECK-AVX2-NEXT: movq %rdi, %rcx -; CHECK-AVX2-NEXT: movl $1, %eax -; CHECK-AVX2-NEXT: # kill: def $cl killed $cl killed $rcx -; CHECK-AVX2-NEXT: shlq %cl, %rax -; CHECK-AVX2-NEXT: vmovq %rax, %xmm0 -; CHECK-AVX2-NEXT: vpunpckldq {{.*#+}} xmm0 = xmm0[0],mem[0],xmm0[1],mem[1] -; CHECK-AVX2-NEXT: vsubpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-AVX2-NEXT: vshufpd {{.*#+}} xmm1 = xmm0[1,0] -; CHECK-AVX2-NEXT: vaddsd %xmm0, %xmm1, %xmm0 -; CHECK-AVX2-NEXT: vmulsd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-AVX2-NEXT: retq -; -; CHECK-NO-FASTFMA-LABEL: fmul_pow_shl_cnt_fail_maybe_bad_exp: -; CHECK-NO-FASTFMA: # %bb.0: -; CHECK-NO-FASTFMA-NEXT: movq %rdi, %rcx -; CHECK-NO-FASTFMA-NEXT: movl $1, %eax -; CHECK-NO-FASTFMA-NEXT: # kill: def $cl killed $cl killed $rcx -; CHECK-NO-FASTFMA-NEXT: shlq %cl, %rax -; CHECK-NO-FASTFMA-NEXT: vcvtusi2sd %rax, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vmulsd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: retq -; -; CHECK-FMA-LABEL: fmul_pow_shl_cnt_fail_maybe_bad_exp: -; CHECK-FMA: # %bb.0: -; CHECK-FMA-NEXT: movl $1, %eax -; CHECK-FMA-NEXT: shlxq %rdi, %rax, %rax -; CHECK-FMA-NEXT: vcvtusi2sd %rax, %xmm0, %xmm0 -; CHECK-FMA-NEXT: vmulsd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-FMA-NEXT: retq ; VI-LABEL: fmul_pow_shl_cnt_fail_maybe_bad_exp: ; VI: ; %bb.0: ; VI-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) @@ -1883,47 +1096,6 @@ define double @fmul_pow_shl_cnt_fail_maybe_bad_exp(i64 %cnt) nounwind { } define double @fmul_pow_shl_cnt_safe(i16 %cnt) nounwind { -; CHECK-SSE-LABEL: fmul_pow_shl_cnt_safe: -; CHECK-SSE: # %bb.0: -; CHECK-SSE-NEXT: movl %edi, %ecx -; CHECK-SSE-NEXT: movl $1, %eax -; CHECK-SSE-NEXT: # kill: def $cl killed $cl killed $ecx -; CHECK-SSE-NEXT: shll %cl, %eax -; CHECK-SSE-NEXT: movzwl %ax, %eax -; CHECK-SSE-NEXT: cvtsi2sd %eax, %xmm0 -; CHECK-SSE-NEXT: mulsd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0 -; CHECK-SSE-NEXT: retq -; -; CHECK-AVX2-LABEL: fmul_pow_shl_cnt_safe: -; CHECK-AVX2: # %bb.0: -; CHECK-AVX2-NEXT: movl %edi, %ecx -; CHECK-AVX2-NEXT: movl $1, %eax -; CHECK-AVX2-NEXT: # kill: def $cl killed $cl killed $ecx -; CHECK-AVX2-NEXT: shll %cl, %eax -; CHECK-AVX2-NEXT: movzwl %ax, %eax -; CHECK-AVX2-NEXT: vcvtsi2sd %eax, %xmm0, %xmm0 -; CHECK-AVX2-NEXT: vmulsd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-AVX2-NEXT: retq -; -; CHECK-NO-FASTFMA-LABEL: fmul_pow_shl_cnt_safe: -; CHECK-NO-FASTFMA: # %bb.0: -; CHECK-NO-FASTFMA-NEXT: movl %edi, %ecx -; CHECK-NO-FASTFMA-NEXT: movl $1, %eax -; CHECK-NO-FASTFMA-NEXT: # kill: def $cl killed $cl killed $ecx -; CHECK-NO-FASTFMA-NEXT: shll %cl, %eax -; CHECK-NO-FASTFMA-NEXT: movzwl %ax, %eax -; CHECK-NO-FASTFMA-NEXT: vcvtsi2sd %eax, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vmulsd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: retq -; -; CHECK-FMA-LABEL: fmul_pow_shl_cnt_safe: -; CHECK-FMA: # %bb.0: -; CHECK-FMA-NEXT: movl $1, %eax -; CHECK-FMA-NEXT: shlxl %edi, %eax, %eax -; CHECK-FMA-NEXT: movzwl %ax, %eax -; CHECK-FMA-NEXT: vcvtsi2sd %eax, %xmm0, %xmm0 -; CHECK-FMA-NEXT: vmulsd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-FMA-NEXT: retq ; VI-LABEL: fmul_pow_shl_cnt_safe: ; VI: ; %bb.0: ; VI-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) @@ -1964,66 +1136,6 @@ define double @fmul_pow_shl_cnt_safe(i16 %cnt) nounwind { } define <2 x double> @fdiv_pow_shl_cnt_vec(<2 x i64> %cnt) nounwind { -; CHECK-SSE-LABEL: fdiv_pow_shl_cnt_vec: -; CHECK-SSE: # %bb.0: -; CHECK-SSE-NEXT: movdqa {{.*#+}} xmm1 = [1,1] -; CHECK-SSE-NEXT: movdqa %xmm1, %xmm2 -; CHECK-SSE-NEXT: psllq %xmm0, %xmm2 -; CHECK-SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,3,2,3] -; CHECK-SSE-NEXT: psllq %xmm0, %xmm1 -; CHECK-SSE-NEXT: movsd {{.*#+}} xmm1 = xmm2[0],xmm1[1] -; CHECK-SSE-NEXT: movapd {{.*#+}} xmm0 = [4294967295,4294967295] -; CHECK-SSE-NEXT: andpd %xmm1, %xmm0 -; CHECK-SSE-NEXT: orpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0 -; CHECK-SSE-NEXT: psrlq $32, %xmm1 -; CHECK-SSE-NEXT: por {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1 -; CHECK-SSE-NEXT: subpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1 -; CHECK-SSE-NEXT: addpd %xmm0, %xmm1 -; CHECK-SSE-NEXT: movapd {{.*#+}} xmm0 = [1.0E+0,1.0E+0] -; CHECK-SSE-NEXT: divpd %xmm1, %xmm0 -; CHECK-SSE-NEXT: retq -; -; CHECK-AVX2-LABEL: fdiv_pow_shl_cnt_vec: -; CHECK-AVX2: # %bb.0: -; CHECK-AVX2-NEXT: vpbroadcastq {{.*#+}} xmm1 = [1,1] -; CHECK-AVX2-NEXT: vpsllvq %xmm0, %xmm1, %xmm0 -; CHECK-AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1 -; CHECK-AVX2-NEXT: vpblendd {{.*#+}} xmm1 = xmm0[0],xmm1[1],xmm0[2],xmm1[3] -; CHECK-AVX2-NEXT: vpor {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1, %xmm1 -; CHECK-AVX2-NEXT: vpsrlq $32, %xmm0, %xmm0 -; CHECK-AVX2-NEXT: vpor {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-AVX2-NEXT: vsubpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-AVX2-NEXT: vaddpd %xmm0, %xmm1, %xmm0 -; CHECK-AVX2-NEXT: vmovddup {{.*#+}} xmm1 = [1.0E+0,1.0E+0] -; CHECK-AVX2-NEXT: # xmm1 = mem[0,0] -; CHECK-AVX2-NEXT: vdivpd %xmm0, %xmm1, %xmm0 -; CHECK-AVX2-NEXT: retq -; -; CHECK-NO-FASTFMA-LABEL: fdiv_pow_shl_cnt_vec: -; CHECK-NO-FASTFMA: # %bb.0: -; CHECK-NO-FASTFMA-NEXT: vpbroadcastq {{.*#+}} xmm1 = [1,1] -; CHECK-NO-FASTFMA-NEXT: vpsllvq %xmm0, %xmm1, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vpxor %xmm1, %xmm1, %xmm1 -; CHECK-NO-FASTFMA-NEXT: vpblendd {{.*#+}} xmm1 = xmm0[0],xmm1[1],xmm0[2],xmm1[3] -; CHECK-NO-FASTFMA-NEXT: vpor {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1, %xmm1 -; CHECK-NO-FASTFMA-NEXT: vpsrlq $32, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vpor {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vsubpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vaddpd %xmm0, %xmm1, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vmovddup {{.*#+}} xmm1 = [1.0E+0,1.0E+0] -; CHECK-NO-FASTFMA-NEXT: # xmm1 = mem[0,0] -; CHECK-NO-FASTFMA-NEXT: vdivpd %xmm0, %xmm1, %xmm0 -; CHECK-NO-FASTFMA-NEXT: retq -; -; CHECK-FMA-LABEL: fdiv_pow_shl_cnt_vec: -; CHECK-FMA: # %bb.0: -; CHECK-FMA-NEXT: vpbroadcastq {{.*#+}} xmm1 = [1,1] -; CHECK-FMA-NEXT: vpsllvq %xmm0, %xmm1, %xmm0 -; CHECK-FMA-NEXT: vcvtuqq2pd %xmm0, %xmm0 -; CHECK-FMA-NEXT: vmovddup {{.*#+}} xmm1 = [1.0E+0,1.0E+0] -; CHECK-FMA-NEXT: # xmm1 = mem[0,0] -; CHECK-FMA-NEXT: vdivpd %xmm0, %xmm1, %xmm0 -; CHECK-FMA-NEXT: retq ; VI-LABEL: fdiv_pow_shl_cnt_vec: ; VI: ; %bb.0: ; VI-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) @@ -2065,94 +1177,6 @@ define <2 x double> @fdiv_pow_shl_cnt_vec(<2 x i64> %cnt) nounwind { } define <2 x float> @fdiv_pow_shl_cnt_vec_with_expensive_cast(<2 x i64> %cnt) nounwind { -; CHECK-SSE-LABEL: fdiv_pow_shl_cnt_vec_with_expensive_cast: -; CHECK-SSE: # %bb.0: -; CHECK-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3] -; CHECK-SSE-NEXT: movdqa {{.*#+}} xmm3 = [1,1] -; CHECK-SSE-NEXT: movdqa %xmm3, %xmm2 -; CHECK-SSE-NEXT: psllq %xmm1, %xmm2 -; CHECK-SSE-NEXT: psllq %xmm0, %xmm3 -; CHECK-SSE-NEXT: movq %xmm3, %rax -; CHECK-SSE-NEXT: testq %rax, %rax -; CHECK-SSE-NEXT: js .LBB15_1 -; CHECK-SSE-NEXT: # %bb.2: -; CHECK-SSE-NEXT: xorps %xmm1, %xmm1 -; CHECK-SSE-NEXT: cvtsi2ss %rax, %xmm1 -; CHECK-SSE-NEXT: jmp .LBB15_3 -; CHECK-SSE-NEXT: .LBB15_1: -; CHECK-SSE-NEXT: movq %rax, %rcx -; CHECK-SSE-NEXT: shrq %rcx -; CHECK-SSE-NEXT: andl $1, %eax -; CHECK-SSE-NEXT: orq %rcx, %rax -; CHECK-SSE-NEXT: xorps %xmm1, %xmm1 -; CHECK-SSE-NEXT: cvtsi2ss %rax, %xmm1 -; CHECK-SSE-NEXT: addss %xmm1, %xmm1 -; CHECK-SSE-NEXT: .LBB15_3: -; CHECK-SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm2[2,3,2,3] -; CHECK-SSE-NEXT: movq %xmm0, %rax -; CHECK-SSE-NEXT: testq %rax, %rax -; CHECK-SSE-NEXT: js .LBB15_4 -; CHECK-SSE-NEXT: # %bb.5: -; CHECK-SSE-NEXT: xorps %xmm0, %xmm0 -; CHECK-SSE-NEXT: cvtsi2ss %rax, %xmm0 -; CHECK-SSE-NEXT: jmp .LBB15_6 -; CHECK-SSE-NEXT: .LBB15_4: -; CHECK-SSE-NEXT: movq %rax, %rcx -; CHECK-SSE-NEXT: shrq %rcx -; CHECK-SSE-NEXT: andl $1, %eax -; CHECK-SSE-NEXT: orq %rcx, %rax -; CHECK-SSE-NEXT: xorps %xmm0, %xmm0 -; CHECK-SSE-NEXT: cvtsi2ss %rax, %xmm0 -; CHECK-SSE-NEXT: addss %xmm0, %xmm0 -; CHECK-SSE-NEXT: .LBB15_6: -; CHECK-SSE-NEXT: unpcklps {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1] -; CHECK-SSE-NEXT: movaps {{.*#+}} xmm0 = <1.0E+0,1.0E+0,u,u> -; CHECK-SSE-NEXT: divps %xmm1, %xmm0 -; CHECK-SSE-NEXT: retq -; -; CHECK-AVX2-LABEL: fdiv_pow_shl_cnt_vec_with_expensive_cast: -; CHECK-AVX2: # %bb.0: -; CHECK-AVX2-NEXT: vpbroadcastq {{.*#+}} xmm1 = [1,1] -; CHECK-AVX2-NEXT: vpsllvq %xmm0, %xmm1, %xmm0 -; CHECK-AVX2-NEXT: vpand %xmm1, %xmm0, %xmm1 -; CHECK-AVX2-NEXT: vpsrlq $1, %xmm0, %xmm2 -; CHECK-AVX2-NEXT: vpor %xmm1, %xmm2, %xmm1 -; CHECK-AVX2-NEXT: vblendvpd %xmm0, %xmm1, %xmm0, %xmm1 -; CHECK-AVX2-NEXT: vpextrq $1, %xmm1, %rax -; CHECK-AVX2-NEXT: vcvtsi2ss %rax, %xmm3, %xmm2 -; CHECK-AVX2-NEXT: vmovq %xmm1, %rax -; CHECK-AVX2-NEXT: vcvtsi2ss %rax, %xmm3, %xmm1 -; CHECK-AVX2-NEXT: vinsertps {{.*#+}} xmm1 = xmm1[0],xmm2[0],zero,zero -; CHECK-AVX2-NEXT: vaddps %xmm1, %xmm1, %xmm2 -; CHECK-AVX2-NEXT: vpxor %xmm3, %xmm3, %xmm3 -; CHECK-AVX2-NEXT: vpcmpgtq %xmm0, %xmm3, %xmm0 -; CHECK-AVX2-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[1,3,2,3] -; CHECK-AVX2-NEXT: vblendvps %xmm0, %xmm2, %xmm1, %xmm0 -; CHECK-AVX2-NEXT: vbroadcastss {{.*#+}} xmm1 = [1.0E+0,1.0E+0,1.0E+0,1.0E+0] -; CHECK-AVX2-NEXT: vdivps %xmm0, %xmm1, %xmm0 -; CHECK-AVX2-NEXT: retq -; -; CHECK-NO-FASTFMA-LABEL: fdiv_pow_shl_cnt_vec_with_expensive_cast: -; CHECK-NO-FASTFMA: # %bb.0: -; CHECK-NO-FASTFMA-NEXT: vpbroadcastq {{.*#+}} xmm1 = [1,1] -; CHECK-NO-FASTFMA-NEXT: vpsllvq %xmm0, %xmm1, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vpextrq $1, %xmm0, %rax -; CHECK-NO-FASTFMA-NEXT: vcvtusi2ss %rax, %xmm2, %xmm1 -; CHECK-NO-FASTFMA-NEXT: vmovq %xmm0, %rax -; CHECK-NO-FASTFMA-NEXT: vcvtusi2ss %rax, %xmm2, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vinsertps {{.*#+}} xmm0 = xmm0[0],xmm1[0],zero,zero -; CHECK-NO-FASTFMA-NEXT: vbroadcastss {{.*#+}} xmm1 = [1.0E+0,1.0E+0,1.0E+0,1.0E+0] -; CHECK-NO-FASTFMA-NEXT: vdivps %xmm0, %xmm1, %xmm0 -; CHECK-NO-FASTFMA-NEXT: retq -; -; CHECK-FMA-LABEL: fdiv_pow_shl_cnt_vec_with_expensive_cast: -; CHECK-FMA: # %bb.0: -; CHECK-FMA-NEXT: vpbroadcastq {{.*#+}} xmm1 = [1,1] -; CHECK-FMA-NEXT: vpsllvq %xmm0, %xmm1, %xmm0 -; CHECK-FMA-NEXT: vcvtuqq2ps %xmm0, %xmm0 -; CHECK-FMA-NEXT: vbroadcastss {{.*#+}} xmm1 = [1.0E+0,1.0E+0,1.0E+0,1.0E+0] -; CHECK-FMA-NEXT: vdivps %xmm0, %xmm1, %xmm0 -; CHECK-FMA-NEXT: retq ; VI-LABEL: fdiv_pow_shl_cnt_vec_with_expensive_cast: ; VI: ; %bb.0: ; VI-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) @@ -2187,65 +1211,6 @@ define <2 x float> @fdiv_pow_shl_cnt_vec_with_expensive_cast(<2 x i64> %cnt) nou } define float @fdiv_pow_shl_cnt_fail_maybe_z(i64 %cnt) nounwind { -; CHECK-SSE-LABEL: fdiv_pow_shl_cnt_fail_maybe_z: -; CHECK-SSE: # %bb.0: -; CHECK-SSE-NEXT: movq %rdi, %rcx -; CHECK-SSE-NEXT: movl $8, %eax -; CHECK-SSE-NEXT: # kill: def $cl killed $cl killed $rcx -; CHECK-SSE-NEXT: shlq %cl, %rax -; CHECK-SSE-NEXT: testq %rax, %rax -; CHECK-SSE-NEXT: js .LBB16_1 -; CHECK-SSE-NEXT: # %bb.2: -; CHECK-SSE-NEXT: cvtsi2ss %rax, %xmm1 -; CHECK-SSE-NEXT: jmp .LBB16_3 -; CHECK-SSE-NEXT: .LBB16_1: -; CHECK-SSE-NEXT: shrq %rax -; CHECK-SSE-NEXT: cvtsi2ss %rax, %xmm1 -; CHECK-SSE-NEXT: addss %xmm1, %xmm1 -; CHECK-SSE-NEXT: .LBB16_3: -; CHECK-SSE-NEXT: movss {{.*#+}} xmm0 = mem[0],zero,zero,zero -; CHECK-SSE-NEXT: divss %xmm1, %xmm0 -; CHECK-SSE-NEXT: retq -; -; CHECK-AVX2-LABEL: fdiv_pow_shl_cnt_fail_maybe_z: -; CHECK-AVX2: # %bb.0: -; CHECK-AVX2-NEXT: movq %rdi, %rcx -; CHECK-AVX2-NEXT: movl $8, %eax -; CHECK-AVX2-NEXT: # kill: def $cl killed $cl killed $rcx -; CHECK-AVX2-NEXT: shlq %cl, %rax -; CHECK-AVX2-NEXT: testq %rax, %rax -; CHECK-AVX2-NEXT: js .LBB16_1 -; CHECK-AVX2-NEXT: # %bb.2: -; CHECK-AVX2-NEXT: vcvtsi2ss %rax, %xmm0, %xmm0 -; CHECK-AVX2-NEXT: jmp .LBB16_3 -; CHECK-AVX2-NEXT: .LBB16_1: -; CHECK-AVX2-NEXT: shrq %rax -; CHECK-AVX2-NEXT: vcvtsi2ss %rax, %xmm0, %xmm0 -; CHECK-AVX2-NEXT: vaddss %xmm0, %xmm0, %xmm0 -; CHECK-AVX2-NEXT: .LBB16_3: -; CHECK-AVX2-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-AVX2-NEXT: vdivss %xmm0, %xmm1, %xmm0 -; CHECK-AVX2-NEXT: retq -; -; CHECK-NO-FASTFMA-LABEL: fdiv_pow_shl_cnt_fail_maybe_z: -; CHECK-NO-FASTFMA: # %bb.0: -; CHECK-NO-FASTFMA-NEXT: movq %rdi, %rcx -; CHECK-NO-FASTFMA-NEXT: movl $8, %eax -; CHECK-NO-FASTFMA-NEXT: # kill: def $cl killed $cl killed $rcx -; CHECK-NO-FASTFMA-NEXT: shlq %cl, %rax -; CHECK-NO-FASTFMA-NEXT: vcvtusi2ss %rax, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-NO-FASTFMA-NEXT: vdivss %xmm0, %xmm1, %xmm0 -; CHECK-NO-FASTFMA-NEXT: retq -; -; CHECK-FMA-LABEL: fdiv_pow_shl_cnt_fail_maybe_z: -; CHECK-FMA: # %bb.0: -; CHECK-FMA-NEXT: movl $8, %eax -; CHECK-FMA-NEXT: shlxq %rdi, %rax, %rax -; CHECK-FMA-NEXT: vcvtusi2ss %rax, %xmm0, %xmm0 -; CHECK-FMA-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-FMA-NEXT: vdivss %xmm0, %xmm1, %xmm0 -; CHECK-FMA-NEXT: retq ; VI-LABEL: fdiv_pow_shl_cnt_fail_maybe_z: ; VI: ; %bb.0: ; VI-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) @@ -2337,47 +1302,6 @@ define float @fdiv_pow_shl_cnt_fail_maybe_z(i64 %cnt) nounwind { } define float @fdiv_pow_shl_cnt_fail_neg_int(i64 %cnt) nounwind { -; CHECK-SSE-LABEL: fdiv_pow_shl_cnt_fail_neg_int: -; CHECK-SSE: # %bb.0: -; CHECK-SSE-NEXT: movq %rdi, %rcx -; CHECK-SSE-NEXT: movl $8, %eax -; CHECK-SSE-NEXT: # kill: def $cl killed $cl killed $rcx -; CHECK-SSE-NEXT: shlq %cl, %rax -; CHECK-SSE-NEXT: cvtsi2ss %rax, %xmm1 -; CHECK-SSE-NEXT: movss {{.*#+}} xmm0 = mem[0],zero,zero,zero -; CHECK-SSE-NEXT: divss %xmm1, %xmm0 -; CHECK-SSE-NEXT: retq -; -; CHECK-AVX2-LABEL: fdiv_pow_shl_cnt_fail_neg_int: -; CHECK-AVX2: # %bb.0: -; CHECK-AVX2-NEXT: movq %rdi, %rcx -; CHECK-AVX2-NEXT: movl $8, %eax -; CHECK-AVX2-NEXT: # kill: def $cl killed $cl killed $rcx -; CHECK-AVX2-NEXT: shlq %cl, %rax -; CHECK-AVX2-NEXT: vcvtsi2ss %rax, %xmm0, %xmm0 -; CHECK-AVX2-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-AVX2-NEXT: vdivss %xmm0, %xmm1, %xmm0 -; CHECK-AVX2-NEXT: retq -; -; CHECK-NO-FASTFMA-LABEL: fdiv_pow_shl_cnt_fail_neg_int: -; CHECK-NO-FASTFMA: # %bb.0: -; CHECK-NO-FASTFMA-NEXT: movq %rdi, %rcx -; CHECK-NO-FASTFMA-NEXT: movl $8, %eax -; CHECK-NO-FASTFMA-NEXT: # kill: def $cl killed $cl killed $rcx -; CHECK-NO-FASTFMA-NEXT: shlq %cl, %rax -; CHECK-NO-FASTFMA-NEXT: vcvtsi2ss %rax, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-NO-FASTFMA-NEXT: vdivss %xmm0, %xmm1, %xmm0 -; CHECK-NO-FASTFMA-NEXT: retq -; -; CHECK-FMA-LABEL: fdiv_pow_shl_cnt_fail_neg_int: -; CHECK-FMA: # %bb.0: -; CHECK-FMA-NEXT: movl $8, %eax -; CHECK-FMA-NEXT: shlxq %rdi, %rax, %rax -; CHECK-FMA-NEXT: vcvtsi2ss %rax, %xmm0, %xmm0 -; CHECK-FMA-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-FMA-NEXT: vdivss %xmm0, %xmm1, %xmm0 -; CHECK-FMA-NEXT: retq ; VI-LABEL: fdiv_pow_shl_cnt_fail_neg_int: ; VI: ; %bb.0: ; VI-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) @@ -2483,51 +1407,6 @@ define float @fdiv_pow_shl_cnt_fail_neg_int(i64 %cnt) nounwind { } define float @fdiv_pow_shl_cnt(i64 %cnt_in) nounwind { -; CHECK-SSE-LABEL: fdiv_pow_shl_cnt: -; CHECK-SSE: # %bb.0: -; CHECK-SSE-NEXT: movq %rdi, %rcx -; CHECK-SSE-NEXT: andb $31, %cl -; CHECK-SSE-NEXT: movl $8, %eax -; CHECK-SSE-NEXT: # kill: def $cl killed $cl killed $rcx -; CHECK-SSE-NEXT: shlq %cl, %rax -; CHECK-SSE-NEXT: cvtsi2ss %rax, %xmm1 -; CHECK-SSE-NEXT: movss {{.*#+}} xmm0 = mem[0],zero,zero,zero -; CHECK-SSE-NEXT: divss %xmm1, %xmm0 -; CHECK-SSE-NEXT: retq -; -; CHECK-AVX2-LABEL: fdiv_pow_shl_cnt: -; CHECK-AVX2: # %bb.0: -; CHECK-AVX2-NEXT: movq %rdi, %rcx -; CHECK-AVX2-NEXT: andb $31, %cl -; CHECK-AVX2-NEXT: movl $8, %eax -; CHECK-AVX2-NEXT: # kill: def $cl killed $cl killed $rcx -; CHECK-AVX2-NEXT: shlq %cl, %rax -; CHECK-AVX2-NEXT: vcvtsi2ss %rax, %xmm0, %xmm0 -; CHECK-AVX2-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-AVX2-NEXT: vdivss %xmm0, %xmm1, %xmm0 -; CHECK-AVX2-NEXT: retq -; -; CHECK-NO-FASTFMA-LABEL: fdiv_pow_shl_cnt: -; CHECK-NO-FASTFMA: # %bb.0: -; CHECK-NO-FASTFMA-NEXT: movq %rdi, %rcx -; CHECK-NO-FASTFMA-NEXT: andb $31, %cl -; CHECK-NO-FASTFMA-NEXT: movl $8, %eax -; CHECK-NO-FASTFMA-NEXT: # kill: def $cl killed $cl killed $rcx -; CHECK-NO-FASTFMA-NEXT: shlq %cl, %rax -; CHECK-NO-FASTFMA-NEXT: vcvtsi2ss %rax, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-NO-FASTFMA-NEXT: vdivss %xmm0, %xmm1, %xmm0 -; CHECK-NO-FASTFMA-NEXT: retq -; -; CHECK-FMA-LABEL: fdiv_pow_shl_cnt: -; CHECK-FMA: # %bb.0: -; CHECK-FMA-NEXT: andb $31, %dil -; CHECK-FMA-NEXT: movl $8, %eax -; CHECK-FMA-NEXT: shlxq %rdi, %rax, %rax -; CHECK-FMA-NEXT: vcvtsi2ss %rax, %xmm0, %xmm0 -; CHECK-FMA-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-FMA-NEXT: vdivss %xmm0, %xmm1, %xmm0 -; CHECK-FMA-NEXT: retq ; VI-LABEL: fdiv_pow_shl_cnt: ; VI: ; %bb.0: ; VI-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) @@ -2560,70 +1439,6 @@ define float @fdiv_pow_shl_cnt(i64 %cnt_in) nounwind { } define half @fdiv_pow_shl_cnt_fail_out_of_bounds(i32 %cnt) nounwind { -; CHECK-SSE-LABEL: fdiv_pow_shl_cnt_fail_out_of_bounds: -; CHECK-SSE: # %bb.0: -; CHECK-SSE-NEXT: pushq %rax -; CHECK-SSE-NEXT: movl %edi, %ecx -; CHECK-SSE-NEXT: movl $1, %eax -; CHECK-SSE-NEXT: # kill: def $cl killed $cl killed $ecx -; CHECK-SSE-NEXT: shll %cl, %eax -; CHECK-SSE-NEXT: cvtsi2ss %rax, %xmm0 -; CHECK-SSE-NEXT: callq __truncsfhf2@PLT -; CHECK-SSE-NEXT: callq __extendhfsf2@PLT -; CHECK-SSE-NEXT: movss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-SSE-NEXT: divss %xmm0, %xmm1 -; CHECK-SSE-NEXT: movaps %xmm1, %xmm0 -; CHECK-SSE-NEXT: callq __truncsfhf2@PLT -; CHECK-SSE-NEXT: popq %rax -; CHECK-SSE-NEXT: retq -; -; CHECK-AVX2-LABEL: fdiv_pow_shl_cnt_fail_out_of_bounds: -; CHECK-AVX2: # %bb.0: -; CHECK-AVX2-NEXT: pushq %rax -; CHECK-AVX2-NEXT: movl %edi, %ecx -; CHECK-AVX2-NEXT: movl $1, %eax -; CHECK-AVX2-NEXT: # kill: def $cl killed $cl killed $ecx -; CHECK-AVX2-NEXT: shll %cl, %eax -; CHECK-AVX2-NEXT: vcvtsi2ss %rax, %xmm0, %xmm0 -; CHECK-AVX2-NEXT: callq __truncsfhf2@PLT -; CHECK-AVX2-NEXT: callq __extendhfsf2@PLT -; CHECK-AVX2-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-AVX2-NEXT: vdivss %xmm0, %xmm1, %xmm0 -; CHECK-AVX2-NEXT: callq __truncsfhf2@PLT -; CHECK-AVX2-NEXT: popq %rax -; CHECK-AVX2-NEXT: retq -; -; CHECK-NO-FASTFMA-LABEL: fdiv_pow_shl_cnt_fail_out_of_bounds: -; CHECK-NO-FASTFMA: # %bb.0: -; CHECK-NO-FASTFMA-NEXT: movl %edi, %ecx -; CHECK-NO-FASTFMA-NEXT: movl $1, %eax -; CHECK-NO-FASTFMA-NEXT: # kill: def $cl killed $cl killed $ecx -; CHECK-NO-FASTFMA-NEXT: shll %cl, %eax -; CHECK-NO-FASTFMA-NEXT: vcvtusi2ss %eax, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vcvtps2ph $4, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vpmovzxwq {{.*#+}} xmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero -; CHECK-NO-FASTFMA-NEXT: vcvtph2ps %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-NO-FASTFMA-NEXT: vdivss %xmm0, %xmm1, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vcvtps2ph $4, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vmovd %xmm0, %eax -; CHECK-NO-FASTFMA-NEXT: vpinsrw $0, %eax, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: retq -; -; CHECK-FMA-LABEL: fdiv_pow_shl_cnt_fail_out_of_bounds: -; CHECK-FMA: # %bb.0: -; CHECK-FMA-NEXT: movl $1, %eax -; CHECK-FMA-NEXT: shlxl %edi, %eax, %eax -; CHECK-FMA-NEXT: vcvtusi2ss %eax, %xmm0, %xmm0 -; CHECK-FMA-NEXT: vcvtps2ph $4, %xmm0, %xmm0 -; CHECK-FMA-NEXT: vpmovzxwq {{.*#+}} xmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero -; CHECK-FMA-NEXT: vcvtph2ps %xmm0, %xmm0 -; CHECK-FMA-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-FMA-NEXT: vdivss %xmm0, %xmm1, %xmm0 -; CHECK-FMA-NEXT: vcvtps2ph $4, %xmm0, %xmm0 -; CHECK-FMA-NEXT: vmovd %xmm0, %eax -; CHECK-FMA-NEXT: vpinsrw $0, %eax, %xmm0, %xmm0 -; CHECK-FMA-NEXT: retq ; VI-LABEL: fdiv_pow_shl_cnt_fail_out_of_bounds: ; VI: ; %bb.0: ; VI-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) @@ -2674,74 +1489,6 @@ define half @fdiv_pow_shl_cnt_fail_out_of_bounds(i32 %cnt) nounwind { } define half @fdiv_pow_shl_cnt_in_bounds(i16 %cnt) nounwind { -; CHECK-SSE-LABEL: fdiv_pow_shl_cnt_in_bounds: -; CHECK-SSE: # %bb.0: -; CHECK-SSE-NEXT: pushq %rax -; CHECK-SSE-NEXT: movl %edi, %ecx -; CHECK-SSE-NEXT: movl $1, %eax -; CHECK-SSE-NEXT: # kill: def $cl killed $cl killed $ecx -; CHECK-SSE-NEXT: shll %cl, %eax -; CHECK-SSE-NEXT: movzwl %ax, %eax -; CHECK-SSE-NEXT: cvtsi2ss %eax, %xmm0 -; CHECK-SSE-NEXT: callq __truncsfhf2@PLT -; CHECK-SSE-NEXT: callq __extendhfsf2@PLT -; CHECK-SSE-NEXT: movss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-SSE-NEXT: divss %xmm0, %xmm1 -; CHECK-SSE-NEXT: movaps %xmm1, %xmm0 -; CHECK-SSE-NEXT: callq __truncsfhf2@PLT -; CHECK-SSE-NEXT: popq %rax -; CHECK-SSE-NEXT: retq -; -; CHECK-AVX2-LABEL: fdiv_pow_shl_cnt_in_bounds: -; CHECK-AVX2: # %bb.0: -; CHECK-AVX2-NEXT: pushq %rax -; CHECK-AVX2-NEXT: movl %edi, %ecx -; CHECK-AVX2-NEXT: movl $1, %eax -; CHECK-AVX2-NEXT: # kill: def $cl killed $cl killed $ecx -; CHECK-AVX2-NEXT: shll %cl, %eax -; CHECK-AVX2-NEXT: movzwl %ax, %eax -; CHECK-AVX2-NEXT: vcvtsi2ss %eax, %xmm0, %xmm0 -; CHECK-AVX2-NEXT: callq __truncsfhf2@PLT -; CHECK-AVX2-NEXT: callq __extendhfsf2@PLT -; CHECK-AVX2-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-AVX2-NEXT: vdivss %xmm0, %xmm1, %xmm0 -; CHECK-AVX2-NEXT: callq __truncsfhf2@PLT -; CHECK-AVX2-NEXT: popq %rax -; CHECK-AVX2-NEXT: retq -; -; CHECK-NO-FASTFMA-LABEL: fdiv_pow_shl_cnt_in_bounds: -; CHECK-NO-FASTFMA: # %bb.0: -; CHECK-NO-FASTFMA-NEXT: movl %edi, %ecx -; CHECK-NO-FASTFMA-NEXT: movl $1, %eax -; CHECK-NO-FASTFMA-NEXT: # kill: def $cl killed $cl killed $ecx -; CHECK-NO-FASTFMA-NEXT: shll %cl, %eax -; CHECK-NO-FASTFMA-NEXT: movzwl %ax, %eax -; CHECK-NO-FASTFMA-NEXT: vcvtsi2ss %eax, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vcvtps2ph $4, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vpmovzxwq {{.*#+}} xmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero -; CHECK-NO-FASTFMA-NEXT: vcvtph2ps %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-NO-FASTFMA-NEXT: vdivss %xmm0, %xmm1, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vcvtps2ph $4, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vmovd %xmm0, %eax -; CHECK-NO-FASTFMA-NEXT: vpinsrw $0, %eax, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: retq -; -; CHECK-FMA-LABEL: fdiv_pow_shl_cnt_in_bounds: -; CHECK-FMA: # %bb.0: -; CHECK-FMA-NEXT: movl $1, %eax -; CHECK-FMA-NEXT: shlxl %edi, %eax, %eax -; CHECK-FMA-NEXT: movzwl %ax, %eax -; CHECK-FMA-NEXT: vcvtsi2ss %eax, %xmm0, %xmm0 -; CHECK-FMA-NEXT: vcvtps2ph $4, %xmm0, %xmm0 -; CHECK-FMA-NEXT: vpmovzxwq {{.*#+}} xmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero -; CHECK-FMA-NEXT: vcvtph2ps %xmm0, %xmm0 -; CHECK-FMA-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-FMA-NEXT: vdivss %xmm0, %xmm1, %xmm0 -; CHECK-FMA-NEXT: vcvtps2ph $4, %xmm0, %xmm0 -; CHECK-FMA-NEXT: vmovd %xmm0, %eax -; CHECK-FMA-NEXT: vpinsrw $0, %eax, %xmm0, %xmm0 -; CHECK-FMA-NEXT: retq ; VI-LABEL: fdiv_pow_shl_cnt_in_bounds: ; VI: ; %bb.0: ; VI-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) @@ -2770,74 +1517,6 @@ define half @fdiv_pow_shl_cnt_in_bounds(i16 %cnt) nounwind { } define half @fdiv_pow_shl_cnt_in_bounds2(i16 %cnt) nounwind { -; CHECK-SSE-LABEL: fdiv_pow_shl_cnt_in_bounds2: -; CHECK-SSE: # %bb.0: -; CHECK-SSE-NEXT: pushq %rax -; CHECK-SSE-NEXT: movl %edi, %ecx -; CHECK-SSE-NEXT: movl $1, %eax -; CHECK-SSE-NEXT: # kill: def $cl killed $cl killed $ecx -; CHECK-SSE-NEXT: shll %cl, %eax -; CHECK-SSE-NEXT: movzwl %ax, %eax -; CHECK-SSE-NEXT: cvtsi2ss %eax, %xmm0 -; CHECK-SSE-NEXT: callq __truncsfhf2@PLT -; CHECK-SSE-NEXT: callq __extendhfsf2@PLT -; CHECK-SSE-NEXT: movss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-SSE-NEXT: divss %xmm0, %xmm1 -; CHECK-SSE-NEXT: movaps %xmm1, %xmm0 -; CHECK-SSE-NEXT: callq __truncsfhf2@PLT -; CHECK-SSE-NEXT: popq %rax -; CHECK-SSE-NEXT: retq -; -; CHECK-AVX2-LABEL: fdiv_pow_shl_cnt_in_bounds2: -; CHECK-AVX2: # %bb.0: -; CHECK-AVX2-NEXT: pushq %rax -; CHECK-AVX2-NEXT: movl %edi, %ecx -; CHECK-AVX2-NEXT: movl $1, %eax -; CHECK-AVX2-NEXT: # kill: def $cl killed $cl killed $ecx -; CHECK-AVX2-NEXT: shll %cl, %eax -; CHECK-AVX2-NEXT: movzwl %ax, %eax -; CHECK-AVX2-NEXT: vcvtsi2ss %eax, %xmm0, %xmm0 -; CHECK-AVX2-NEXT: callq __truncsfhf2@PLT -; CHECK-AVX2-NEXT: callq __extendhfsf2@PLT -; CHECK-AVX2-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-AVX2-NEXT: vdivss %xmm0, %xmm1, %xmm0 -; CHECK-AVX2-NEXT: callq __truncsfhf2@PLT -; CHECK-AVX2-NEXT: popq %rax -; CHECK-AVX2-NEXT: retq -; -; CHECK-NO-FASTFMA-LABEL: fdiv_pow_shl_cnt_in_bounds2: -; CHECK-NO-FASTFMA: # %bb.0: -; CHECK-NO-FASTFMA-NEXT: movl %edi, %ecx -; CHECK-NO-FASTFMA-NEXT: movl $1, %eax -; CHECK-NO-FASTFMA-NEXT: # kill: def $cl killed $cl killed $ecx -; CHECK-NO-FASTFMA-NEXT: shll %cl, %eax -; CHECK-NO-FASTFMA-NEXT: movzwl %ax, %eax -; CHECK-NO-FASTFMA-NEXT: vcvtsi2ss %eax, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vcvtps2ph $4, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vpmovzxwq {{.*#+}} xmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero -; CHECK-NO-FASTFMA-NEXT: vcvtph2ps %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-NO-FASTFMA-NEXT: vdivss %xmm0, %xmm1, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vcvtps2ph $4, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vmovd %xmm0, %eax -; CHECK-NO-FASTFMA-NEXT: vpinsrw $0, %eax, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: retq -; -; CHECK-FMA-LABEL: fdiv_pow_shl_cnt_in_bounds2: -; CHECK-FMA: # %bb.0: -; CHECK-FMA-NEXT: movl $1, %eax -; CHECK-FMA-NEXT: shlxl %edi, %eax, %eax -; CHECK-FMA-NEXT: movzwl %ax, %eax -; CHECK-FMA-NEXT: vcvtsi2ss %eax, %xmm0, %xmm0 -; CHECK-FMA-NEXT: vcvtps2ph $4, %xmm0, %xmm0 -; CHECK-FMA-NEXT: vpmovzxwq {{.*#+}} xmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero -; CHECK-FMA-NEXT: vcvtph2ps %xmm0, %xmm0 -; CHECK-FMA-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-FMA-NEXT: vdivss %xmm0, %xmm1, %xmm0 -; CHECK-FMA-NEXT: vcvtps2ph $4, %xmm0, %xmm0 -; CHECK-FMA-NEXT: vmovd %xmm0, %eax -; CHECK-FMA-NEXT: vpinsrw $0, %eax, %xmm0, %xmm0 -; CHECK-FMA-NEXT: retq ; VI-LABEL: fdiv_pow_shl_cnt_in_bounds2: ; VI: ; %bb.0: ; VI-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) @@ -2866,74 +1545,6 @@ define half @fdiv_pow_shl_cnt_in_bounds2(i16 %cnt) nounwind { } define half @fdiv_pow_shl_cnt_fail_out_of_bound2(i16 %cnt) nounwind { -; CHECK-SSE-LABEL: fdiv_pow_shl_cnt_fail_out_of_bound2: -; CHECK-SSE: # %bb.0: -; CHECK-SSE-NEXT: pushq %rax -; CHECK-SSE-NEXT: movl %edi, %ecx -; CHECK-SSE-NEXT: movl $1, %eax -; CHECK-SSE-NEXT: # kill: def $cl killed $cl killed $ecx -; CHECK-SSE-NEXT: shll %cl, %eax -; CHECK-SSE-NEXT: movzwl %ax, %eax -; CHECK-SSE-NEXT: cvtsi2ss %eax, %xmm0 -; CHECK-SSE-NEXT: callq __truncsfhf2@PLT -; CHECK-SSE-NEXT: callq __extendhfsf2@PLT -; CHECK-SSE-NEXT: movss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-SSE-NEXT: divss %xmm0, %xmm1 -; CHECK-SSE-NEXT: movaps %xmm1, %xmm0 -; CHECK-SSE-NEXT: callq __truncsfhf2@PLT -; CHECK-SSE-NEXT: popq %rax -; CHECK-SSE-NEXT: retq -; -; CHECK-AVX2-LABEL: fdiv_pow_shl_cnt_fail_out_of_bound2: -; CHECK-AVX2: # %bb.0: -; CHECK-AVX2-NEXT: pushq %rax -; CHECK-AVX2-NEXT: movl %edi, %ecx -; CHECK-AVX2-NEXT: movl $1, %eax -; CHECK-AVX2-NEXT: # kill: def $cl killed $cl killed $ecx -; CHECK-AVX2-NEXT: shll %cl, %eax -; CHECK-AVX2-NEXT: movzwl %ax, %eax -; CHECK-AVX2-NEXT: vcvtsi2ss %eax, %xmm0, %xmm0 -; CHECK-AVX2-NEXT: callq __truncsfhf2@PLT -; CHECK-AVX2-NEXT: callq __extendhfsf2@PLT -; CHECK-AVX2-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-AVX2-NEXT: vdivss %xmm0, %xmm1, %xmm0 -; CHECK-AVX2-NEXT: callq __truncsfhf2@PLT -; CHECK-AVX2-NEXT: popq %rax -; CHECK-AVX2-NEXT: retq -; -; CHECK-NO-FASTFMA-LABEL: fdiv_pow_shl_cnt_fail_out_of_bound2: -; CHECK-NO-FASTFMA: # %bb.0: -; CHECK-NO-FASTFMA-NEXT: movl %edi, %ecx -; CHECK-NO-FASTFMA-NEXT: movl $1, %eax -; CHECK-NO-FASTFMA-NEXT: # kill: def $cl killed $cl killed $ecx -; CHECK-NO-FASTFMA-NEXT: shll %cl, %eax -; CHECK-NO-FASTFMA-NEXT: movzwl %ax, %eax -; CHECK-NO-FASTFMA-NEXT: vcvtsi2ss %eax, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vcvtps2ph $4, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vpmovzxwq {{.*#+}} xmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero -; CHECK-NO-FASTFMA-NEXT: vcvtph2ps %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-NO-FASTFMA-NEXT: vdivss %xmm0, %xmm1, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vcvtps2ph $4, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vmovd %xmm0, %eax -; CHECK-NO-FASTFMA-NEXT: vpinsrw $0, %eax, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: retq -; -; CHECK-FMA-LABEL: fdiv_pow_shl_cnt_fail_out_of_bound2: -; CHECK-FMA: # %bb.0: -; CHECK-FMA-NEXT: movl $1, %eax -; CHECK-FMA-NEXT: shlxl %edi, %eax, %eax -; CHECK-FMA-NEXT: movzwl %ax, %eax -; CHECK-FMA-NEXT: vcvtsi2ss %eax, %xmm0, %xmm0 -; CHECK-FMA-NEXT: vcvtps2ph $4, %xmm0, %xmm0 -; CHECK-FMA-NEXT: vpmovzxwq {{.*#+}} xmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero -; CHECK-FMA-NEXT: vcvtph2ps %xmm0, %xmm0 -; CHECK-FMA-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-FMA-NEXT: vdivss %xmm0, %xmm1, %xmm0 -; CHECK-FMA-NEXT: vcvtps2ph $4, %xmm0, %xmm0 -; CHECK-FMA-NEXT: vmovd %xmm0, %eax -; CHECK-FMA-NEXT: vpinsrw $0, %eax, %xmm0, %xmm0 -; CHECK-FMA-NEXT: retq ; VI-LABEL: fdiv_pow_shl_cnt_fail_out_of_bound2: ; VI: ; %bb.0: ; VI-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) @@ -2980,47 +1591,6 @@ define half @fdiv_pow_shl_cnt_fail_out_of_bound2(i16 %cnt) nounwind { } define double @fdiv_pow_shl_cnt32_to_dbl_okay(i32 %cnt) nounwind { -; CHECK-SSE-LABEL: fdiv_pow_shl_cnt32_to_dbl_okay: -; CHECK-SSE: # %bb.0: -; CHECK-SSE-NEXT: movl %edi, %ecx -; CHECK-SSE-NEXT: movl $1, %eax -; CHECK-SSE-NEXT: # kill: def $cl killed $cl killed $ecx -; CHECK-SSE-NEXT: shll %cl, %eax -; CHECK-SSE-NEXT: cvtsi2sd %rax, %xmm1 -; CHECK-SSE-NEXT: movsd {{.*#+}} xmm0 = mem[0],zero -; CHECK-SSE-NEXT: divsd %xmm1, %xmm0 -; CHECK-SSE-NEXT: retq -; -; CHECK-AVX2-LABEL: fdiv_pow_shl_cnt32_to_dbl_okay: -; CHECK-AVX2: # %bb.0: -; CHECK-AVX2-NEXT: movl %edi, %ecx -; CHECK-AVX2-NEXT: movl $1, %eax -; CHECK-AVX2-NEXT: # kill: def $cl killed $cl killed $ecx -; CHECK-AVX2-NEXT: shll %cl, %eax -; CHECK-AVX2-NEXT: vcvtsi2sd %rax, %xmm0, %xmm0 -; CHECK-AVX2-NEXT: vmovsd {{.*#+}} xmm1 = mem[0],zero -; CHECK-AVX2-NEXT: vdivsd %xmm0, %xmm1, %xmm0 -; CHECK-AVX2-NEXT: retq -; -; CHECK-NO-FASTFMA-LABEL: fdiv_pow_shl_cnt32_to_dbl_okay: -; CHECK-NO-FASTFMA: # %bb.0: -; CHECK-NO-FASTFMA-NEXT: movl %edi, %ecx -; CHECK-NO-FASTFMA-NEXT: movl $1, %eax -; CHECK-NO-FASTFMA-NEXT: # kill: def $cl killed $cl killed $ecx -; CHECK-NO-FASTFMA-NEXT: shll %cl, %eax -; CHECK-NO-FASTFMA-NEXT: vcvtusi2sd %eax, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vmovsd {{.*#+}} xmm1 = mem[0],zero -; CHECK-NO-FASTFMA-NEXT: vdivsd %xmm0, %xmm1, %xmm0 -; CHECK-NO-FASTFMA-NEXT: retq -; -; CHECK-FMA-LABEL: fdiv_pow_shl_cnt32_to_dbl_okay: -; CHECK-FMA: # %bb.0: -; CHECK-FMA-NEXT: movl $1, %eax -; CHECK-FMA-NEXT: shlxl %edi, %eax, %eax -; CHECK-FMA-NEXT: vcvtusi2sd %eax, %xmm0, %xmm0 -; CHECK-FMA-NEXT: vmovsd {{.*#+}} xmm1 = mem[0],zero -; CHECK-FMA-NEXT: vdivsd %xmm0, %xmm1, %xmm0 -; CHECK-FMA-NEXT: retq ; VI-LABEL: fdiv_pow_shl_cnt32_to_dbl_okay: ; VI: ; %bb.0: ; VI-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) @@ -3056,47 +1626,6 @@ define double @fdiv_pow_shl_cnt32_to_dbl_okay(i32 %cnt) nounwind { } define float @fdiv_pow_shl_cnt32_out_of_bounds2(i32 %cnt) nounwind { -; CHECK-SSE-LABEL: fdiv_pow_shl_cnt32_out_of_bounds2: -; CHECK-SSE: # %bb.0: -; CHECK-SSE-NEXT: movl %edi, %ecx -; CHECK-SSE-NEXT: movl $1, %eax -; CHECK-SSE-NEXT: # kill: def $cl killed $cl killed $ecx -; CHECK-SSE-NEXT: shll %cl, %eax -; CHECK-SSE-NEXT: cvtsi2ss %rax, %xmm1 -; CHECK-SSE-NEXT: movss {{.*#+}} xmm0 = mem[0],zero,zero,zero -; CHECK-SSE-NEXT: divss %xmm1, %xmm0 -; CHECK-SSE-NEXT: retq -; -; CHECK-AVX2-LABEL: fdiv_pow_shl_cnt32_out_of_bounds2: -; CHECK-AVX2: # %bb.0: -; CHECK-AVX2-NEXT: movl %edi, %ecx -; CHECK-AVX2-NEXT: movl $1, %eax -; CHECK-AVX2-NEXT: # kill: def $cl killed $cl killed $ecx -; CHECK-AVX2-NEXT: shll %cl, %eax -; CHECK-AVX2-NEXT: vcvtsi2ss %rax, %xmm0, %xmm0 -; CHECK-AVX2-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-AVX2-NEXT: vdivss %xmm0, %xmm1, %xmm0 -; CHECK-AVX2-NEXT: retq -; -; CHECK-NO-FASTFMA-LABEL: fdiv_pow_shl_cnt32_out_of_bounds2: -; CHECK-NO-FASTFMA: # %bb.0: -; CHECK-NO-FASTFMA-NEXT: movl %edi, %ecx -; CHECK-NO-FASTFMA-NEXT: movl $1, %eax -; CHECK-NO-FASTFMA-NEXT: # kill: def $cl killed $cl killed $ecx -; CHECK-NO-FASTFMA-NEXT: shll %cl, %eax -; CHECK-NO-FASTFMA-NEXT: vcvtusi2ss %eax, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-NO-FASTFMA-NEXT: vdivss %xmm0, %xmm1, %xmm0 -; CHECK-NO-FASTFMA-NEXT: retq -; -; CHECK-FMA-LABEL: fdiv_pow_shl_cnt32_out_of_bounds2: -; CHECK-FMA: # %bb.0: -; CHECK-FMA-NEXT: movl $1, %eax -; CHECK-FMA-NEXT: shlxl %edi, %eax, %eax -; CHECK-FMA-NEXT: vcvtusi2ss %eax, %xmm0, %xmm0 -; CHECK-FMA-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-FMA-NEXT: vdivss %xmm0, %xmm1, %xmm0 -; CHECK-FMA-NEXT: retq ; VI-LABEL: fdiv_pow_shl_cnt32_out_of_bounds2: ; VI: ; %bb.0: ; VI-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) @@ -3164,47 +1693,6 @@ define float @fdiv_pow_shl_cnt32_out_of_bounds2(i32 %cnt) nounwind { } define float @fdiv_pow_shl_cnt32_okay(i32 %cnt) nounwind { -; CHECK-SSE-LABEL: fdiv_pow_shl_cnt32_okay: -; CHECK-SSE: # %bb.0: -; CHECK-SSE-NEXT: movl %edi, %ecx -; CHECK-SSE-NEXT: movl $1, %eax -; CHECK-SSE-NEXT: # kill: def $cl killed $cl killed $ecx -; CHECK-SSE-NEXT: shll %cl, %eax -; CHECK-SSE-NEXT: cvtsi2ss %rax, %xmm1 -; CHECK-SSE-NEXT: movss {{.*#+}} xmm0 = mem[0],zero,zero,zero -; CHECK-SSE-NEXT: divss %xmm1, %xmm0 -; CHECK-SSE-NEXT: retq -; -; CHECK-AVX2-LABEL: fdiv_pow_shl_cnt32_okay: -; CHECK-AVX2: # %bb.0: -; CHECK-AVX2-NEXT: movl %edi, %ecx -; CHECK-AVX2-NEXT: movl $1, %eax -; CHECK-AVX2-NEXT: # kill: def $cl killed $cl killed $ecx -; CHECK-AVX2-NEXT: shll %cl, %eax -; CHECK-AVX2-NEXT: vcvtsi2ss %rax, %xmm0, %xmm0 -; CHECK-AVX2-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-AVX2-NEXT: vdivss %xmm0, %xmm1, %xmm0 -; CHECK-AVX2-NEXT: retq -; -; CHECK-NO-FASTFMA-LABEL: fdiv_pow_shl_cnt32_okay: -; CHECK-NO-FASTFMA: # %bb.0: -; CHECK-NO-FASTFMA-NEXT: movl %edi, %ecx -; CHECK-NO-FASTFMA-NEXT: movl $1, %eax -; CHECK-NO-FASTFMA-NEXT: # kill: def $cl killed $cl killed $ecx -; CHECK-NO-FASTFMA-NEXT: shll %cl, %eax -; CHECK-NO-FASTFMA-NEXT: vcvtusi2ss %eax, %xmm0, %xmm0 -; CHECK-NO-FASTFMA-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-NO-FASTFMA-NEXT: vdivss %xmm0, %xmm1, %xmm0 -; CHECK-NO-FASTFMA-NEXT: retq -; -; CHECK-FMA-LABEL: fdiv_pow_shl_cnt32_okay: -; CHECK-FMA: # %bb.0: -; CHECK-FMA-NEXT: movl $1, %eax -; CHECK-FMA-NEXT: shlxl %edi, %eax, %eax -; CHECK-FMA-NEXT: vcvtusi2ss %eax, %xmm0, %xmm0 -; CHECK-FMA-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; CHECK-FMA-NEXT: vdivss %xmm0, %xmm1, %xmm0 -; CHECK-FMA-NEXT: retq ; VI-LABEL: fdiv_pow_shl_cnt32_okay: ; VI: ; %bb.0: ; VI-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) From 415a82c64afa0047383caf471aa872fdb0e2658d Mon Sep 17 00:00:00 2001 From: Clement Courbet Date: Mon, 10 Jun 2024 11:15:11 +0200 Subject: [PATCH 30/36] =?UTF-8?q?[clang-tidy]=20`doesNotMutateObject`:=20H?= =?UTF-8?q?andle=20calls=20to=20member=20functions=20=E2=80=A6=20(#94362)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit …and operators that have non-const overloads. This allows `unnecessary-copy-initialization` to warn on more cases. The common case is a class with a a set of const/non-sconst overloads (e.g. std::vector::operator[]). ``` void F() { std::vector v; // ... const Expensive e = v[i]; } ``` --- .../UnnecessaryCopyInitialization.cpp | 25 +-- .../clang-tidy/utils/DeclRefExprUtils.cpp | 168 +++++++++++++++++- clang-tools-extra/docs/ReleaseNotes.rst | 4 +- .../unnecessary-copy-initialization.cpp | 29 ++- .../clang-tidy/DeclRefExprUtilsTest.cpp | 48 ++++- 5 files changed, 247 insertions(+), 27 deletions(-) diff --git a/clang-tools-extra/clang-tidy/performance/UnnecessaryCopyInitialization.cpp b/clang-tools-extra/clang-tidy/performance/UnnecessaryCopyInitialization.cpp index 9beb185cba929de..61240fa4b0eb8e3 100644 --- a/clang-tools-extra/clang-tidy/performance/UnnecessaryCopyInitialization.cpp +++ b/clang-tools-extra/clang-tidy/performance/UnnecessaryCopyInitialization.cpp @@ -75,16 +75,16 @@ void recordRemoval(const DeclStmt &Stmt, ASTContext &Context, } } -AST_MATCHER_FUNCTION_P(StatementMatcher, isConstRefReturningMethodCall, +AST_MATCHER_FUNCTION_P(StatementMatcher, + isRefReturningMethodCallWithConstOverloads, std::vector, ExcludedContainerTypes) { // Match method call expressions where the `this` argument is only used as - // const, this will be checked in `check()` part. This returned const - // reference is highly likely to outlive the local const reference of the - // variable being declared. The assumption is that the const reference being - // returned either points to a global static variable or to a member of the - // called object. + // const, this will be checked in `check()` part. This returned reference is + // highly likely to outlive the local const reference of the variable being + // declared. The assumption is that the reference being returned either points + // to a global static variable or to a member of the called object. const auto MethodDecl = - cxxMethodDecl(returns(hasCanonicalType(matchers::isReferenceToConst()))) + cxxMethodDecl(returns(hasCanonicalType(referenceType()))) .bind(MethodDeclId); const auto ReceiverExpr = ignoringParenImpCasts(declRefExpr(to(varDecl().bind(ObjectArgId)))); @@ -121,7 +121,7 @@ AST_MATCHER_FUNCTION_P(StatementMatcher, initializerReturnsReferenceToConst, declRefExpr(to(varDecl(hasLocalStorage()).bind(OldVarDeclId))); return expr( anyOf(isConstRefReturningFunctionCall(), - isConstRefReturningMethodCall(ExcludedContainerTypes), + isRefReturningMethodCallWithConstOverloads(ExcludedContainerTypes), ignoringImpCasts(OldVarDeclRef), ignoringImpCasts(unaryOperator(hasOperatorName("&"), hasUnaryOperand(OldVarDeclRef))))); @@ -259,10 +259,11 @@ void UnnecessaryCopyInitialization::registerMatchers(MatchFinder *Finder) { .bind("blockStmt"); }; - Finder->addMatcher(LocalVarCopiedFrom(anyOf(isConstRefReturningFunctionCall(), - isConstRefReturningMethodCall( - ExcludedContainerTypes))), - this); + Finder->addMatcher( + LocalVarCopiedFrom(anyOf( + isConstRefReturningFunctionCall(), + isRefReturningMethodCallWithConstOverloads(ExcludedContainerTypes))), + this); Finder->addMatcher(LocalVarCopiedFrom(declRefExpr( to(varDecl(hasLocalStorage()).bind(OldVarDeclId)))), diff --git a/clang-tools-extra/clang-tidy/utils/DeclRefExprUtils.cpp b/clang-tools-extra/clang-tidy/utils/DeclRefExprUtils.cpp index a48e45e13568134..a6062ccf42230bf 100644 --- a/clang-tools-extra/clang-tidy/utils/DeclRefExprUtils.cpp +++ b/clang-tools-extra/clang-tidy/utils/DeclRefExprUtils.cpp @@ -36,6 +36,116 @@ void extractNodesByIdTo(ArrayRef Matches, StringRef ID, Nodes.insert(Match.getNodeAs(ID)); } +// Returns true if both types refer to the same type, +// ignoring the const-qualifier. +bool isSameTypeIgnoringConst(QualType A, QualType B) { + A = A.getCanonicalType(); + B = B.getCanonicalType(); + A.addConst(); + B.addConst(); + return A == B; +} + +// Returns true if `D` and `O` have the same parameter types. +bool hasSameParameterTypes(const CXXMethodDecl &D, const CXXMethodDecl &O) { + if (D.getNumParams() != O.getNumParams()) + return false; + for (int I = 0, E = D.getNumParams(); I < E; ++I) { + if (!isSameTypeIgnoringConst(D.getParamDecl(I)->getType(), + O.getParamDecl(I)->getType())) + return false; + } + return true; +} + +// If `D` has a const-qualified overload with otherwise identical +// ref-qualifiers and parameter types, returns that overload. +const CXXMethodDecl *findConstOverload(const CXXMethodDecl &D) { + assert(!D.isConst()); + + DeclContext::lookup_result LookupResult = + D.getParent()->lookup(D.getNameInfo().getName()); + if (LookupResult.isSingleResult()) { + // No overload. + return nullptr; + } + for (const Decl *Overload : LookupResult) { + const auto *O = dyn_cast(Overload); + if (O && !O->isDeleted() && O->isConst() && + O->getRefQualifier() == D.getRefQualifier() && + hasSameParameterTypes(D, *O)) + return O; + } + return nullptr; +} + +// Returns true if both types are pointers or reference to the same type, +// ignoring the const-qualifier. +bool pointsToSameTypeIgnoringConst(QualType A, QualType B) { + assert(A->isPointerType() || A->isReferenceType()); + assert(B->isPointerType() || B->isReferenceType()); + return isSameTypeIgnoringConst(A->getPointeeType(), B->getPointeeType()); +} + +// Return true if non-const member function `M` likely does not mutate `*this`. +// +// Note that if the member call selects a method/operator `f` that +// is not const-qualified, then we also consider that the object is +// not mutated if: +// - (A) there is a const-qualified overload `cf` of `f` that has +// the +// same ref-qualifiers; +// - (B) * `f` returns a value, or +// * if `f` returns a `T&`, `cf` returns a `const T&` (up to +// possible aliases such as `reference` and +// `const_reference`), or +// * if `f` returns a `T*`, `cf` returns a `const T*` (up to +// possible aliases). +// - (C) the result of the call is not mutated. +// +// The assumption that `cf` has the same semantics as `f`. +// For example: +// - In `std::vector v; const T t = v[...];`, we consider that +// expression `v[...]` does not mutate `v` as +// `T& std::vector::operator[]` has a const overload +// `const T& std::vector::operator[] const`, and the +// result expression of type `T&` is only used as a `const T&`; +// - In `std::map m; V v = m.at(...);`, we consider +// `m.at(...)` to be an immutable access for the same reason. +// However: +// - In `std::map m; const V v = m[...];`, We consider that +// `m[...]` mutates `m` as `V& std::map::operator[]` does +// not have a const overload. +// - In `std::vector v; T& t = v[...];`, we consider that +// expression `v[...]` mutates `v` as the result is kept as a +// mutable reference. +// +// This function checks (A) ad (B), but the caller should make sure that the +// object is not mutated through the return value. +bool isLikelyShallowConst(const CXXMethodDecl &M) { + assert(!M.isConst()); + // The method can mutate our variable. + + // (A) + const CXXMethodDecl *ConstOverload = findConstOverload(M); + if (ConstOverload == nullptr) { + return false; + } + + // (B) + const QualType CallTy = M.getReturnType().getCanonicalType(); + const QualType OverloadTy = ConstOverload->getReturnType().getCanonicalType(); + if (CallTy->isReferenceType()) { + return OverloadTy->isReferenceType() && + pointsToSameTypeIgnoringConst(CallTy, OverloadTy); + } + if (CallTy->isPointerType()) { + return OverloadTy->isPointerType() && + pointsToSameTypeIgnoringConst(CallTy, OverloadTy); + } + return isSameTypeIgnoringConst(CallTy, OverloadTy); +} + // A matcher that matches DeclRefExprs that are used in ways such that the // underlying declaration is not modified. // If the declaration is of pointer type, `Indirections` specifies the level @@ -54,16 +164,15 @@ void extractNodesByIdTo(ArrayRef Matches, StringRef ID, // matches (A). // AST_MATCHER_P(DeclRefExpr, doesNotMutateObject, int, Indirections) { - // We walk up the parents of the DeclRefExpr recursively until we end up on a - // parent that cannot modify the underlying object. There are a few kinds of - // expressions: - // - Those that cannot be used to mutate the underlying object. We can stop + // We walk up the parents of the DeclRefExpr recursively. There are a few + // kinds of expressions: + // - Those that cannot be used to mutate the underlying variable. We can stop // recursion there. - // - Those that can be used to mutate the underlying object in analyzable + // - Those that can be used to mutate the underlying variable in analyzable // ways (such as taking the address or accessing a subobject). We have to // examine the parents. // - Those that we don't know how to analyze. In that case we stop there and - // we assume that they can mutate the underlying expression. + // we assume that they can modify the expression. struct StackEntry { StackEntry(const Expr *E, int Indirections) @@ -90,7 +199,7 @@ AST_MATCHER_P(DeclRefExpr, doesNotMutateObject, int, Indirections) { assert(Ty->isPointerType()); Ty = Ty->getPointeeType().getCanonicalType(); } - if (Ty.isConstQualified()) + if (Ty->isVoidType() || Ty.isConstQualified()) continue; // Otherwise we have to look at the parents to see how the expression is @@ -159,11 +268,56 @@ AST_MATCHER_P(DeclRefExpr, doesNotMutateObject, int, Indirections) { // The method call cannot mutate our variable. continue; } + if (isLikelyShallowConst(*Method)) { + // We still have to check that the object is not modified through + // the method's return value (C). + const auto MemberParents = Ctx.getParents(*Member); + assert(MemberParents.size() == 1); + const auto *Call = MemberParents[0].get(); + // If `o` is an object of class type and `f` is a member function, + // then `o.f` has to be used as part of a call expression. + assert(Call != nullptr && "member function has to be called"); + Stack.emplace_back( + Call, + Method->getReturnType().getCanonicalType()->isPointerType() + ? 1 + : 0); + continue; + } return false; } Stack.emplace_back(Member, 0); continue; } + if (const auto *const OpCall = dyn_cast(P)) { + // Operator calls have function call syntax. The `*this` parameter + // is the first parameter. + if (OpCall->getNumArgs() == 0 || OpCall->getArg(0) != Entry.E) { + return false; + } + const auto *const Method = + dyn_cast(OpCall->getDirectCallee()); + + if (Method == nullptr) { + // This is not a member operator. Typically, a friend operator. These + // are handled like function calls. + return false; + } + + if (Method->isConst() || Method->isStatic()) { + continue; + } + if (isLikelyShallowConst(*Method)) { + // We still have to check that the object is not modified through + // the operator's return value (C). + Stack.emplace_back( + OpCall, + Method->getReturnType().getCanonicalType()->isPointerType() ? 1 + : 0); + continue; + } + return false; + } if (const auto *const Op = dyn_cast(P)) { switch (Op->getOpcode()) { diff --git a/clang-tools-extra/docs/ReleaseNotes.rst b/clang-tools-extra/docs/ReleaseNotes.rst index 50edc10654dab32..0c0c10605a83070 100644 --- a/clang-tools-extra/docs/ReleaseNotes.rst +++ b/clang-tools-extra/docs/ReleaseNotes.rst @@ -376,8 +376,10 @@ Changes in existing checks - Improved :doc:`performance-unnecessary-copy-initialization ` check by detecting more cases of constant access. In particular, pointers can be - analyzed, se the check now handles the common patterns + analyzed, so the check now handles the common patterns `const auto e = (*vector_ptr)[i]` and `const auto e = vector_ptr->at(i);`. + Calls to mutable function where there exists a `const` overload are also + handled. - Improved :doc:`readability-avoid-return-with-void-value ` check by adding diff --git a/clang-tools-extra/test/clang-tidy/checkers/performance/unnecessary-copy-initialization.cpp b/clang-tools-extra/test/clang-tidy/checkers/performance/unnecessary-copy-initialization.cpp index 92625cc1332e286..f259552dc8f1d89 100644 --- a/clang-tools-extra/test/clang-tidy/checkers/performance/unnecessary-copy-initialization.cpp +++ b/clang-tools-extra/test/clang-tidy/checkers/performance/unnecessary-copy-initialization.cpp @@ -32,6 +32,9 @@ struct ExpensiveToCopyType { template struct Container { + using reference = T&; + using const_reference = const T&; + bool empty() const; const T& operator[](int) const; const T& operator[](int); @@ -42,8 +45,8 @@ struct Container { void nonConstMethod(); bool constMethod() const; - const T& at(int) const; - T& at(int); + reference at(int) const; + const_reference at(int); }; @@ -207,6 +210,28 @@ void PositiveOperatorCallConstValueParam(const Container C) VarCopyConstructed.constMethod(); } +void PositiveOperatorValueParam(Container C) { + const auto AutoAssigned = C[42]; + // CHECK-MESSAGES: [[@LINE-1]]:14: warning: the const qualified variable 'AutoAssigned' + // CHECK-FIXES: const auto& AutoAssigned = C[42]; + AutoAssigned.constMethod(); + + const auto AutoCopyConstructed(C[42]); + // CHECK-MESSAGES: [[@LINE-1]]:14: warning: the const qualified variable 'AutoCopyConstructed' + // CHECK-FIXES: const auto& AutoCopyConstructed(C[42]); + AutoCopyConstructed.constMethod(); + + const ExpensiveToCopyType VarAssigned = C.at(42); + // CHECK-MESSAGES: [[@LINE-1]]:29: warning: the const qualified variable 'VarAssigned' + // CHECK-FIXES: const ExpensiveToCopyType& VarAssigned = C.at(42); + VarAssigned.constMethod(); + + const ExpensiveToCopyType VarCopyConstructed(C.at(42)); + // CHECK-MESSAGES: [[@LINE-1]]:29: warning: the const qualified variable 'VarCopyConstructed' + // CHECK-FIXES: const ExpensiveToCopyType& VarCopyConstructed(C.at(42)); + VarCopyConstructed.constMethod(); +} + void PositiveOperatorCallConstValueParamAlias(const ExpensiveToCopyContainerAlias C) { const auto AutoAssigned = C[42]; // CHECK-MESSAGES: [[@LINE-1]]:14: warning: the const qualified variable 'AutoAssigned' diff --git a/clang-tools-extra/unittests/clang-tidy/DeclRefExprUtilsTest.cpp b/clang-tools-extra/unittests/clang-tidy/DeclRefExprUtilsTest.cpp index 3d9f51e2e17b094..064e04c932de83e 100644 --- a/clang-tools-extra/unittests/clang-tidy/DeclRefExprUtilsTest.cpp +++ b/clang-tools-extra/unittests/clang-tidy/DeclRefExprUtilsTest.cpp @@ -46,6 +46,7 @@ template void RunTest(StringRef Snippet) { StringRef CommonCode = R"( struct ConstTag{}; struct NonConstTag{}; + struct Tag1{}; struct S { void constMethod() const; @@ -59,6 +60,13 @@ template void RunTest(StringRef Snippet) { void operator[](int); void operator[](int) const; + int& at(int); + const int& at(int) const; + const int& at(Tag1); + + int& weird_overload(); + const double& weird_overload() const; + bool operator==(const S&) const; int int_member; @@ -161,9 +169,11 @@ TEST(ConstReferenceDeclRefExprsTest, ConstRefVar) { useIntConstRef(/*const*/target.int_member); useIntPtr(/*const*/target.ptr_member); useIntConstPtr(&/*const*/target.int_member); + (void)/*const*/target.at(3); const S& const_target_ref = /*const*/target; const S* const_target_ptr = &/*const*/target; + (void)/*const*/target.at(3); } )"); } @@ -187,7 +197,7 @@ TEST(ConstReferenceDeclRefExprsTest, ValueVar) { /*const*/target.staticMethod(); target.nonConstMethod(); /*const*/target(ConstTag{}); - target[42]; + /*const*/target[42]; /*const*/target(ConstTag{}); target(NonConstTag{}); useRef(target); @@ -211,6 +221,14 @@ TEST(ConstReferenceDeclRefExprsTest, ValueVar) { const S& const_target_ref = /*const*/target; const S* const_target_ptr = &/*const*/target; S* target_ptr = ⌖ + + (void)/*const*/target.at(3); + ++target.at(3); + const int civ = /*const*/target.at(3); + const int& cir = /*const*/target.at(3); + int& ir = target.at(3); + target.at(Tag1{}); + target.weird_overload(); } )"); } @@ -227,7 +245,7 @@ TEST(ConstReferenceDeclRefExprsTest, RefVar) { /*const*/target.staticMethod(); target.nonConstMethod(); /*const*/target(ConstTag{}); - target[42]; + /*const*/target[42]; useConstRef((/*const*/target)); (/*const*/target).constMethod(); (void)(/*const*/target == /*const*/target); @@ -249,6 +267,14 @@ TEST(ConstReferenceDeclRefExprsTest, RefVar) { const S& const_target_ref = /*const*/target; const S* const_target_ptr = &/*const*/target; S* target_ptr = ⌖ + + (void)/*const*/target.at(3); + ++target.at(3); + const int civ = /*const*/target.at(3); + const int& cir = /*const*/target.at(3); + int& ir = target.at(3); + target.at(Tag1{}); + target.weird_overload(); } )"); } @@ -266,8 +292,8 @@ TEST(ConstReferenceDeclRefExprsTest, PtrVar) { /*const*/target->staticMethod(); target->nonConstMethod(); (*/*const*/target)(ConstTag{}); - (*target)[42]; - target->operator[](42); + (*/*const*/target)[42]; + /*const*/target->operator[](42); useConstRef((*/*const*/target)); (/*const*/target)->constMethod(); (void)(*/*const*/target == */*const*/target); @@ -284,7 +310,15 @@ TEST(ConstReferenceDeclRefExprsTest, PtrVar) { const S& const_target_ref = */*const*/target; const S* const_target_ptr = /*const*/target; - S* target_ptr = target; // FIXME: we could chect const usage of `target_ptr`. + S* target_ptr = target; // FIXME: we could chect const usage of `target_ptr` + + (void)/*const*/target->at(3); + ++target->at(3); + const int civ = /*const*/target->at(3); + const int& cir = /*const*/target->at(3); + int& ir = target->at(3); + target->at(Tag1{}); + target->weird_overload(); } )"); } @@ -319,6 +353,10 @@ TEST(ConstReferenceDeclRefExprsTest, ConstPtrVar) { const S& const_target_ref = */*const*/target; const S* const_target_ptr = /*const*/target; + + (void)/*const*/target->at(3); + const int civ = /*const*/target->at(3); + const int& cir = /*const*/target->at(3); } )"); } From 317ed770e3ad4ac663ad4d95c644e83347b78fa4 Mon Sep 17 00:00:00 2001 From: Jay Foad Date: Mon, 10 Jun 2024 10:24:10 +0100 Subject: [PATCH 31/36] [AMDGPU] Fix typos in GCN-PROMOTE check prefix --- llvm/test/CodeGen/AMDGPU/vector-alloca-bitcast.ll | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/llvm/test/CodeGen/AMDGPU/vector-alloca-bitcast.ll b/llvm/test/CodeGen/AMDGPU/vector-alloca-bitcast.ll index 769db215cf0349e..85c34e036e1fd99 100644 --- a/llvm/test/CodeGen/AMDGPU/vector-alloca-bitcast.ll +++ b/llvm/test/CodeGen/AMDGPU/vector-alloca-bitcast.ll @@ -14,12 +14,12 @@ target datalayout = "A5" ; GCN-ALLOCA-COUNT-4: buffer_store_dword ; GCN-ALLOCA: buffer_load_dword -; GCN_PROMOTE: s_cmp_lg_u32 s{{[0-9]+}}, 2 ; GCN-PROMOTE: s_cmp_eq_u32 s{{[0-9]+}}, 1 ; GCN-PROMOTE: s_cselect_b64 [[CC1:[^,]+]], -1, 0 +; GCN-PROMOTE: s_cmp_lg_u32 s{{[0-9]+}}, 2 ; GCN-PROMOTE: v_cndmask_b32_e{{32|64}} [[IND1:v[0-9]+]], 0, 1, [[CC1]] ; GCN-PROMOTE: s_cselect_b64 vcc, -1, 0 -; GCN_PROMOTE: s_cmp_lg_u32 s{{[0-9]+}}, 3 +; GCN-PROMOTE: s_cmp_lg_u32 s{{[0-9]+}}, 3 ; GCN-PROMOTE: v_cndmask_b32_e{{32|64}} [[IND2:v[0-9]+]], 2, [[IND1]], vcc ; GCN-PROMOTE: s_cselect_b64 vcc, -1, 0 ; GCN-PROMOTE: v_cndmask_b32_e{{32|64}} [[IND3:v[0-9]+]], 3, [[IND2]], vcc @@ -294,10 +294,10 @@ entry: ; GCN-PROMOTE: s_cmp_eq_u32 s{{[0-9]+}}, 1 ; GCN-PROMOTE: s_cselect_b64 [[CC1:[^,]+]], -1, 0 -; GCN_PROMOTE: s_cmp_lg_u32 s{{[0-9]+}}, 2 +; GCN-PROMOTE: s_cmp_lg_u32 s{{[0-9]+}}, 2 ; GCN-PROMOTE: v_cndmask_b32_e{{32|64}} [[IND1:v[0-9]+]], 0, 1, [[CC1]] ; GCN-PROMOTE: s_cselect_b64 vcc, -1, 0 -; GCN_PROMOTE: s_cmp_lg_u32 s{{[0-9]+}}, 3 +; GCN-PROMOTE: s_cmp_lg_u32 s{{[0-9]+}}, 3 ; GCN-PROMOTE: v_cndmask_b32_e{{32|64}} [[IND2:v[0-9]+]], 2, [[IND1]], vcc ; GCN-PROMOTE: s_cselect_b64 vcc, -1, 0 ; GCN-PROMOTE: v_cndmask_b32_e{{32|64}} [[IND3:v[0-9]+]], 3, [[IND2]], vcc From 81469a2ee804778185e177314314f637bcb026c9 Mon Sep 17 00:00:00 2001 From: jeanPerier Date: Mon, 10 Jun 2024 11:40:26 +0200 Subject: [PATCH 32/36] [flang] use hlfir base when translating assumed-rank entity to fir::ExtendedValue (#94822) The hlfir::Entity to fir::ExtendedValue conversion usually uses the "fir base" output of hlfir.declare (which is the same as the input) to avoid introducing temporary descriptors for the sole purpose of introducing updating lower bound information. This is possible because local lower bounds, if any, are tracked in a vector inside the fir::ExtendedValue. With assumed-ranks, the lower bounds cannot be tracked inside the fir::ExtendedValue vector (their numbers is unknown at compile time). Hence, the fir.box/fir.class used in fir::ExtendedValue in lowering must always contain accurate local lower bound information. --- flang/lib/Optimizer/Builder/HLFIRTools.cpp | 13 ++++-- .../Lower/HLFIR/assumed-rank-inquiries-2.f90 | 12 ++--- .../Lower/HLFIR/assumed-rank-inquiries.f90 | 46 +++++++++---------- .../HLFIR/convert-variable-assumed-rank.f90 | 4 +- flang/test/Lower/HLFIR/select-rank.f90 | 30 ++++++------ 5 files changed, 55 insertions(+), 50 deletions(-) diff --git a/flang/lib/Optimizer/Builder/HLFIRTools.cpp b/flang/lib/Optimizer/Builder/HLFIRTools.cpp index 31e2c4082b32cae..8d0ae2f195178c8 100644 --- a/flang/lib/Optimizer/Builder/HLFIRTools.cpp +++ b/flang/lib/Optimizer/Builder/HLFIRTools.cpp @@ -882,10 +882,15 @@ static fir::ExtendedValue translateVariableToExtendedValue( mlir::Location loc, fir::FirOpBuilder &builder, hlfir::Entity variable, bool forceHlfirBase = false, bool contiguousHint = false) { assert(variable.isVariable() && "must be a variable"); - /// When going towards FIR, use the original base value to avoid - /// introducing descriptors at runtime when they are not required. - mlir::Value base = - forceHlfirBase ? variable.getBase() : variable.getFirBase(); + // When going towards FIR, use the original base value to avoid + // introducing descriptors at runtime when they are not required. + // This is not done for assumed-rank since the fir::ExtendedValue cannot + // held the related lower bounds in an vector. The lower bounds of the + // descriptor must always be used instead. + + mlir::Value base = (forceHlfirBase || variable.isAssumedRank()) + ? variable.getBase() + : variable.getFirBase(); if (variable.isMutableBox()) return fir::MutableBoxValue(base, getExplicitTypeParams(variable), fir::MutableProperties{}); diff --git a/flang/test/Lower/HLFIR/assumed-rank-inquiries-2.f90 b/flang/test/Lower/HLFIR/assumed-rank-inquiries-2.f90 index 353e944a8ac18c3..6c8f5ba48d1bf90 100644 --- a/flang/test/Lower/HLFIR/assumed-rank-inquiries-2.f90 +++ b/flang/test/Lower/HLFIR/assumed-rank-inquiries-2.f90 @@ -29,7 +29,7 @@ subroutine test_size_4(x) ! CHECK-SAME: %[[VAL_0:.*]]: !fir.box> {fir.bindc_name = "x"}) { ! CHECK: %[[VAL_1:.*]] = fir.dummy_scope : !fir.dscope ! CHECK: %[[VAL_2:.*]]:2 = hlfir.declare %[[VAL_0]] dummy_scope %[[VAL_1]] {uniq_name = "_QFtest_size_1Ex"} : (!fir.box>, !fir.dscope) -> (!fir.box>, !fir.box>) -! CHECK: %[[VAL_5:.*]] = fir.convert %[[VAL_2]]#1 : (!fir.box>) -> !fir.box +! CHECK: %[[VAL_5:.*]] = fir.convert %[[VAL_2]]#0 : (!fir.box>) -> !fir.box ! CHECK: %[[VAL_7:.*]] = fir.call @_FortranASize(%[[VAL_5]] ! CHECK: %[[VAL_8:.*]] = fir.convert %[[VAL_7]] : (i64) -> i32 ! CHECK: %[[VAL_9:.*]]:3 = hlfir.associate %[[VAL_8]] {adapt.valuebyref} : (i32) -> (!fir.ref, !fir.ref, i1) @@ -49,13 +49,13 @@ subroutine test_size_4(x) ! CHECK: %[[VAL_6:.*]] = arith.constant 0 : i64 ! CHECK: %[[VAL_7:.*]] = arith.cmpi eq, %[[VAL_5]], %[[VAL_6]] : i64 ! CHECK: %[[VAL_8:.*]] = fir.if %[[VAL_7]] -> (i32) { -! CHECK: %[[VAL_11:.*]] = fir.convert %[[VAL_3]]#1 : (!fir.box>) -> !fir.box +! CHECK: %[[VAL_11:.*]] = fir.convert %[[VAL_3]]#0 : (!fir.box>) -> !fir.box ! CHECK: %[[VAL_13:.*]] = fir.call @_FortranASize(%[[VAL_11]] ! CHECK: %[[VAL_14:.*]] = fir.convert %[[VAL_13]] : (i64) -> i32 ! CHECK: fir.result %[[VAL_14]] : i32 ! CHECK: } else { ! CHECK: %[[VAL_15:.*]] = fir.load %[[VAL_1]] : !fir.ref -! CHECK: %[[VAL_18:.*]] = fir.convert %[[VAL_3]]#1 : (!fir.box>) -> !fir.box +! CHECK: %[[VAL_18:.*]] = fir.convert %[[VAL_3]]#0 : (!fir.box>) -> !fir.box ! CHECK: %[[VAL_20:.*]] = fir.call @_FortranASizeDim(%[[VAL_18]] ! CHECK: %[[VAL_21:.*]] = fir.convert %[[VAL_20]] : (i64) -> i32 ! CHECK: fir.result %[[VAL_21]] : i32 @@ -76,13 +76,13 @@ subroutine test_size_4(x) ! CHECK: %[[VAL_6:.*]] = arith.constant 0 : i64 ! CHECK: %[[VAL_7:.*]] = arith.cmpi eq, %[[VAL_5]], %[[VAL_6]] : i64 ! CHECK: %[[VAL_8:.*]] = fir.if %[[VAL_7]] -> (i32) { -! CHECK: %[[VAL_11:.*]] = fir.convert %[[VAL_4]]#1 : (!fir.box>) -> !fir.box +! CHECK: %[[VAL_11:.*]] = fir.convert %[[VAL_4]]#0 : (!fir.box>) -> !fir.box ! CHECK: %[[VAL_13:.*]] = fir.call @_FortranASize(%[[VAL_11]], ! CHECK: %[[VAL_14:.*]] = fir.convert %[[VAL_13]] : (i64) -> i32 ! CHECK: fir.result %[[VAL_14]] : i32 ! CHECK: } else { ! CHECK: %[[VAL_15:.*]] = fir.load %[[VAL_3]]#1 : !fir.ref -! CHECK: %[[VAL_18:.*]] = fir.convert %[[VAL_4]]#1 : (!fir.box>) -> !fir.box +! CHECK: %[[VAL_18:.*]] = fir.convert %[[VAL_4]]#0 : (!fir.box>) -> !fir.box ! CHECK: %[[VAL_20:.*]] = fir.call @_FortranASizeDim(%[[VAL_18]] ! CHECK: %[[VAL_21:.*]] = fir.convert %[[VAL_20]] : (i64) -> i32 ! CHECK: fir.result %[[VAL_21]] : i32 @@ -97,7 +97,7 @@ subroutine test_size_4(x) ! CHECK-SAME: %[[VAL_0:.*]]: !fir.ref>>> {fir.bindc_name = "x"}) { ! CHECK: %[[VAL_1:.*]] = fir.dummy_scope : !fir.dscope ! CHECK: %[[VAL_2:.*]]:2 = hlfir.declare %[[VAL_0]] dummy_scope %[[VAL_1]] {fortran_attrs = #fir.var_attrs, uniq_name = "_QFtest_size_4Ex"} : (!fir.ref>>>, !fir.dscope) -> (!fir.ref>>>, !fir.ref>>>) -! CHECK: %[[VAL_3:.*]] = fir.load %[[VAL_2]]#1 : !fir.ref>>> +! CHECK: %[[VAL_3:.*]] = fir.load %[[VAL_2]]#0 : !fir.ref>>> ! CHECK: %[[VAL_6:.*]] = fir.convert %[[VAL_3]] : (!fir.box>>) -> !fir.box ! CHECK: %[[VAL_8:.*]] = fir.call @_FortranASize(%[[VAL_6]] ! CHECK: %[[VAL_9:.*]] = fir.convert %[[VAL_8]] : (i64) -> i32 diff --git a/flang/test/Lower/HLFIR/assumed-rank-inquiries.f90 b/flang/test/Lower/HLFIR/assumed-rank-inquiries.f90 index e8610aa7d0e78cf..a1d150a21d1497f 100644 --- a/flang/test/Lower/HLFIR/assumed-rank-inquiries.f90 +++ b/flang/test/Lower/HLFIR/assumed-rank-inquiries.f90 @@ -99,7 +99,7 @@ subroutine c_loc_2(x) ! CHECK-SAME: %[[VAL_0:.*]]: !fir.ref>>> {fir.bindc_name = "x"}) { ! CHECK: %[[VAL_1:.*]] = fir.dummy_scope : !fir.dscope ! CHECK: %[[VAL_2:.*]]:2 = hlfir.declare %[[VAL_0]] dummy_scope %[[VAL_1]] {fortran_attrs = #fir.var_attrs, uniq_name = "_QFtest_allocatedEx"} : (!fir.ref>>>, !fir.dscope) -> (!fir.ref>>>, !fir.ref>>>) -! CHECK: %[[VAL_3:.*]] = fir.load %[[VAL_2]]#1 : !fir.ref>>> +! CHECK: %[[VAL_3:.*]] = fir.load %[[VAL_2]]#0 : !fir.ref>>> ! CHECK: %[[VAL_4:.*]] = fir.box_addr %[[VAL_3]] : (!fir.box>>) -> !fir.heap> ! CHECK: %[[VAL_5:.*]] = fir.convert %[[VAL_4]] : (!fir.heap>) -> i64 ! CHECK: %[[VAL_6:.*]] = arith.constant 0 : i64 @@ -115,7 +115,7 @@ subroutine c_loc_2(x) ! CHECK-SAME: %[[VAL_0:.*]]: !fir.ref>>> {fir.bindc_name = "x"}) { ! CHECK: %[[VAL_1:.*]] = fir.dummy_scope : !fir.dscope ! CHECK: %[[VAL_2:.*]]:2 = hlfir.declare %[[VAL_0]] dummy_scope %[[VAL_1]] {fortran_attrs = #fir.var_attrs, uniq_name = "_QFtest_associated_1Ex"} : (!fir.ref>>>, !fir.dscope) -> (!fir.ref>>>, !fir.ref>>>) -! CHECK: %[[VAL_3:.*]] = fir.load %[[VAL_2]]#1 : !fir.ref>>> +! CHECK: %[[VAL_3:.*]] = fir.load %[[VAL_2]]#0 : !fir.ref>>> ! CHECK: %[[VAL_4:.*]] = fir.box_addr %[[VAL_3]] : (!fir.box>>) -> !fir.ptr> ! CHECK: %[[VAL_5:.*]] = fir.convert %[[VAL_4]] : (!fir.ptr>) -> i64 ! CHECK: %[[VAL_6:.*]] = arith.constant 0 : i64 @@ -133,7 +133,7 @@ subroutine c_loc_2(x) ! CHECK: %[[VAL_2:.*]] = fir.dummy_scope : !fir.dscope ! CHECK: %[[VAL_3:.*]]:2 = hlfir.declare %[[VAL_0]] dummy_scope %[[VAL_2]] {fortran_attrs = #fir.var_attrs, uniq_name = "_QFtest_associated_2Ex"} : (!fir.ref>>>, !fir.dscope) -> (!fir.ref>>>, !fir.ref>>>) ! CHECK: %[[VAL_4:.*]]:2 = hlfir.declare %[[VAL_1]] dummy_scope %[[VAL_2]] {fortran_attrs = #fir.var_attrs, uniq_name = "_QFtest_associated_2Ey"} : (!fir.box>, !fir.dscope) -> (!fir.box>, !fir.box>) -! CHECK: %[[VAL_5:.*]] = fir.load %[[VAL_3]]#1 : !fir.ref>>> +! CHECK: %[[VAL_5:.*]] = fir.load %[[VAL_3]]#0 : !fir.ref>>> ! CHECK: %[[VAL_6:.*]] = fir.convert %[[VAL_5]] : (!fir.box>>) -> !fir.box ! CHECK: %[[VAL_7:.*]] = fir.convert %[[VAL_4]]#1 : (!fir.box>) -> !fir.box ! CHECK: %[[VAL_8:.*]] = fir.call @_FortranAPointerIsAssociatedWith(%[[VAL_6]], %[[VAL_7]]) fastmath : (!fir.box, !fir.box) -> i1 @@ -150,8 +150,8 @@ subroutine c_loc_2(x) ! CHECK: %[[VAL_2:.*]] = fir.dummy_scope : !fir.dscope ! CHECK: %[[VAL_3:.*]]:2 = hlfir.declare %[[VAL_0]] dummy_scope %[[VAL_2]] {fortran_attrs = #fir.var_attrs, uniq_name = "_QFtest_associated_3Ex"} : (!fir.ref>>>, !fir.dscope) -> (!fir.ref>>>, !fir.ref>>>) ! CHECK: %[[VAL_4:.*]]:2 = hlfir.declare %[[VAL_1]] dummy_scope %[[VAL_2]] {fortran_attrs = #fir.var_attrs, uniq_name = "_QFtest_associated_3Ey"} : (!fir.ref>>>, !fir.dscope) -> (!fir.ref>>>, !fir.ref>>>) -! CHECK: %[[VAL_5:.*]] = fir.load %[[VAL_4]]#1 : !fir.ref>>> -! CHECK: %[[VAL_6:.*]] = fir.load %[[VAL_3]]#1 : !fir.ref>>> +! CHECK: %[[VAL_5:.*]] = fir.load %[[VAL_4]]#0 : !fir.ref>>> +! CHECK: %[[VAL_6:.*]] = fir.load %[[VAL_3]]#0 : !fir.ref>>> ! CHECK: %[[VAL_7:.*]] = fir.convert %[[VAL_6]] : (!fir.box>>) -> !fir.box ! CHECK: %[[VAL_8:.*]] = fir.convert %[[VAL_5]] : (!fir.box>>) -> !fir.box ! CHECK: %[[VAL_9:.*]] = fir.call @_FortranAPointerIsAssociatedWith(%[[VAL_7]], %[[VAL_8]]) fastmath : (!fir.box, !fir.box) -> i1 @@ -166,7 +166,7 @@ subroutine c_loc_2(x) ! CHECK-SAME: %[[VAL_0:.*]]: !fir.box>> {fir.bindc_name = "x"}) { ! CHECK: %[[VAL_1:.*]] = fir.dummy_scope : !fir.dscope ! CHECK: %[[VAL_2:.*]]:2 = hlfir.declare %[[VAL_0]] dummy_scope %[[VAL_1]] {uniq_name = "_QFtest_len_1Ex"} : (!fir.box>>, !fir.dscope) -> (!fir.box>>, !fir.box>>) -! CHECK: %[[VAL_3:.*]] = fir.box_elesize %[[VAL_2]]#1 : (!fir.box>>) -> index +! CHECK: %[[VAL_3:.*]] = fir.box_elesize %[[VAL_2]]#0 : (!fir.box>>) -> index ! CHECK: %[[VAL_4:.*]] = fir.convert %[[VAL_3]] : (index) -> i32 ! CHECK: %[[VAL_5:.*]]:3 = hlfir.associate %[[VAL_4]] {adapt.valuebyref} : (i32) -> (!fir.ref, !fir.ref, i1) ! CHECK: fir.call @_QPtakes_integer(%[[VAL_5]]#1) fastmath : (!fir.ref) -> () @@ -191,7 +191,7 @@ subroutine c_loc_2(x) ! CHECK-SAME: %[[VAL_0:.*]]: !fir.class> {fir.bindc_name = "x"}) { ! CHECK: %[[VAL_1:.*]] = fir.dummy_scope : !fir.dscope ! CHECK: %[[VAL_2:.*]]:2 = hlfir.declare %[[VAL_0]] dummy_scope %[[VAL_1]] {uniq_name = "_QFtest_storage_size_1Ex"} : (!fir.class>, !fir.dscope) -> (!fir.class>, !fir.class>) -! CHECK: %[[VAL_3:.*]] = fir.box_elesize %[[VAL_2]]#1 : (!fir.class>) -> i32 +! CHECK: %[[VAL_3:.*]] = fir.box_elesize %[[VAL_2]]#0 : (!fir.class>) -> i32 ! CHECK: %[[VAL_4:.*]] = arith.constant 8 : i32 ! CHECK: %[[VAL_5:.*]] = arith.muli %[[VAL_3]], %[[VAL_4]] : i32 ! CHECK: %[[VAL_6:.*]]:3 = hlfir.associate %[[VAL_5]] {adapt.valuebyref} : (i32) -> (!fir.ref, !fir.ref, i1) @@ -204,7 +204,7 @@ subroutine c_loc_2(x) ! CHECK-SAME: %[[VAL_0:.*]]: !fir.ref>>> {fir.bindc_name = "x"}) { ! CHECK: %[[VAL_1:.*]] = fir.dummy_scope : !fir.dscope ! CHECK: %[[VAL_2:.*]]:2 = hlfir.declare %[[VAL_0]] dummy_scope %[[VAL_1]] {fortran_attrs = #fir.var_attrs, uniq_name = "_QFtest_storage_size_2Ex"} : (!fir.ref>>>, !fir.dscope) -> (!fir.ref>>>, !fir.ref>>>) -! CHECK: %[[VAL_3:.*]] = fir.load %[[VAL_2]]#1 : !fir.ref>>> +! CHECK: %[[VAL_3:.*]] = fir.load %[[VAL_2]]#0 : !fir.ref>>> ! CHECK: %[[VAL_4:.*]] = fir.box_addr %[[VAL_3]] : (!fir.class>>) -> !fir.ptr> ! CHECK: %[[VAL_5:.*]] = fir.convert %[[VAL_4]] : (!fir.ptr>) -> i64 ! CHECK: %[[VAL_6:.*]] = arith.constant 0 : i64 @@ -212,7 +212,7 @@ subroutine c_loc_2(x) ! CHECK: fir.if %[[VAL_7]] { ! CHECK: %[[VAL_13:.*]] = fir.call @_FortranAReportFatalUserError ! CHECK: } -! CHECK: %[[VAL_14:.*]] = fir.load %[[VAL_2]]#1 : !fir.ref>>> +! CHECK: %[[VAL_14:.*]] = fir.load %[[VAL_2]]#0 : !fir.ref>>> ! CHECK: %[[VAL_15:.*]] = fir.box_elesize %[[VAL_14]] : (!fir.class>>) -> i32 ! CHECK: %[[VAL_16:.*]] = arith.constant 8 : i32 ! CHECK: %[[VAL_17:.*]] = arith.muli %[[VAL_15]], %[[VAL_16]] : i32 @@ -226,7 +226,7 @@ subroutine c_loc_2(x) ! CHECK-SAME: %[[VAL_0:.*]]: !fir.class> {fir.bindc_name = "x", fir.optional}) { ! CHECK: %[[VAL_1:.*]] = fir.dummy_scope : !fir.dscope ! CHECK: %[[VAL_2:.*]]:2 = hlfir.declare %[[VAL_0]] dummy_scope %[[VAL_1]] {fortran_attrs = #fir.var_attrs, uniq_name = "_QFtest_present_1Ex"} : (!fir.class>, !fir.dscope) -> (!fir.class>, !fir.class>) -! CHECK: %[[VAL_3:.*]] = fir.is_present %[[VAL_2]]#1 : (!fir.class>) -> i1 +! CHECK: %[[VAL_3:.*]] = fir.is_present %[[VAL_2]]#0 : (!fir.class>) -> i1 ! CHECK: %[[VAL_4:.*]] = fir.convert %[[VAL_3]] : (i1) -> !fir.logical<4> ! CHECK: %[[VAL_5:.*]]:3 = hlfir.associate %[[VAL_4]] {adapt.valuebyref} : (!fir.logical<4>) -> (!fir.ref>, !fir.ref>, i1) ! CHECK: fir.call @_QPtakes_logical(%[[VAL_5]]#1) fastmath : (!fir.ref>) -> () @@ -238,7 +238,7 @@ subroutine c_loc_2(x) ! CHECK-SAME: %[[VAL_0:.*]]: !fir.ref>>> {fir.bindc_name = "x", fir.optional}) { ! CHECK: %[[VAL_1:.*]] = fir.dummy_scope : !fir.dscope ! CHECK: %[[VAL_2:.*]]:2 = hlfir.declare %[[VAL_0]] dummy_scope %[[VAL_1]] {fortran_attrs = #fir.var_attrs, uniq_name = "_QFtest_present_2Ex"} : (!fir.ref>>>, !fir.dscope) -> (!fir.ref>>>, !fir.ref>>>) -! CHECK: %[[VAL_3:.*]] = fir.is_present %[[VAL_2]]#1 : (!fir.ref>>>) -> i1 +! CHECK: %[[VAL_3:.*]] = fir.is_present %[[VAL_2]]#0 : (!fir.ref>>>) -> i1 ! CHECK: %[[VAL_4:.*]] = fir.convert %[[VAL_3]] : (i1) -> !fir.logical<4> ! CHECK: %[[VAL_5:.*]]:3 = hlfir.associate %[[VAL_4]] {adapt.valuebyref} : (!fir.logical<4>) -> (!fir.ref>, !fir.ref>, i1) ! CHECK: fir.call @_QPtakes_logical(%[[VAL_5]]#1) fastmath : (!fir.ref>) -> () @@ -250,7 +250,7 @@ subroutine c_loc_2(x) ! CHECK-SAME: %[[VAL_0:.*]]: !fir.class> {fir.bindc_name = "x"}) { ! CHECK: %[[VAL_1:.*]] = fir.dummy_scope : !fir.dscope ! CHECK: %[[VAL_2:.*]]:2 = hlfir.declare %[[VAL_0]] dummy_scope %[[VAL_1]] {uniq_name = "_QFtest_is_contiguous_1Ex"} : (!fir.class>, !fir.dscope) -> (!fir.class>, !fir.class>) -! CHECK: %[[VAL_3:.*]] = fir.convert %[[VAL_2]]#1 : (!fir.class>) -> !fir.box +! CHECK: %[[VAL_3:.*]] = fir.convert %[[VAL_2]]#0 : (!fir.class>) -> !fir.box ! CHECK: %[[VAL_4:.*]] = fir.call @_FortranAIsContiguous(%[[VAL_3]]) fastmath : (!fir.box) -> i1 ! CHECK: %[[VAL_5:.*]] = fir.convert %[[VAL_4]] : (i1) -> !fir.logical<4> ! CHECK: %[[VAL_6:.*]]:3 = hlfir.associate %[[VAL_5]] {adapt.valuebyref} : (!fir.logical<4>) -> (!fir.ref>, !fir.ref>, i1) @@ -263,7 +263,7 @@ subroutine c_loc_2(x) ! CHECK-SAME: %[[VAL_0:.*]]: !fir.ref>>> {fir.bindc_name = "x"}) { ! CHECK: %[[VAL_1:.*]] = fir.dummy_scope : !fir.dscope ! CHECK: %[[VAL_2:.*]]:2 = hlfir.declare %[[VAL_0]] dummy_scope %[[VAL_1]] {fortran_attrs = #fir.var_attrs, uniq_name = "_QFtest_is_contiguous_2Ex"} : (!fir.ref>>>, !fir.dscope) -> (!fir.ref>>>, !fir.ref>>>) -! CHECK: %[[VAL_3:.*]] = fir.load %[[VAL_2]]#1 : !fir.ref>>> +! CHECK: %[[VAL_3:.*]] = fir.load %[[VAL_2]]#0 : !fir.ref>>> ! CHECK: %[[VAL_4:.*]] = fir.convert %[[VAL_3]] : (!fir.class>>) -> !fir.box ! CHECK: %[[VAL_5:.*]] = fir.call @_FortranAIsContiguous(%[[VAL_4]]) fastmath : (!fir.box) -> i1 ! CHECK: %[[VAL_6:.*]] = fir.convert %[[VAL_5]] : (i1) -> !fir.logical<4> @@ -279,8 +279,8 @@ subroutine c_loc_2(x) ! CHECK: %[[VAL_2:.*]] = fir.dummy_scope : !fir.dscope ! CHECK: %[[VAL_3:.*]]:2 = hlfir.declare %[[VAL_0]] dummy_scope %[[VAL_2]] {uniq_name = "_QFtest_same_type_as_1Ex"} : (!fir.class>, !fir.dscope) -> (!fir.class>, !fir.class>) ! CHECK: %[[VAL_4:.*]]:2 = hlfir.declare %[[VAL_1]] dummy_scope %[[VAL_2]] {uniq_name = "_QFtest_same_type_as_1Ey"} : (!fir.class>, !fir.dscope) -> (!fir.class>, !fir.class>) -! CHECK: %[[VAL_5:.*]] = fir.convert %[[VAL_3]]#1 : (!fir.class>) -> !fir.box -! CHECK: %[[VAL_6:.*]] = fir.convert %[[VAL_4]]#1 : (!fir.class>) -> !fir.box +! CHECK: %[[VAL_5:.*]] = fir.convert %[[VAL_3]]#0 : (!fir.class>) -> !fir.box +! CHECK: %[[VAL_6:.*]] = fir.convert %[[VAL_4]]#0 : (!fir.class>) -> !fir.box ! CHECK: %[[VAL_7:.*]] = fir.call @_FortranASameTypeAs(%[[VAL_5]], %[[VAL_6]]) fastmath : (!fir.box, !fir.box) -> i1 ! CHECK: %[[VAL_8:.*]] = fir.convert %[[VAL_7]] : (i1) -> !fir.logical<4> ! CHECK: %[[VAL_9:.*]]:3 = hlfir.associate %[[VAL_8]] {adapt.valuebyref} : (!fir.logical<4>) -> (!fir.ref>, !fir.ref>, i1) @@ -295,8 +295,8 @@ subroutine c_loc_2(x) ! CHECK: %[[VAL_2:.*]] = fir.dummy_scope : !fir.dscope ! CHECK: %[[VAL_3:.*]]:2 = hlfir.declare %[[VAL_0]] dummy_scope %[[VAL_2]] {fortran_attrs = #fir.var_attrs, uniq_name = "_QFtest_same_type_as_2Ex"} : (!fir.ref>>>, !fir.dscope) -> (!fir.ref>>>, !fir.ref>>>) ! CHECK: %[[VAL_4:.*]]:2 = hlfir.declare %[[VAL_1]] dummy_scope %[[VAL_2]] {fortran_attrs = #fir.var_attrs, uniq_name = "_QFtest_same_type_as_2Ey"} : (!fir.ref>>>, !fir.dscope) -> (!fir.ref>>>, !fir.ref>>>) -! CHECK: %[[VAL_5:.*]] = fir.load %[[VAL_3]]#1 : !fir.ref>>> -! CHECK: %[[VAL_6:.*]] = fir.load %[[VAL_4]]#1 : !fir.ref>>> +! CHECK: %[[VAL_5:.*]] = fir.load %[[VAL_3]]#0 : !fir.ref>>> +! CHECK: %[[VAL_6:.*]] = fir.load %[[VAL_4]]#0 : !fir.ref>>> ! CHECK: %[[VAL_7:.*]] = fir.convert %[[VAL_5]] : (!fir.class>>) -> !fir.box ! CHECK: %[[VAL_8:.*]] = fir.convert %[[VAL_6]] : (!fir.class>>) -> !fir.box ! CHECK: %[[VAL_9:.*]] = fir.call @_FortranASameTypeAs(%[[VAL_7]], %[[VAL_8]]) fastmath : (!fir.box, !fir.box) -> i1 @@ -313,8 +313,8 @@ subroutine c_loc_2(x) ! CHECK: %[[VAL_2:.*]] = fir.dummy_scope : !fir.dscope ! CHECK: %[[VAL_3:.*]]:2 = hlfir.declare %[[VAL_0]] dummy_scope %[[VAL_2]] {uniq_name = "_QFtest_extends_type_of_1Ex"} : (!fir.class>, !fir.dscope) -> (!fir.class>, !fir.class>) ! CHECK: %[[VAL_4:.*]]:2 = hlfir.declare %[[VAL_1]] dummy_scope %[[VAL_2]] {uniq_name = "_QFtest_extends_type_of_1Ey"} : (!fir.class>, !fir.dscope) -> (!fir.class>, !fir.class>) -! CHECK: %[[VAL_5:.*]] = fir.convert %[[VAL_3]]#1 : (!fir.class>) -> !fir.box -! CHECK: %[[VAL_6:.*]] = fir.convert %[[VAL_4]]#1 : (!fir.class>) -> !fir.box +! CHECK: %[[VAL_5:.*]] = fir.convert %[[VAL_3]]#0 : (!fir.class>) -> !fir.box +! CHECK: %[[VAL_6:.*]] = fir.convert %[[VAL_4]]#0 : (!fir.class>) -> !fir.box ! CHECK: %[[VAL_7:.*]] = fir.call @_FortranAExtendsTypeOf(%[[VAL_5]], %[[VAL_6]]) fastmath : (!fir.box, !fir.box) -> i1 ! CHECK: %[[VAL_8:.*]] = fir.convert %[[VAL_7]] : (i1) -> !fir.logical<4> ! CHECK: %[[VAL_9:.*]]:3 = hlfir.associate %[[VAL_8]] {adapt.valuebyref} : (!fir.logical<4>) -> (!fir.ref>, !fir.ref>, i1) @@ -329,8 +329,8 @@ subroutine c_loc_2(x) ! CHECK: %[[VAL_2:.*]] = fir.dummy_scope : !fir.dscope ! CHECK: %[[VAL_3:.*]]:2 = hlfir.declare %[[VAL_0]] dummy_scope %[[VAL_2]] {fortran_attrs = #fir.var_attrs, uniq_name = "_QFtest_extends_type_of_2Ex"} : (!fir.ref>>>, !fir.dscope) -> (!fir.ref>>>, !fir.ref>>>) ! CHECK: %[[VAL_4:.*]]:2 = hlfir.declare %[[VAL_1]] dummy_scope %[[VAL_2]] {fortran_attrs = #fir.var_attrs, uniq_name = "_QFtest_extends_type_of_2Ey"} : (!fir.ref>>>, !fir.dscope) -> (!fir.ref>>>, !fir.ref>>>) -! CHECK: %[[VAL_5:.*]] = fir.load %[[VAL_3]]#1 : !fir.ref>>> -! CHECK: %[[VAL_6:.*]] = fir.load %[[VAL_4]]#1 : !fir.ref>>> +! CHECK: %[[VAL_5:.*]] = fir.load %[[VAL_3]]#0 : !fir.ref>>> +! CHECK: %[[VAL_6:.*]] = fir.load %[[VAL_4]]#0 : !fir.ref>>> ! CHECK: %[[VAL_7:.*]] = fir.convert %[[VAL_5]] : (!fir.class>>) -> !fir.box ! CHECK: %[[VAL_8:.*]] = fir.convert %[[VAL_6]] : (!fir.class>>) -> !fir.box ! CHECK: %[[VAL_9:.*]] = fir.call @_FortranAExtendsTypeOf(%[[VAL_7]], %[[VAL_8]]) fastmath : (!fir.box, !fir.box) -> i1 @@ -348,7 +348,7 @@ subroutine c_loc_2(x) ! CHECK: %[[VAL_3:.*]] = fir.alloca !fir.type<_QM__fortran_builtinsT__builtin_c_ptr{__address:i64}> ! CHECK: %[[VAL_4:.*]] = fir.field_index __address, !fir.type<_QM__fortran_builtinsT__builtin_c_ptr{__address:i64}> ! CHECK: %[[VAL_5:.*]] = fir.coordinate_of %[[VAL_3]], %[[VAL_4]] : (!fir.ref>, !fir.field) -> !fir.ref -! CHECK: %[[VAL_6:.*]] = fir.box_addr %[[VAL_2]]#1 : (!fir.box>) -> !fir.ref> +! CHECK: %[[VAL_6:.*]] = fir.box_addr %[[VAL_2]]#0 : (!fir.box>) -> !fir.ref> ! CHECK: %[[VAL_7:.*]] = fir.convert %[[VAL_6]] : (!fir.ref>) -> i64 ! CHECK: fir.store %[[VAL_7]] to %[[VAL_5]] : !fir.ref ! CHECK: %[[VAL_8:.*]]:2 = hlfir.declare %[[VAL_3]] {uniq_name = ".tmp.intrinsic_result"} : (!fir.ref>) -> (!fir.ref>, !fir.ref>) @@ -365,7 +365,7 @@ subroutine c_loc_2(x) ! CHECK-SAME: %[[VAL_0:.*]]: !fir.ref>>> {fir.bindc_name = "x"}) { ! CHECK: %[[VAL_1:.*]] = fir.dummy_scope : !fir.dscope ! CHECK: %[[VAL_2:.*]]:2 = hlfir.declare %[[VAL_0]] dummy_scope %[[VAL_1]] {fortran_attrs = #fir.var_attrs, uniq_name = "_QFc_loc_2Ex"} : (!fir.ref>>>, !fir.dscope) -> (!fir.ref>>>, !fir.ref>>>) -! CHECK: %[[VAL_3:.*]] = fir.load %[[VAL_2]]#1 : !fir.ref>>> +! CHECK: %[[VAL_3:.*]] = fir.load %[[VAL_2]]#0 : !fir.ref>>> ! CHECK: %[[VAL_4:.*]] = fir.alloca !fir.type<_QM__fortran_builtinsT__builtin_c_ptr{__address:i64}> ! CHECK: %[[VAL_5:.*]] = fir.field_index __address, !fir.type<_QM__fortran_builtinsT__builtin_c_ptr{__address:i64}> ! CHECK: %[[VAL_6:.*]] = fir.coordinate_of %[[VAL_4]], %[[VAL_5]] : (!fir.ref>, !fir.field) -> !fir.ref diff --git a/flang/test/Lower/HLFIR/convert-variable-assumed-rank.f90 b/flang/test/Lower/HLFIR/convert-variable-assumed-rank.f90 index cd65696ca5ede91..3b60b0d04e9fa3b 100644 --- a/flang/test/Lower/HLFIR/convert-variable-assumed-rank.f90 +++ b/flang/test/Lower/HLFIR/convert-variable-assumed-rank.f90 @@ -103,7 +103,7 @@ subroutine test_assumed_length_alloc(x) ! CHECK-SAME: %[[VAL_0:.*]]: !fir.ref>>> {fir.bindc_name = "x"}) { ! CHECK: %[[VAL_1:.*]] = fir.dummy_scope : !fir.dscope ! CHECK: %[[VAL_2:.*]]:2 = hlfir.declare %[[VAL_0]] dummy_scope %[[VAL_1]] {fortran_attrs = #fir.var_attrs, uniq_name = "_QMassumed_rank_testsFtest_intentoutEx"} : (!fir.ref>>>, !fir.dscope) -> (!fir.ref>>>, !fir.ref>>>) -! CHECK: %[[VAL_3:.*]] = fir.load %[[VAL_2]]#1 : !fir.ref>>> +! CHECK: %[[VAL_3:.*]] = fir.load %[[VAL_2]]#0 : !fir.ref>>> ! CHECK: %[[VAL_4:.*]] = fir.box_addr %[[VAL_3]] : (!fir.box>>) -> !fir.heap> ! CHECK: %[[VAL_5:.*]] = fir.convert %[[VAL_4]] : (!fir.heap>) -> i64 ! CHECK: %[[VAL_6:.*]] = arith.constant 0 : i64 @@ -111,7 +111,7 @@ subroutine test_assumed_length_alloc(x) ! CHECK: fir.if %[[VAL_7]] { ! CHECK: %[[VAL_8:.*]] = arith.constant false ! CHECK: %[[VAL_9:.*]] = fir.absent !fir.box -! CHECK: %[[VAL_12:.*]] = fir.convert %[[VAL_2]]#1 : (!fir.ref>>>) -> !fir.ref> +! CHECK: %[[VAL_12:.*]] = fir.convert %[[VAL_2]]#0 : (!fir.ref>>>) -> !fir.ref> ! CHECK: %[[VAL_14:.*]] = fir.call @_FortranAAllocatableDeallocate(%[[VAL_12]], %[[VAL_8]], %[[VAL_9]], %{{.*}}, %{{.*}}) fastmath : (!fir.ref>, i1, !fir.box, !fir.ref, i32) -> i32 ! CHECK: } ! CHECK: return diff --git a/flang/test/Lower/HLFIR/select-rank.f90 b/flang/test/Lower/HLFIR/select-rank.f90 index 84987531f1419e8..211b7565bab8a36 100644 --- a/flang/test/Lower/HLFIR/select-rank.f90 +++ b/flang/test/Lower/HLFIR/select-rank.f90 @@ -319,7 +319,7 @@ subroutine test_branching(x) ! CHECK: %[[VAL_6:.*]] = fir.is_assumed_size %[[VAL_2]]#0 : (!fir.box>) -> i1 ! CHECK: cf.cond_br %[[VAL_6]], ^bb1, ^bb2 ! CHECK: ^bb1: -! CHECK: %[[VAL_7:.*]]:2 = hlfir.declare %[[VAL_2]]#1 {uniq_name = "_QFtest_simple_caseEx"} : (!fir.box>) -> (!fir.box>, !fir.box>) +! CHECK: %[[VAL_7:.*]]:2 = hlfir.declare %[[VAL_2]]#0 {uniq_name = "_QFtest_simple_caseEx"} : (!fir.box>) -> (!fir.box>, !fir.box>) ! CHECK: fir.call @_QPrdefault(%[[VAL_7]]#0) fastmath : (!fir.box>) -> () ! CHECK: cf.br ^bb6 ! CHECK: ^bb2: @@ -362,7 +362,7 @@ subroutine test_branching(x) ! CHECK: fir.call @_QPr2(%[[VAL_8]]#0) fastmath : (!fir.box>) -> () ! CHECK: cf.br ^bb6 ! CHECK: ^bb3: -! CHECK: %[[VAL_9:.*]]:2 = hlfir.declare %[[VAL_2]]#1 {uniq_name = "_QFtest_rank_starEx"} : (!fir.box>) -> (!fir.box>, !fir.box>) +! CHECK: %[[VAL_9:.*]]:2 = hlfir.declare %[[VAL_2]]#0 {uniq_name = "_QFtest_rank_starEx"} : (!fir.box>) -> (!fir.box>, !fir.box>) ! CHECK: fir.call @_QPrdefault(%[[VAL_9]]#0) fastmath : (!fir.box>) -> () ! CHECK: cf.br ^bb6 ! CHECK: ^bb4: @@ -431,7 +431,7 @@ subroutine test_branching(x) ! CHECK: fir.call @_QPr2(%[[VAL_7]]#0) fastmath : (!fir.box>) -> () ! CHECK: cf.br ^bb5 ! CHECK: ^bb3: -! CHECK: %[[VAL_8:.*]]:2 = hlfir.declare %[[VAL_2]]#1 {fortran_attrs = #fir.var_attrs, uniq_name = "_QFtest_rank_star_attributesEx"} : (!fir.box>) -> (!fir.box>, !fir.box>) +! CHECK: %[[VAL_8:.*]]:2 = hlfir.declare %[[VAL_2]]#0 {fortran_attrs = #fir.var_attrs, uniq_name = "_QFtest_rank_star_attributesEx"} : (!fir.box>) -> (!fir.box>, !fir.box>) ! CHECK: fir.call @_QPrdefault(%[[VAL_8]]#0) fastmath : (!fir.box>) -> () ! CHECK: cf.br ^bb5 ! CHECK: ^bb4: @@ -469,7 +469,7 @@ subroutine test_branching(x) ! CHECK: fir.call @_QPr2_implicit(%[[VAL_14]]#1) fastmath : (!fir.ref>) -> () ! CHECK: cf.br ^bb6 ! CHECK: ^bb3: -! CHECK: %[[VAL_15:.*]]:2 = hlfir.declare %[[VAL_2]]#1 {fortran_attrs = #fir.var_attrs, uniq_name = "_QFtest_rank_star_contiguousEx"} : (!fir.box>) -> (!fir.box>, !fir.box>) +! CHECK: %[[VAL_15:.*]]:2 = hlfir.declare %[[VAL_2]]#0 {fortran_attrs = #fir.var_attrs, uniq_name = "_QFtest_rank_star_contiguousEx"} : (!fir.box>) -> (!fir.box>, !fir.box>) ! CHECK: fir.call @_QPrdefault(%[[VAL_15]]#0) fastmath : (!fir.box>) -> () ! CHECK: cf.br ^bb6 ! CHECK: ^bb4: @@ -518,7 +518,7 @@ subroutine test_branching(x) ! CHECK: fir.call @_QPrc0_implicit(%[[VAL_16]]#0) fastmath : (!fir.boxchar<1>) -> () ! CHECK: cf.br ^bb6 ! CHECK: ^bb3: -! CHECK: %[[VAL_17:.*]]:2 = hlfir.declare %[[VAL_8]]#1 typeparams %[[VAL_7]] {uniq_name = "_QFtest_rank_star_contiguous_characterEx"} : (!fir.box>>, i64) -> (!fir.box>>, !fir.box>>) +! CHECK: %[[VAL_17:.*]]:2 = hlfir.declare %[[VAL_8]]#0 typeparams %[[VAL_7]] {uniq_name = "_QFtest_rank_star_contiguous_characterEx"} : (!fir.box>>, i64) -> (!fir.box>>, !fir.box>>) ! CHECK: fir.call @_QPrcdefault(%[[VAL_17]]#0) fastmath : (!fir.box>>) -> () ! CHECK: cf.br ^bb6 ! CHECK: ^bb4: @@ -568,7 +568,7 @@ subroutine test_branching(x) ! CHECK: fir.call @_QPra0(%[[VAL_10]]#0) fastmath : (!fir.ref>>) -> () ! CHECK: cf.br ^bb5 ! CHECK: ^bb3: -! CHECK: %[[VAL_11:.*]]:2 = hlfir.declare %[[VAL_2]]#1 {fortran_attrs = #fir.var_attrs, uniq_name = "_QFtest_simple_allocEx"} : (!fir.ref>>>) -> (!fir.ref>>>, !fir.ref>>>) +! CHECK: %[[VAL_11:.*]]:2 = hlfir.declare %[[VAL_2]]#0 {fortran_attrs = #fir.var_attrs, uniq_name = "_QFtest_simple_allocEx"} : (!fir.ref>>>) -> (!fir.ref>>>, !fir.ref>>>) ! CHECK: fir.call @_QPradefault(%[[VAL_11]]#0) fastmath : (!fir.ref>>>) -> () ! CHECK: cf.br ^bb5 ! CHECK: ^bb4: @@ -588,7 +588,7 @@ subroutine test_branching(x) ! CHECK: %[[VAL_4:.*]] = fir.box_rank %[[VAL_2]]#0 : (!fir.ref>>>>) -> i8 ! CHECK: fir.select_case %[[VAL_4]] : i8 [#fir.point, %[[VAL_3]], ^bb2, unit, ^bb1] ! CHECK: ^bb1: -! CHECK: %[[VAL_5:.*]]:2 = hlfir.declare %[[VAL_2]]#1 {fortran_attrs = #fir.var_attrs, uniq_name = "_QFtest_character_allocEx"} : (!fir.ref>>>>) -> (!fir.ref>>>>, !fir.ref>>>>) +! CHECK: %[[VAL_5:.*]]:2 = hlfir.declare %[[VAL_2]]#0 {fortran_attrs = #fir.var_attrs, uniq_name = "_QFtest_character_allocEx"} : (!fir.ref>>>>) -> (!fir.ref>>>>, !fir.ref>>>>) ! CHECK: cf.br ^bb3 ! CHECK: ^bb2: ! CHECK: %[[VAL_6:.*]] = fir.convert %[[VAL_2]]#0 : (!fir.ref>>>>) -> !fir.ref>>>> @@ -614,7 +614,7 @@ subroutine test_branching(x) ! CHECK: %[[VAL_10:.*]] = fir.box_rank %[[VAL_8]]#0 : (!fir.ref>>>>) -> i8 ! CHECK: fir.select_case %[[VAL_10]] : i8 [#fir.point, %[[VAL_9]], ^bb2, unit, ^bb1] ! CHECK: ^bb1: -! CHECK: %[[VAL_11:.*]]:2 = hlfir.declare %[[VAL_8]]#1 typeparams %[[VAL_7]] {fortran_attrs = #fir.var_attrs, uniq_name = "_QFtest_explicit_character_ptrEx"} : (!fir.ref>>>>, i64) -> (!fir.ref>>>>, !fir.ref>>>>) +! CHECK: %[[VAL_11:.*]]:2 = hlfir.declare %[[VAL_8]]#0 typeparams %[[VAL_7]] {fortran_attrs = #fir.var_attrs, uniq_name = "_QFtest_explicit_character_ptrEx"} : (!fir.ref>>>>, i64) -> (!fir.ref>>>>, !fir.ref>>>>) ! CHECK: cf.br ^bb3 ! CHECK: ^bb2: ! CHECK: %[[VAL_12:.*]] = fir.convert %[[VAL_8]]#0 : (!fir.ref>>>>) -> !fir.ref>>> @@ -634,7 +634,7 @@ subroutine test_branching(x) ! CHECK: %[[VAL_6:.*]] = fir.box_rank %[[VAL_4]]#0 : (!fir.ref>>>>) -> i8 ! CHECK: fir.select_case %[[VAL_6]] : i8 [#fir.point, %[[VAL_5]], ^bb2, unit, ^bb1] ! CHECK: ^bb1: -! CHECK: %[[VAL_7:.*]]:2 = hlfir.declare %[[VAL_4]]#1 typeparams %[[VAL_3]] {fortran_attrs = #fir.var_attrs, uniq_name = "_QFtest_assumed_character_ptrEx"} : (!fir.ref>>>>, index) -> (!fir.ref>>>>, !fir.ref>>>>) +! CHECK: %[[VAL_7:.*]]:2 = hlfir.declare %[[VAL_4]]#0 typeparams %[[VAL_3]] {fortran_attrs = #fir.var_attrs, uniq_name = "_QFtest_assumed_character_ptrEx"} : (!fir.ref>>>>, index) -> (!fir.ref>>>>, !fir.ref>>>>) ! CHECK: cf.br ^bb3 ! CHECK: ^bb2: ! CHECK: %[[VAL_8:.*]] = fir.convert %[[VAL_4]]#0 : (!fir.ref>>>>) -> !fir.ref>>> @@ -661,7 +661,7 @@ subroutine test_branching(x) ! CHECK: fir.call @_QPrup1(%[[VAL_8]]#0) fastmath : (!fir.class>) -> () ! CHECK: cf.br ^bb5 ! CHECK: ^bb3: -! CHECK: %[[VAL_9:.*]]:2 = hlfir.declare %[[VAL_2]]#1 {uniq_name = "_QFtest_polymorphicEx"} : (!fir.class>) -> (!fir.class>, !fir.class>) +! CHECK: %[[VAL_9:.*]]:2 = hlfir.declare %[[VAL_2]]#0 {uniq_name = "_QFtest_polymorphicEx"} : (!fir.class>) -> (!fir.class>, !fir.class>) ! CHECK: fir.call @_QPrupdefault(%[[VAL_9]]#0) fastmath : (!fir.class>) -> () ! CHECK: cf.br ^bb5 ! CHECK: ^bb4: @@ -711,7 +711,7 @@ subroutine test_branching(x) ! CHECK: fir.call @_QPr1(%[[VAL_20]]#0) fastmath : (!fir.box>) -> () ! CHECK: cf.br ^bb7 ! CHECK: ^bb6: -! CHECK: %[[VAL_21:.*]]:2 = hlfir.declare %[[VAL_4]]#1 {uniq_name = "_QFtest_nested_select_rankEx2"} : (!fir.box>) -> (!fir.box>, !fir.box>) +! CHECK: %[[VAL_21:.*]]:2 = hlfir.declare %[[VAL_4]]#0 {uniq_name = "_QFtest_nested_select_rankEx2"} : (!fir.box>) -> (!fir.box>, !fir.box>) ! CHECK: fir.call @_QPr0(%[[VAL_11]]#1) fastmath : (!fir.ref) -> () ! CHECK: fir.call @_QPrdefault(%[[VAL_21]]#0) fastmath : (!fir.box>) -> () ! CHECK: cf.br ^bb7 @@ -741,14 +741,14 @@ subroutine test_branching(x) ! CHECK: fir.call @_QPr1(%[[VAL_32]]#0) fastmath : (!fir.box>) -> () ! CHECK: cf.br ^bb13 ! CHECK: ^bb12: -! CHECK: %[[VAL_33:.*]]:2 = hlfir.declare %[[VAL_4]]#1 {uniq_name = "_QFtest_nested_select_rankEx2"} : (!fir.box>) -> (!fir.box>, !fir.box>) +! CHECK: %[[VAL_33:.*]]:2 = hlfir.declare %[[VAL_4]]#0 {uniq_name = "_QFtest_nested_select_rankEx2"} : (!fir.box>) -> (!fir.box>, !fir.box>) ! CHECK: fir.call @_QPr1(%[[VAL_23]]#0) fastmath : (!fir.box>) -> () ! CHECK: fir.call @_QPrdefault(%[[VAL_33]]#0) fastmath : (!fir.box>) -> () ! CHECK: cf.br ^bb13 ! CHECK: ^bb13: ! CHECK: cf.br ^bb20 ! CHECK: ^bb14: -! CHECK: %[[VAL_34:.*]]:2 = hlfir.declare %[[VAL_3]]#1 {uniq_name = "_QFtest_nested_select_rankEx1"} : (!fir.box>) -> (!fir.box>, !fir.box>) +! CHECK: %[[VAL_34:.*]]:2 = hlfir.declare %[[VAL_3]]#0 {uniq_name = "_QFtest_nested_select_rankEx1"} : (!fir.box>) -> (!fir.box>, !fir.box>) ! CHECK: %[[VAL_35:.*]] = arith.constant 0 : i8 ! CHECK: %[[VAL_36:.*]] = arith.constant 1 : i8 ! CHECK: %[[VAL_37:.*]] = fir.is_assumed_size %[[VAL_4]]#0 : (!fir.box>) -> i1 @@ -770,7 +770,7 @@ subroutine test_branching(x) ! CHECK: fir.call @_QPr1(%[[VAL_43]]#0) fastmath : (!fir.box>) -> () ! CHECK: cf.br ^bb19 ! CHECK: ^bb18: -! CHECK: %[[VAL_44:.*]]:2 = hlfir.declare %[[VAL_4]]#1 {uniq_name = "_QFtest_nested_select_rankEx2"} : (!fir.box>) -> (!fir.box>, !fir.box>) +! CHECK: %[[VAL_44:.*]]:2 = hlfir.declare %[[VAL_4]]#0 {uniq_name = "_QFtest_nested_select_rankEx2"} : (!fir.box>) -> (!fir.box>, !fir.box>) ! CHECK: fir.call @_QPrdefault(%[[VAL_34]]#0) fastmath : (!fir.box>) -> () ! CHECK: fir.call @_QPrdefault(%[[VAL_44]]#0) fastmath : (!fir.box>) -> () ! CHECK: cf.br ^bb19 @@ -789,7 +789,7 @@ subroutine test_branching(x) ! CHECK: %[[VAL_5:.*]] = fir.is_assumed_size %[[VAL_2]]#0 : (!fir.box>) -> i1 ! CHECK: cf.cond_br %[[VAL_5]], ^bb1, ^bb2 ! CHECK: ^bb1: -! CHECK: %[[VAL_6:.*]]:2 = hlfir.declare %[[VAL_2]]#1 {uniq_name = "_QFtest_branchingEx"} : (!fir.box>) -> (!fir.box>, !fir.box>) +! CHECK: %[[VAL_6:.*]]:2 = hlfir.declare %[[VAL_2]]#0 {uniq_name = "_QFtest_branchingEx"} : (!fir.box>) -> (!fir.box>, !fir.box>) ! CHECK: %[[VAL_7:.*]] = fir.call @_QPjump() fastmath : () -> !fir.logical<4> ! CHECK: %[[VAL_8:.*]] = fir.convert %[[VAL_7]] : (!fir.logical<4>) -> i1 ! CHECK: %[[VAL_9:.*]] = arith.constant true From a6129a56d3640e70198d91fc97b4fc65525cb59c Mon Sep 17 00:00:00 2001 From: Tom Eccles Date: Mon, 10 Jun 2024 10:43:56 +0100 Subject: [PATCH 33/36] [flang][Transforms][NFC] reduce boilerplate in func attr pass (#94739) Use tablegen to automatically create the pass constructor. The purpose of this pass is to add attributes to functions, so it doesn't need to work on other top level operations. --- .../flang/Optimizer/Transforms/Passes.h | 12 +---------- .../flang/Optimizer/Transforms/Passes.td | 1 - flang/include/flang/Tools/CLOptions.inc | 14 ++++++------- .../lib/Optimizer/Transforms/FunctionAttr.cpp | 20 ------------------- 4 files changed, 7 insertions(+), 40 deletions(-) diff --git a/flang/include/flang/Optimizer/Transforms/Passes.h b/flang/include/flang/Optimizer/Transforms/Passes.h index 4e978e6c9cdec22..9fa819e2bf502e2 100644 --- a/flang/include/flang/Optimizer/Transforms/Passes.h +++ b/flang/include/flang/Optimizer/Transforms/Passes.h @@ -55,6 +55,7 @@ namespace fir { #define GEN_PASS_DECL_OMPMARKDECLARETARGETPASS #define GEN_PASS_DECL_OMPFUNCTIONFILTERING #define GEN_PASS_DECL_VSCALEATTR +#define GEN_PASS_DECL_FUNCTIONATTR #include "flang/Optimizer/Transforms/Passes.h.inc" std::unique_ptr createAffineDemotionPass(); @@ -75,17 +76,6 @@ std::unique_ptr createVScaleAttrPass(); std::unique_ptr createVScaleAttrPass(std::pair vscaleAttr); -struct FunctionAttrTypes { - mlir::LLVM::framePointerKind::FramePointerKind framePointerKind = - mlir::LLVM::framePointerKind::FramePointerKind::None; -}; - -std::unique_ptr createFunctionAttrPass(); -std::unique_ptr -createFunctionAttrPass(FunctionAttrTypes &functionAttr, bool noInfsFPMath, - bool noNaNsFPMath, bool approxFuncFPMath, - bool noSignedZerosFPMath, bool unsafeFPMath); - void populateCfgConversionRewrites(mlir::RewritePatternSet &patterns, bool forceLoopToExecuteOnce = false, bool setNSW = false); diff --git a/flang/include/flang/Optimizer/Transforms/Passes.td b/flang/include/flang/Optimizer/Transforms/Passes.td index 4e281e2846c56df..7a3baca4c19dac9 100644 --- a/flang/include/flang/Optimizer/Transforms/Passes.td +++ b/flang/include/flang/Optimizer/Transforms/Passes.td @@ -394,7 +394,6 @@ def FunctionAttr : Pass<"function-attr", "mlir::func::FuncOp"> { "bool", /*default=*/"false", "Set the unsafe-fp-math attribute on functions in the module.">, ]; - let constructor = "::fir::createFunctionAttrPass()"; } def AssumedRankOpConversion : Pass<"fir-assumed-rank-op", "mlir::ModuleOp"> { diff --git a/flang/include/flang/Tools/CLOptions.inc b/flang/include/flang/Tools/CLOptions.inc index 528c51d8c1c4411..2a0cfc04aa35099 100644 --- a/flang/include/flang/Tools/CLOptions.inc +++ b/flang/include/flang/Tools/CLOptions.inc @@ -372,24 +372,22 @@ inline void createDefaultFIRCodeGenPassPipeline(mlir::PassManager &pm, pm.addPass(fir::createVScaleAttr({{config.VScaleMin, config.VScaleMax}})); // Add function attributes - fir::FunctionAttrTypes functionAttrs; + mlir::LLVM::framePointerKind::FramePointerKind framePointerKind; if (config.FramePointerKind != llvm::FramePointerKind::None || config.NoInfsFPMath || config.NoNaNsFPMath || config.ApproxFuncFPMath || config.NoSignedZerosFPMath || config.UnsafeFPMath) { if (config.FramePointerKind == llvm::FramePointerKind::NonLeaf) - functionAttrs.framePointerKind = + framePointerKind = mlir::LLVM::framePointerKind::FramePointerKind::NonLeaf; else if (config.FramePointerKind == llvm::FramePointerKind::All) - functionAttrs.framePointerKind = - mlir::LLVM::framePointerKind::FramePointerKind::All; + framePointerKind = mlir::LLVM::framePointerKind::FramePointerKind::All; else - functionAttrs.framePointerKind = - mlir::LLVM::framePointerKind::FramePointerKind::None; + framePointerKind = mlir::LLVM::framePointerKind::FramePointerKind::None; - pm.addPass(fir::createFunctionAttrPass(functionAttrs, config.NoInfsFPMath, + pm.addPass(fir::createFunctionAttr({framePointerKind, config.NoInfsFPMath, config.NoNaNsFPMath, config.ApproxFuncFPMath, - config.NoSignedZerosFPMath, config.UnsafeFPMath)); + config.NoSignedZerosFPMath, config.UnsafeFPMath})); } fir::addFIRToLLVMPass(pm, config); diff --git a/flang/lib/Optimizer/Transforms/FunctionAttr.cpp b/flang/lib/Optimizer/Transforms/FunctionAttr.cpp index f54080f21b9675a..69c41595974ef9d 100644 --- a/flang/lib/Optimizer/Transforms/FunctionAttr.cpp +++ b/flang/lib/Optimizer/Transforms/FunctionAttr.cpp @@ -15,7 +15,6 @@ #include "mlir/Dialect/LLVMIR/LLVMDialect.h" namespace fir { -#define GEN_PASS_DECL_FUNCTIONATTR #define GEN_PASS_DEF_FUNCTIONATTR #include "flang/Optimizer/Transforms/Passes.h.inc" } // namespace fir @@ -76,22 +75,3 @@ void FunctionAttrPass::runOnOperation() { LLVM_DEBUG(llvm::dbgs() << "=== End " DEBUG_TYPE " ===\n"); } - -std::unique_ptr fir::createFunctionAttrPass( - fir::FunctionAttrTypes &functionAttr, bool noInfsFPMath, bool noNaNsFPMath, - bool approxFuncFPMath, bool noSignedZerosFPMath, bool unsafeFPMath) { - FunctionAttrOptions opts; - // Frame pointer - opts.framePointerKind = functionAttr.framePointerKind; - opts.noInfsFPMath = noInfsFPMath; - opts.noNaNsFPMath = noNaNsFPMath; - opts.approxFuncFPMath = approxFuncFPMath; - opts.noSignedZerosFPMath = noSignedZerosFPMath; - opts.unsafeFPMath = unsafeFPMath; - - return std::make_unique(opts); -} - -std::unique_ptr fir::createFunctionAttrPass() { - return std::make_unique(); -} From ae9d89d53e64bad514171c99f16d553febfbac12 Mon Sep 17 00:00:00 2001 From: Oleksandr T Date: Mon, 10 Jun 2024 12:58:56 +0300 Subject: [PATCH 34/36] [Clang][C++23] update constexpr diagnostics for missing return statements per P2448 (#94123) Fixes #92583 --- clang/lib/Sema/SemaDeclCXX.cpp | 36 ++++++++++++------- .../CXX/dcl.dcl/dcl.spec/dcl.constexpr/p3.cpp | 2 +- .../SemaCXX/constant-expression-cxx14.cpp | 2 +- .../constexpr-return-non-void-cxx2b.cpp | 7 ++++ 4 files changed, 33 insertions(+), 14 deletions(-) create mode 100644 clang/test/SemaCXX/constexpr-return-non-void-cxx2b.cpp diff --git a/clang/lib/Sema/SemaDeclCXX.cpp b/clang/lib/Sema/SemaDeclCXX.cpp index e224e79fdb8ea1a..37f0df2a6a27d5f 100644 --- a/clang/lib/Sema/SemaDeclCXX.cpp +++ b/clang/lib/Sema/SemaDeclCXX.cpp @@ -1806,6 +1806,7 @@ static unsigned getRecordDiagFromTagKind(TagTypeKind Tag) { static bool CheckConstexprFunctionBody(Sema &SemaRef, const FunctionDecl *Dcl, Stmt *Body, Sema::CheckConstexprKind Kind); +static bool CheckConstexprMissingReturn(Sema &SemaRef, const FunctionDecl *Dcl); // Check whether a function declaration satisfies the requirements of a // constexpr function definition or a constexpr constructor definition. If so, @@ -2411,20 +2412,9 @@ static bool CheckConstexprFunctionBody(Sema &SemaRef, const FunctionDecl *Dcl, } } else { if (ReturnStmts.empty()) { - // C++1y doesn't require constexpr functions to contain a 'return' - // statement. We still do, unless the return type might be void, because - // otherwise if there's no return statement, the function cannot - // be used in a core constant expression. - bool OK = SemaRef.getLangOpts().CPlusPlus14 && - (Dcl->getReturnType()->isVoidType() || - Dcl->getReturnType()->isDependentType()); switch (Kind) { case Sema::CheckConstexprKind::Diagnose: - SemaRef.Diag(Dcl->getLocation(), - OK ? diag::warn_cxx11_compat_constexpr_body_no_return - : diag::err_constexpr_body_no_return) - << Dcl->isConsteval(); - if (!OK) + if (!CheckConstexprMissingReturn(SemaRef, Dcl)) return false; break; @@ -2494,6 +2484,28 @@ static bool CheckConstexprFunctionBody(Sema &SemaRef, const FunctionDecl *Dcl, return true; } +static bool CheckConstexprMissingReturn(Sema &SemaRef, + const FunctionDecl *Dcl) { + bool IsVoidOrDependentType = Dcl->getReturnType()->isVoidType() || + Dcl->getReturnType()->isDependentType(); + // Skip emitting a missing return error diagnostic for non-void functions + // since C++23 no longer mandates constexpr functions to yield constant + // expressions. + if (SemaRef.getLangOpts().CPlusPlus23 && !IsVoidOrDependentType) + return true; + + // C++14 doesn't require constexpr functions to contain a 'return' + // statement. We still do, unless the return type might be void, because + // otherwise if there's no return statement, the function cannot + // be used in a core constant expression. + bool OK = SemaRef.getLangOpts().CPlusPlus14 && IsVoidOrDependentType; + SemaRef.Diag(Dcl->getLocation(), + OK ? diag::warn_cxx11_compat_constexpr_body_no_return + : diag::err_constexpr_body_no_return) + << Dcl->isConsteval(); + return OK; +} + bool Sema::CheckImmediateEscalatingFunctionDefinition( FunctionDecl *FD, const sema::FunctionScopeInfo *FSI) { if (!getLangOpts().CPlusPlus20 || !FD->isImmediateEscalating()) diff --git a/clang/test/CXX/dcl.dcl/dcl.spec/dcl.constexpr/p3.cpp b/clang/test/CXX/dcl.dcl/dcl.spec/dcl.constexpr/p3.cpp index 4416c825226494b..51990ee4341d29b 100644 --- a/clang/test/CXX/dcl.dcl/dcl.spec/dcl.constexpr/p3.cpp +++ b/clang/test/CXX/dcl.dcl/dcl.spec/dcl.constexpr/p3.cpp @@ -212,7 +212,7 @@ constexpr int ClassDecl3() { return 0; } -constexpr int NoReturn() {} // expected-error {{no return statement in constexpr function}} +constexpr int NoReturn() {} // beforecxx23-error {{no return statement in constexpr function}} constexpr int MultiReturn() { return 0; // beforecxx14-note {{return statement}} return 0; // beforecxx14-warning {{multiple return statements in constexpr function}} diff --git a/clang/test/SemaCXX/constant-expression-cxx14.cpp b/clang/test/SemaCXX/constant-expression-cxx14.cpp index 80a7a2dd31531c4..70ab5dcd357c1ca 100644 --- a/clang/test/SemaCXX/constant-expression-cxx14.cpp +++ b/clang/test/SemaCXX/constant-expression-cxx14.cpp @@ -82,7 +82,7 @@ constexpr void k() { // If the return type is not 'void', no return statements => never a constant // expression, so still diagnose that case. -[[noreturn]] constexpr int fn() { // expected-error {{no return statement in constexpr function}} +[[noreturn]] constexpr int fn() { // cxx14_20-error {{no return statement in constexpr function}} fn(); } diff --git a/clang/test/SemaCXX/constexpr-return-non-void-cxx2b.cpp b/clang/test/SemaCXX/constexpr-return-non-void-cxx2b.cpp new file mode 100644 index 000000000000000..25d1f8df7f71663 --- /dev/null +++ b/clang/test/SemaCXX/constexpr-return-non-void-cxx2b.cpp @@ -0,0 +1,7 @@ +// RUN: %clang_cc1 -std=c++23 -fsyntax-only -verify %s + +constexpr int f() { } // expected-warning {{non-void function does not return a value}} +static_assert(__is_same(decltype([] constexpr -> int { }( )), int)); // expected-warning {{non-void lambda does not return a value}} + +consteval int g() { } // expected-warning {{non-void function does not return a value}} +static_assert(__is_same(decltype([] consteval -> int { }( )), int)); // expected-warning {{non-void lambda does not return a value}} From ecb9d94a8bfb07591fdfcf3418c5cfafcad5a13d Mon Sep 17 00:00:00 2001 From: Jay Foad Date: Mon, 10 Jun 2024 11:03:18 +0100 Subject: [PATCH 35/36] [KnownBits] Speed up conflict handling in ForeachKnownBits in unit test. (#94943) Exit early if known bits have a conflict. This gives me a ~15% speed up when running this in my Release+Asserts build: $ unittests/Support/SupportTests --gtest_filter=KnownBitsTest.*Exhaustive --- llvm/unittests/Support/KnownBitsTest.h | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/llvm/unittests/Support/KnownBitsTest.h b/llvm/unittests/Support/KnownBitsTest.h index 974051891d48e7f..807ce31381efbf4 100644 --- a/llvm/unittests/Support/KnownBitsTest.h +++ b/llvm/unittests/Support/KnownBitsTest.h @@ -38,6 +38,12 @@ void ForeachNumInKnownBits(const KnownBits &Known, FnTy Fn) { unsigned Max = 1u << Bits; unsigned Zero = Known.Zero.getZExtValue(); unsigned One = Known.One.getZExtValue(); + + if (Zero & One) { + // Known has a conflict. No values will satisfy it. + return; + } + for (unsigned N = 0; N < Max; ++N) { if ((N & Zero) == 0 && (~N & One) == 0) Fn(APInt(Bits, N)); From 8dc8b9f168f283beed02867f0917b62f1e2c4a4b Mon Sep 17 00:00:00 2001 From: Tom Eccles Date: Mon, 10 Jun 2024 11:14:27 +0100 Subject: [PATCH 36/36] [flang][OpenMP] Fix unused prefixes in function-filtering-2 test (#94330) Co-authored-by: Andrew Gozillon --- flang/test/Lower/OpenMP/function-filtering-2.f90 | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/flang/test/Lower/OpenMP/function-filtering-2.f90 b/flang/test/Lower/OpenMP/function-filtering-2.f90 index e1d0f72b9f525a9..f367069efb3d920 100644 --- a/flang/test/Lower/OpenMP/function-filtering-2.f90 +++ b/flang/test/Lower/OpenMP/function-filtering-2.f90 @@ -1,6 +1,6 @@ -! RUN: %flang_fc1 -fopenmp -flang-experimental-hlfir -emit-llvm %s -o - | FileCheck --check-prefix=LLVM %s +! RUN: %flang_fc1 -fopenmp -flang-experimental-hlfir -emit-llvm %s -o - | FileCheck --check-prefixes=LLVM,LLVM-HOST %s ! RUN: %flang_fc1 -fopenmp -emit-hlfir %s -o - | FileCheck --check-prefix=MLIR %s -! RUN: %flang_fc1 -fopenmp -fopenmp-is-target-device -flang-experimental-hlfir -emit-llvm %s -o - | FileCheck --check-prefix=LLVM %s +! RUN: %flang_fc1 -fopenmp -fopenmp-is-target-device -flang-experimental-hlfir -emit-llvm %s -o - | FileCheck --check-prefixes=LLVM,LLVM-DEVICE %s ! RUN: %flang_fc1 -fopenmp -fopenmp-is-target-device -emit-hlfir %s -o - | FileCheck --check-prefix=MLIR %s ! RUN: bbc -fopenmp -emit-hlfir %s -o - | FileCheck --check-prefixes=MLIR-HOST,MLIR-ALL %s ! RUN: bbc -fopenmp -fopenmp-is-target-device -emit-hlfir %s -o - | FileCheck --check-prefixes=MLIR-DEVICE,MLIR-ALL %s @@ -38,7 +38,7 @@ end subroutine no_declaretarget ! MLIR-ALL: return ! LLVM-HOST: define {{.*}} @{{.*}}main{{.*}}( -! LLVM-HOST-NOT: {{.*}} @{{.*}}__omp_offloading{{.*}}main_{{.*}}( +! LLVM-HOST: {{.*}} @{{.*}}__omp_offloading{{.*}}main_{{.*}}( ! LLVM-DEVICE-NOT: {{.*}} @{{.*}}main{{.*}}( ! LLVM-DEVICE: define {{.*}} @{{.*}}__omp_offloading{{.*}}main_{{.*}}( program main