Merge branch 'main' into xegpu-reduction-unit-dim

Author: nbpatel

.github/dependabot.yml

@@ -1,9 +1,8 @@
 // Test that compiling using a PCH doesn't leak file descriptors.
 // https://bugs.chromium.org/p/chromium/issues/detail?id=924225
 //
-// This test requires bash loops and ulimit.
-// REQUIRES: shell
-// UNSUPPORTED: target={{.*win32.*}}
+// This test uses ulimit.
+// UNSUPPORTED: system-windows
 //
 // Set up source files. lib/lib.h includes lots of lib*.h files in that dir.
 // client.c includes lib/lib.h, and also the individual files directly.
@@ -12,10 +11,10 @@
 // RUN: mkdir %t
 // RUN: cd %t
 // RUN: mkdir lib
-// RUN: for i in {1..300}; do touch lib/lib$i.h; done
-// RUN: for i in {1..300}; do echo "#include \"lib$i.h\"" >> lib/lib.h; done
+// RUN: %python -c "from pathlib import Path; list(map(lambda i: Path(f'lib/lib{i}.h').touch(), range(1, 301)))"
+// RUN: %python -c "for i in range(1, 301): print(f'#include \"lib{i}.h\"')" > lib/lib.h
 // RUN: echo "#include \"lib/lib.h\"" > client.c
-// RUN: for i in {1..300}; do echo "#include \"lib/lib$i.h\"" >> client.c; done
+// RUN: %python -c "for i in range(1, 301): print(f'#include \"lib/lib{i}.h\"')" > client.c
 //
 // We want to verify that we don't hold all the files open at the same time.
 // This is important e.g. on mac, which has a low default FD limit.

clang/test/PCH/leakfiles.test

@@ -1,7 +1,6 @@
 // Only run this test where ulimit is known to work well.
 // (There's nothing really platform-specific being tested, this is just ulimit).
 //
-// REQUIRES: shell
 // REQUIRES: system-linux
 // UNSUPPORTED: msan
 // UNSUPPORTED: asan

clang/test/SemaCXX/PR51712-large-array-constexpr-check-oom.cpp

@@ -1161,24 +1161,6 @@ vector<_Tp, _Allocator>::__emplace_back_slow_path(_Args&&... __args) {
   return this->__end_;
 }
 
-// This makes the compiler inline `__else()` if `__cond` is known to be false. Currently LLVM doesn't do that without
-// the `__builtin_constant_p`, since it considers `__else` unlikely even through it's known to be run.
-// See https://llvm.org/PR154292
-template <class _If, class _Else>
-_LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX14 void __if_likely_else(bool __cond, _If __if, _Else __else) {
-  if (__builtin_constant_p(__cond)) {
-    if (__cond)
-      __if();
-    else
-      __else();
-  } else {
-    if (__cond) [[__likely__]]
-      __if();
-    else
-      __else();
-  }
-}
-
 template <class _Tp, class _Allocator>
 template <class... _Args>
 _LIBCPP_CONSTEXPR_SINCE_CXX20 inline
@@ -1189,14 +1171,12 @@ _LIBCPP_CONSTEXPR_SINCE_CXX20 inline
 #endif
     vector<_Tp, _Allocator>::emplace_back(_Args&&... __args) {
   pointer __end = this->__end_;
-  std::__if_likely_else(
-      __end < this->__cap_,
-      [&] {
-        __emplace_back_assume_capacity(std::forward<_Args>(__args)...);
-        ++__end;
-      },
-      [&] { __end = __emplace_back_slow_path(std::forward<_Args>(__args)...); });
-
+  if (__end < this->__cap_) {
+    __emplace_back_assume_capacity(std::forward<_Args>(__args)...);
+    ++__end;
+  } else {
+    __end = __emplace_back_slow_path(std::forward<_Args>(__args)...);
+  }
   this->__end_ = __end;
 #if _LIBCPP_STD_VER >= 17
   return *(__end - 1);

libcxx/include/__vector/vector.h

@@ -327,6 +327,11 @@ class ArchSpec {
   ///  \return a boolean value.
   bool IsMIPS() const;
 
+  /// If NVPTX architecture return true.
+  ///
+  ///  \return a boolean value.
+  bool IsNVPTX() const;
+
   /// Returns a string representing current architecture as a target CPU for
   /// tools like compiler, disassembler etc.
   ///

lldb/include/lldb/Utility/ArchSpec.h

@@ -545,6 +545,8 @@ const char *ArchSpec::GetArchitectureName() const {
 
 bool ArchSpec::IsMIPS() const { return GetTriple().isMIPS(); }
 
+bool ArchSpec::IsNVPTX() const { return GetTriple().isNVPTX(); }
+
 std::string ArchSpec::GetTargetABI() const {
 
   std::string abi;

lldb/source/Utility/ArchSpec.cpp

@@ -92,7 +92,7 @@ struct RootDescriptorYaml {
 };
 
 struct DescriptorRangeYaml {
-  uint32_t RangeType;
+  dxil::ResourceClass RangeType;
   uint32_t NumDescriptors;
   uint32_t BaseShaderRegister;
   uint32_t RegisterSpace;
@@ -111,12 +111,12 @@ struct DescriptorTableYaml {
 };
 
 struct RootParameterHeaderYaml {
-  uint32_t Type;
-  uint32_t Visibility;
+  dxbc::RootParameterType Type;
+  dxbc::ShaderVisibility Visibility;
   uint32_t Offset;
 
   RootParameterHeaderYaml(){};
-  RootParameterHeaderYaml(uint32_t T) : Type(T) {}
+  RootParameterHeaderYaml(dxbc::RootParameterType T) : Type(T) {}
 };
 
 struct RootParameterLocationYaml {
@@ -165,21 +165,19 @@ struct RootParameterYamlDesc {
 };
 
 struct StaticSamplerYamlDesc {
-  uint32_t Filter = llvm::to_underlying(dxbc::SamplerFilter::Anisotropic);
-  uint32_t AddressU = llvm::to_underlying(dxbc::TextureAddressMode::Wrap);
-  uint32_t AddressV = llvm::to_underlying(dxbc::TextureAddressMode::Wrap);
-  uint32_t AddressW = llvm::to_underlying(dxbc::TextureAddressMode::Wrap);
+  dxbc::SamplerFilter Filter = dxbc::SamplerFilter::Anisotropic;
+  dxbc::TextureAddressMode AddressU = dxbc::TextureAddressMode::Wrap;
+  dxbc::TextureAddressMode AddressV = dxbc::TextureAddressMode::Wrap;
+  dxbc::TextureAddressMode AddressW = dxbc::TextureAddressMode::Wrap;
   float MipLODBias = 0.f;
   uint32_t MaxAnisotropy = 16u;
-  uint32_t ComparisonFunc =
-      llvm::to_underlying(dxbc::ComparisonFunc::LessEqual);
-  uint32_t BorderColor =
-      llvm::to_underlying(dxbc::StaticBorderColor::OpaqueWhite);
+  dxbc::ComparisonFunc ComparisonFunc = dxbc::ComparisonFunc::LessEqual;
+  dxbc::StaticBorderColor BorderColor = dxbc::StaticBorderColor::OpaqueWhite;
   float MinLOD = 0.f;
   float MaxLOD = std::numeric_limits<float>::max();
   uint32_t ShaderRegister;
   uint32_t RegisterSpace;
-  uint32_t ShaderVisibility;
+  dxbc::ShaderVisibility ShaderVisibility;
 };
 
 struct RootSignatureYamlDesc {
@@ -321,6 +319,13 @@ LLVM_YAML_DECLARE_ENUM_TRAITS(llvm::dxbc::PSV::ResourceKind)
 LLVM_YAML_DECLARE_ENUM_TRAITS(llvm::dxbc::D3DSystemValue)
 LLVM_YAML_DECLARE_ENUM_TRAITS(llvm::dxbc::SigComponentType)
 LLVM_YAML_DECLARE_ENUM_TRAITS(llvm::dxbc::SigMinPrecision)
+LLVM_YAML_DECLARE_ENUM_TRAITS(llvm::dxbc::RootParameterType)
+LLVM_YAML_DECLARE_ENUM_TRAITS(dxil::ResourceClass)
+LLVM_YAML_DECLARE_ENUM_TRAITS(llvm::dxbc::SamplerFilter)
+LLVM_YAML_DECLARE_ENUM_TRAITS(llvm::dxbc::StaticBorderColor)
+LLVM_YAML_DECLARE_ENUM_TRAITS(llvm::dxbc::TextureAddressMode)
+LLVM_YAML_DECLARE_ENUM_TRAITS(llvm::dxbc::ShaderVisibility)
+LLVM_YAML_DECLARE_ENUM_TRAITS(llvm::dxbc::ComparisonFunc)
 
 namespace llvm {
 

llvm/include/llvm/ObjectYAML/DXContainerYAML.h

@@ -102,7 +102,6 @@ const unsigned MinWaveSize = 4;
 const unsigned MaxWaveSize = 128;
 
 LLVM_ABI StringRef getResourceClassName(ResourceClass RC);
-
 } // namespace dxil
 } // namespace llvm
 

llvm/include/llvm/Support/DXILABI.h

@@ -3217,26 +3217,18 @@ const SCEV *ScalarEvolution::getMulExpr(SmallVectorImpl<const SCEV *> &Ops,
       }
 
       // Try to fold (C1 * D /u C2) -> C1/C2 * D, if C1 and C2 are powers-of-2,
-      // D is a multiple of C2, and C1 is a multiple of C2. If C2 is a multiple
-      // of C1, fold to (D /u (C2 /u C1)).
+      // D is a multiple of C2, and C1 is a multiple of C2.
       const SCEV *D;
       APInt C1V = LHSC->getAPInt();
-      // (C1 * D /u C2) == -1 * -C1 * D /u C2 when C1 != INT_MIN. Don't treat -1
-      // as -1 * 1, as it won't enable additional folds.
-      if (C1V.isNegative() && !C1V.isMinSignedValue() && !C1V.isAllOnes())
+      // (C1 * D /u C2) == -1 * -C1 * D /u C2 when C1 != INT_MIN.
+      if (C1V.isNegative() && !C1V.isMinSignedValue())
         C1V = C1V.abs();
       const SCEVConstant *C2;
       if (C1V.isPowerOf2() &&
           match(Ops[1], m_scev_UDiv(m_SCEV(D), m_SCEVConstant(C2))) &&
-          C2->getAPInt().isPowerOf2() &&
+          C2->getAPInt().isPowerOf2() && C1V.uge(C2->getAPInt()) &&
           C1V.logBase2() <= getMinTrailingZeros(D)) {
-        const SCEV *NewMul;
-        if (C1V.uge(C2->getAPInt())) {
-          NewMul = getMulExpr(getUDivExpr(getConstant(C1V), C2), D);
-        } else {
-          assert(C1V.ugt(1) && "C1 <= 1 should have been folded earlier");
-          NewMul = getUDivExpr(D, getUDivExpr(C2, getConstant(C1V)));
-        }
+        const SCEV *NewMul = getMulExpr(getUDivExpr(getConstant(C1V), C2), D);
         return C1V == LHSC->getAPInt() ? NewMul : getNegativeSCEV(NewMul);
       }
     }

llvm/lib/Analysis/ScalarEvolution.cpp

@@ -740,14 +740,6 @@ Intrinsic::ID Intrinsic::lookupIntrinsicID(StringRef Name) {
 #include "llvm/IR/IntrinsicImpl.inc"
 #undef GET_INTRINSIC_ATTRIBUTES
 
-AttributeSet Intrinsic::getFnAttributes(LLVMContext &C, ID id) {
-  if (id == 0)
-    return AttributeSet();
-  uint16_t PackedID = IntrinsicsToAttributesMap[id - 1];
-  uint8_t FnAttrID = PackedID >> 8;
-  return getIntrinsicFnAttributeSet(C, FnAttrID);
-}
-
 Function *Intrinsic::getOrInsertDeclaration(Module *M, ID id,
                                             ArrayRef<Type *> Tys) {
   // There can never be multiple globals with the same name of different types,