Merged main:15d9d0fa8f55936625882a28759f0ec0033cb6de into amd-gfx:df9c746d9591

Local branch amd-gfx df9c746 Merged main:37daff028fcec27f2be1bb990df77e19c0244ccf into amd-gfx:2c4204e0a79d
Remote branch main 15d9d0f [VPlan] Also print final VPlan directly before codegen/execute. (llvm#82269)

Author: SC llvm team

clang/docs/LibASTMatchersReference.html

@@ -7049,7 +7049,7 @@ <h2 id="traversal-matchers">AST Traversal Matchers</h2>
 </pre></td></tr>
 
 
-<tr><td>Matcher&lt;<a href="https://clang.llvm.org/doxygen/classclang_1_1CXXFoldExpr.html">CXXFoldExpr</a>&gt;</td><td class="name" onclick="toggle('hasFoldInit0')"><a name="hasFoldInit0Anchor">hasFoldInit</a></td><td>ast_matchers::Matcher&lt;<a href="https://clang.llvm.org/doxygen/classclang_1_1Expr.html">Expr</a>&gt; InnerMacher</td></tr>
+<tr><td>Matcher&lt;<a href="https://clang.llvm.org/doxygen/classclang_1_1CXXFoldExpr.html">CXXFoldExpr</a>&gt;</td><td class="name" onclick="toggle('hasFoldInit0')"><a name="hasFoldInit0Anchor">hasFoldInit</a></td><td>Matcher&lt;<a href="https://clang.llvm.org/doxygen/classclang_1_1Expr.html">Expr</a>&gt; InnerMacher</td></tr>
 <tr><td colspan="4" class="doc" id="hasFoldInit0"><pre>Matches the operand that does not contain the parameter pack.
 
 Example matches `(0 + ... + args)` and `(args * ... * 1)`
@@ -7089,7 +7089,7 @@ <h2 id="traversal-matchers">AST Traversal Matchers</h2>
 </pre></td></tr>
 
 
-<tr><td>Matcher&lt;<a href="https://clang.llvm.org/doxygen/classclang_1_1CXXFoldExpr.html">CXXFoldExpr</a>&gt;</td><td class="name" onclick="toggle('hasPattern0')"><a name="hasPattern0Anchor">hasPattern</a></td><td>ast_matchers::Matcher&lt;<a href="https://clang.llvm.org/doxygen/classclang_1_1Expr.html">Expr</a>&gt; InnerMacher</td></tr>
+<tr><td>Matcher&lt;<a href="https://clang.llvm.org/doxygen/classclang_1_1CXXFoldExpr.html">CXXFoldExpr</a>&gt;</td><td class="name" onclick="toggle('hasPattern0')"><a name="hasPattern0Anchor">hasPattern</a></td><td>Matcher&lt;<a href="https://clang.llvm.org/doxygen/classclang_1_1Expr.html">Expr</a>&gt; InnerMacher</td></tr>
 <tr><td colspan="4" class="doc" id="hasPattern0"><pre>Matches the operand that contains the parameter pack.
 
 Example matches `(0 + ... + args)`
@@ -7859,7 +7859,7 @@ <h2 id="traversal-matchers">AST Traversal Matchers</h2>
 </pre></td></tr>
 
 
-<tr><td>Matcher&lt;<a href="https://clang.llvm.org/doxygen/classclang_1_1ClassTemplateSpecializationDecl.html">ClassTemplateSpecializationDecl</a>&gt;</td><td class="name" onclick="toggle('forEachTemplateArgument0')"><a name="forEachTemplateArgument0Anchor">forEachTemplateArgument</a></td><td>clang::ast_matchers::Matcher&lt;<a href="https://clang.llvm.org/doxygen/classclang_1_1TemplateArgument.html">TemplateArgument</a>&gt; InnerMatcher</td></tr>
+<tr><td>Matcher&lt;<a href="https://clang.llvm.org/doxygen/classclang_1_1ClassTemplateSpecializationDecl.html">ClassTemplateSpecializationDecl</a>&gt;</td><td class="name" onclick="toggle('forEachTemplateArgument0')"><a name="forEachTemplateArgument0Anchor">forEachTemplateArgument</a></td><td>Matcher&lt;<a href="https://clang.llvm.org/doxygen/classclang_1_1TemplateArgument.html">TemplateArgument</a>&gt; InnerMatcher</td></tr>
 <tr><td colspan="4" class="doc" id="forEachTemplateArgument0"><pre>Matches classTemplateSpecialization, templateSpecializationType and
 functionDecl nodes where the template argument matches the inner matcher.
 This matcher may produce multiple matches.
@@ -8454,7 +8454,7 @@ <h2 id="traversal-matchers">AST Traversal Matchers</h2>
 </pre></td></tr>
 
 
-<tr><td>Matcher&lt;<a href="https://clang.llvm.org/doxygen/classclang_1_1Expr.html">Expr</a>&gt;</td><td class="name" onclick="toggle('ignoringElidableConstructorCall0')"><a name="ignoringElidableConstructorCall0Anchor">ignoringElidableConstructorCall</a></td><td>ast_matchers::Matcher&lt;<a href="https://clang.llvm.org/doxygen/classclang_1_1Expr.html">Expr</a>&gt; InnerMatcher</td></tr>
+<tr><td>Matcher&lt;<a href="https://clang.llvm.org/doxygen/classclang_1_1Expr.html">Expr</a>&gt;</td><td class="name" onclick="toggle('ignoringElidableConstructorCall0')"><a name="ignoringElidableConstructorCall0Anchor">ignoringElidableConstructorCall</a></td><td>Matcher&lt;<a href="https://clang.llvm.org/doxygen/classclang_1_1Expr.html">Expr</a>&gt; InnerMatcher</td></tr>
 <tr><td colspan="4" class="doc" id="ignoringElidableConstructorCall0"><pre>Matches expressions that match InnerMatcher that are possibly wrapped in an
 elidable constructor and other corresponding bookkeeping nodes.
 
@@ -8691,7 +8691,7 @@ <h2 id="traversal-matchers">AST Traversal Matchers</h2>
 </pre></td></tr>
 
 
-<tr><td>Matcher&lt;<a href="https://clang.llvm.org/doxygen/classclang_1_1FunctionDecl.html">FunctionDecl</a>&gt;</td><td class="name" onclick="toggle('forEachTemplateArgument2')"><a name="forEachTemplateArgument2Anchor">forEachTemplateArgument</a></td><td>clang::ast_matchers::Matcher&lt;<a href="https://clang.llvm.org/doxygen/classclang_1_1TemplateArgument.html">TemplateArgument</a>&gt; InnerMatcher</td></tr>
+<tr><td>Matcher&lt;<a href="https://clang.llvm.org/doxygen/classclang_1_1FunctionDecl.html">FunctionDecl</a>&gt;</td><td class="name" onclick="toggle('forEachTemplateArgument2')"><a name="forEachTemplateArgument2Anchor">forEachTemplateArgument</a></td><td>Matcher&lt;<a href="https://clang.llvm.org/doxygen/classclang_1_1TemplateArgument.html">TemplateArgument</a>&gt; InnerMatcher</td></tr>
 <tr><td colspan="4" class="doc" id="forEachTemplateArgument2"><pre>Matches classTemplateSpecialization, templateSpecializationType and
 functionDecl nodes where the template argument matches the inner matcher.
 This matcher may produce multiple matches.
@@ -8959,7 +8959,7 @@ <h2 id="traversal-matchers">AST Traversal Matchers</h2>
 </pre></td></tr>
 
 
-<tr><td>Matcher&lt;<a href="https://clang.llvm.org/doxygen/classclang_1_1InitListExpr.html">InitListExpr</a>&gt;</td><td class="name" onclick="toggle('hasInit0')"><a name="hasInit0Anchor">hasInit</a></td><td>unsigned N, ast_matchers::Matcher&lt;<a href="https://clang.llvm.org/doxygen/classclang_1_1Expr.html">Expr</a>&gt; InnerMatcher</td></tr>
+<tr><td>Matcher&lt;<a href="https://clang.llvm.org/doxygen/classclang_1_1InitListExpr.html">InitListExpr</a>&gt;</td><td class="name" onclick="toggle('hasInit0')"><a name="hasInit0Anchor">hasInit</a></td><td>unsigned N, Matcher&lt;<a href="https://clang.llvm.org/doxygen/classclang_1_1Expr.html">Expr</a>&gt; InnerMatcher</td></tr>
 <tr><td colspan="4" class="doc" id="hasInit0"><pre>Matches the n'th item of an initializer list expression.
 
 Example matches y.
@@ -10026,7 +10026,7 @@ <h2 id="traversal-matchers">AST Traversal Matchers</h2>
 </pre></td></tr>
 
 
-<tr><td>Matcher&lt;<a href="https://clang.llvm.org/doxygen/classclang_1_1TemplateSpecializationType.html">TemplateSpecializationType</a>&gt;</td><td class="name" onclick="toggle('forEachTemplateArgument1')"><a name="forEachTemplateArgument1Anchor">forEachTemplateArgument</a></td><td>clang::ast_matchers::Matcher&lt;<a href="https://clang.llvm.org/doxygen/classclang_1_1TemplateArgument.html">TemplateArgument</a>&gt; InnerMatcher</td></tr>
+<tr><td>Matcher&lt;<a href="https://clang.llvm.org/doxygen/classclang_1_1TemplateSpecializationType.html">TemplateSpecializationType</a>&gt;</td><td class="name" onclick="toggle('forEachTemplateArgument1')"><a name="forEachTemplateArgument1Anchor">forEachTemplateArgument</a></td><td>Matcher&lt;<a href="https://clang.llvm.org/doxygen/classclang_1_1TemplateArgument.html">TemplateArgument</a>&gt; InnerMatcher</td></tr>
 <tr><td colspan="4" class="doc" id="forEachTemplateArgument1"><pre>Matches classTemplateSpecialization, templateSpecializationType and
 functionDecl nodes where the template argument matches the inner matcher.
 This matcher may produce multiple matches.

clang/docs/tools/dump_ast_matchers.py

@@ -116,6 +116,8 @@ def strip_doxygen(comment):
 
 def unify_arguments(args):
     """Gets rid of anything the user doesn't care about in the argument list."""
+    args = re.sub(r"clang::ast_matchers::internal::", r"", args)
+    args = re.sub(r"ast_matchers::internal::", r"", args)
     args = re.sub(r"internal::", r"", args)
     args = re.sub(r"extern const\s+(.*)&", r"\1 ", args)
     args = re.sub(r"&", r" ", args)

clang/include/clang/ASTMatchers/ASTMatchers.h

@@ -4580,8 +4580,7 @@ AST_POLYMORPHIC_MATCHER_P2(hasArgument,
 ///       return (args * ... * 1);
 ///   }
 /// \endcode
-AST_MATCHER_P(CXXFoldExpr, hasFoldInit, ast_matchers::internal::Matcher<Expr>,
-              InnerMacher) {
+AST_MATCHER_P(CXXFoldExpr, hasFoldInit, internal::Matcher<Expr>, InnerMacher) {
   const auto *const Init = Node.getInit();
   return Init && InnerMacher.matches(*Init, Finder, Builder);
 }
@@ -4603,8 +4602,7 @@ AST_MATCHER_P(CXXFoldExpr, hasFoldInit, ast_matchers::internal::Matcher<Expr>,
 ///       return (args * ... * 1);
 ///   }
 /// \endcode
-AST_MATCHER_P(CXXFoldExpr, hasPattern, ast_matchers::internal::Matcher<Expr>,
-              InnerMacher) {
+AST_MATCHER_P(CXXFoldExpr, hasPattern, internal::Matcher<Expr>, InnerMacher) {
   const Expr *const Pattern = Node.getPattern();
   return Pattern && InnerMacher.matches(*Pattern, Finder, Builder);
 }
@@ -4685,8 +4683,8 @@ AST_MATCHER(CXXFoldExpr, isBinaryFold) { return Node.getInit() != nullptr; }
 /// \code
 ///   int x{y}.
 /// \endcode
-AST_MATCHER_P2(InitListExpr, hasInit, unsigned, N,
-               ast_matchers::internal::Matcher<Expr>, InnerMatcher) {
+AST_MATCHER_P2(InitListExpr, hasInit, unsigned, N, internal::Matcher<Expr>,
+               InnerMatcher) {
   return N < Node.getNumInits() &&
           InnerMatcher.matches(*Node.getInit(N), Finder, Builder);
 }
@@ -5309,7 +5307,7 @@ AST_POLYMORPHIC_MATCHER_P(
     forEachTemplateArgument,
     AST_POLYMORPHIC_SUPPORTED_TYPES(ClassTemplateSpecializationDecl,
                                     TemplateSpecializationType, FunctionDecl),
-    clang::ast_matchers::internal::Matcher<TemplateArgument>, InnerMatcher) {
+    internal::Matcher<TemplateArgument>, InnerMatcher) {
   ArrayRef<TemplateArgument> TemplateArgs =
       clang::ast_matchers::internal::getTemplateSpecializationArgs(Node);
   clang::ast_matchers::internal::BoundNodesTreeBuilder Result;
@@ -8525,8 +8523,8 @@ AST_MATCHER(FunctionDecl, hasTrailingReturn) {
 ///
 /// ``varDecl(hasInitializer(ignoringElidableConstructorCall(callExpr())))``
 /// matches ``H D = G()`` in C++11 through C++17 (and beyond).
-AST_MATCHER_P(Expr, ignoringElidableConstructorCall,
-              ast_matchers::internal::Matcher<Expr>, InnerMatcher) {
+AST_MATCHER_P(Expr, ignoringElidableConstructorCall, internal::Matcher<Expr>,
+              InnerMatcher) {
   // E tracks the node that we are examining.
   const Expr *E = &Node;
   // If present, remove an outer `ExprWithCleanups` corresponding to the

clang/lib/CodeGen/CGCall.cpp

@@ -3221,12 +3221,11 @@ void CodeGenFunction::EmitFunctionProlog(const CGFunctionInfo &FI,
 
       llvm::StructType *STy =
           dyn_cast<llvm::StructType>(ArgI.getCoerceToType());
-      llvm::TypeSize StructSize;
-      llvm::TypeSize PtrElementSize;
       if (ArgI.isDirect() && !ArgI.getCanBeFlattened() && STy &&
           STy->getNumElements() > 1) {
-        StructSize = CGM.getDataLayout().getTypeAllocSize(STy);
-        PtrElementSize =
+        [[maybe_unused]] llvm::TypeSize StructSize =
+            CGM.getDataLayout().getTypeAllocSize(STy);
+        [[maybe_unused]] llvm::TypeSize PtrElementSize =
             CGM.getDataLayout().getTypeAllocSize(ConvertTypeForMem(Ty));
         if (STy->containsHomogeneousScalableVectorTypes()) {
           assert(StructSize == PtrElementSize &&
@@ -5310,12 +5309,12 @@ RValue CodeGenFunction::EmitCall(const CGFunctionInfo &CallInfo,
 
       llvm::StructType *STy =
           dyn_cast<llvm::StructType>(ArgInfo.getCoerceToType());
-      llvm::Type *SrcTy = ConvertTypeForMem(I->Ty);
-      llvm::TypeSize SrcTypeSize;
-      llvm::TypeSize DstTypeSize;
       if (STy && ArgInfo.isDirect() && !ArgInfo.getCanBeFlattened()) {
-        SrcTypeSize = CGM.getDataLayout().getTypeAllocSize(SrcTy);
-        DstTypeSize = CGM.getDataLayout().getTypeAllocSize(STy);
+        llvm::Type *SrcTy = ConvertTypeForMem(I->Ty);
+        [[maybe_unused]] llvm::TypeSize SrcTypeSize =
+            CGM.getDataLayout().getTypeAllocSize(SrcTy);
+        [[maybe_unused]] llvm::TypeSize DstTypeSize =
+            CGM.getDataLayout().getTypeAllocSize(STy);
         if (STy->containsHomogeneousScalableVectorTypes()) {
           assert(SrcTypeSize == DstTypeSize &&
                  "Only allow non-fractional movement of structure with "

flang/include/flang/Lower/AbstractConverter.h

@@ -16,7 +16,6 @@
 #include "flang/Common/Fortran.h"
 #include "flang/Lower/LoweringOptions.h"
 #include "flang/Lower/PFTDefs.h"
-#include "flang/Lower/SymbolMap.h"
 #include "flang/Optimizer/Builder/BoxValue.h"
 #include "flang/Semantics/symbol.h"
 #include "mlir/IR/Builders.h"
@@ -54,6 +53,7 @@ class DerivedTypeSpec;
 
 namespace lower {
 class SymMap;
+class SymbolBox;
 namespace pft {
 struct Variable;
 }

flang/lib/Lower/OpenMP/DataSharingProcessor.cpp

@@ -14,6 +14,7 @@
 
 #include "Utils.h"
 #include "flang/Lower/PFTBuilder.h"
+#include "flang/Lower/SymbolMap.h"
 #include "flang/Optimizer/Builder/Todo.h"
 #include "flang/Semantics/tools.h"
 #include "mlir/Dialect/OpenMP/OpenMPDialect.h"

flang/lib/Optimizer/HLFIR/Transforms/LowerHLFIRIntrinsics.cpp

@@ -176,7 +176,14 @@ class HlfirIntrinsicConversion : public mlir::OpRewritePattern<OP> {
           rewriter.eraseOp(use);
       }
     }
-    rewriter.replaceAllUsesWith(op->getResults(), {base});
+    // TODO: This entire pass should be a greedy pattern rewrite or a manual
+    // IR traversal. A dialect conversion cannot be used here because
+    // `replaceAllUsesWith` is not supported. Similarly, `replaceOp` is not
+    // suitable because "op->getResult(0)" and "base" can have different types.
+    // In such a case, the dialect conversion will attempt to convert the type,
+    // but no type converter is specified in this pass. Also note that all
+    // patterns in this pass are actually rewrite patterns.
+    op->getResult(0).replaceAllUsesWith(base);
     rewriter.replaceOp(op, base);
   }
 };

llvm/docs/CommandGuide/llvm-objcopy.rst

@@ -455,6 +455,15 @@ them.
  Set the start address of the output to ``<addr>``. Overrides any previously
  specified :option:`--change-start` or :option:`--adjust-start` options.
 
+.. option:: --set-symbol-visibility <symbol>=<visibility>
+
+ Change the visibility of a symbol to the specified value.
+
+.. option:: --set-symbols-visibility <filename>=<visibility>
+
+ Read a list of symbols from <filename> and change their visibility to the
+ specified value. Visibility values: default, internal, hidden, protected.
+
 .. option:: --split-dwo <dwo-file>
 
  Equivalent to running :program:`llvm-objcopy` with :option:`--extract-dwo` and

llvm/docs/ReleaseNotes.rst

@@ -147,6 +147,10 @@ Changes to the LLVM tools
   if it's not specified with the ``--format`` argument and cannot be inferred from
   input files.
 
+* llvm-objcopy now supports ``--set-symbol-visibility`` and
+  ``--set-symbols-visibility`` options for ELF input to change the
+  visibility of symbols.
+
 Changes to LLDB
 ---------------------------------
 

llvm/include/llvm/Config/llvm-config.h.cmake

@@ -16,7 +16,7 @@
 
 /* Indicate that this is LLVM compiled from the amd-gfx branch. */
 #define LLVM_HAVE_BRANCH_AMD_GFX
-#define LLVM_MAIN_REVISION 490988
+#define LLVM_MAIN_REVISION 491000
 
 /* Define if LLVM_ENABLE_DUMP is enabled */
 #cmakedefine LLVM_ENABLE_DUMP

llvm/include/llvm/ObjCopy/ELF/ELFConfig.h

@@ -9,6 +9,7 @@
 #ifndef LLVM_OBJCOPY_ELF_ELFCONFIG_H
 #define LLVM_OBJCOPY_ELF_ELFCONFIG_H
 
+#include "llvm/ObjCopy/CommonConfig.h"
 #include "llvm/Object/ELFTypes.h"
 
 namespace llvm {
@@ -18,6 +19,8 @@ namespace objcopy {
 struct ELFConfig {
   uint8_t NewSymbolVisibility = (uint8_t)ELF::STV_DEFAULT;
 
+  std::vector<std::pair<NameMatcher, uint8_t>> SymbolsToSetVisibility;
+
   // ELF entry point address expression. The input parameter is an entry point
   // address in the input ELF file. The entry address in the output file is
   // calculated with EntryExpr(input_address), when either --set-start or

llvm/include/llvm/Support/TypeSize.h

@@ -321,8 +321,6 @@ class TypeSize : public details::FixedOrScalableQuantity<TypeSize, uint64_t> {
       : FixedOrScalableQuantity(V) {}
 
 public:
-  constexpr TypeSize() : FixedOrScalableQuantity(0, false) {}
-
   constexpr TypeSize(ScalarTy Quantity, bool Scalable)
       : FixedOrScalableQuantity(Quantity, Scalable) {}
 

llvm/lib/ObjCopy/ELF/ELFObjcopy.cpp

@@ -300,6 +300,10 @@ static Error updateAndRemoveSymbols(const CommonConfig &Config,
          Config.SymbolsToLocalize.matches(Sym.Name)))
       Sym.Binding = STB_LOCAL;
 
+    for (auto &[Matcher, Visibility] : ELFConfig.SymbolsToSetVisibility)
+      if (Matcher.matches(Sym.Name))
+        Sym.Visibility = Visibility;
+
     // Note: these two globalize flags have very similar names but different
     // meanings:
     //

llvm/lib/Target/X86/X86FrameLowering.cpp

@@ -1418,6 +1418,34 @@ bool X86FrameLowering::needsDwarfCFI(const MachineFunction &MF) const {
   return !isWin64Prologue(MF) && MF.needsFrameMoves();
 }
 
+/// Return true if an opcode is part of the REP group of instructions
+static bool isOpcodeRep(unsigned Opcode) {
+  switch (Opcode) {
+  case X86::REPNE_PREFIX:
+  case X86::REP_MOVSB_32:
+  case X86::REP_MOVSB_64:
+  case X86::REP_MOVSD_32:
+  case X86::REP_MOVSD_64:
+  case X86::REP_MOVSQ_32:
+  case X86::REP_MOVSQ_64:
+  case X86::REP_MOVSW_32:
+  case X86::REP_MOVSW_64:
+  case X86::REP_PREFIX:
+  case X86::REP_STOSB_32:
+  case X86::REP_STOSB_64:
+  case X86::REP_STOSD_32:
+  case X86::REP_STOSD_64:
+  case X86::REP_STOSQ_32:
+  case X86::REP_STOSQ_64:
+  case X86::REP_STOSW_32:
+  case X86::REP_STOSW_64:
+    return true;
+  default:
+    break;
+  }
+  return false;
+}
+
 /// emitPrologue - Push callee-saved registers onto the stack, which
 /// automatically adjust the stack pointer. Adjust the stack pointer to allocate
 /// space for local variables. Also emit labels used by the exception handler to
@@ -2194,13 +2222,44 @@ void X86FrameLowering::emitPrologue(MachineFunction &MF,
   // flag (DF in EFLAGS register). Clear this flag by creating "cld" instruction
   // in each prologue of interrupt handler function.
   //
-  // FIXME: Create "cld" instruction only in these cases:
+  // Create "cld" instruction only in these cases:
   // 1. The interrupt handling function uses any of the "rep" instructions.
   // 2. Interrupt handling function calls another function.
+  // 3. If there are any inline asm blocks, as we do not know what they do
   //
-  if (Fn.getCallingConv() == CallingConv::X86_INTR)
-    BuildMI(MBB, MBBI, DL, TII.get(X86::CLD))
-        .setMIFlag(MachineInstr::FrameSetup);
+  // TODO: We should also emit cld if we detect the use of std, but as of now,
+  // the compiler does not even emit that instruction or even define it, so in
+  // practice, this would only happen with inline asm, which we cover anyway.
+  if (Fn.getCallingConv() == CallingConv::X86_INTR) {
+    bool NeedsCLD = false;
+
+    for (const MachineBasicBlock &B : MF) {
+      for (const MachineInstr &MI : B) {
+        if (MI.isCall()) {
+          NeedsCLD = true;
+          break;
+        }
+
+        if (isOpcodeRep(MI.getOpcode())) {
+          NeedsCLD = true;
+          break;
+        }
+
+        if (MI.isInlineAsm()) {
+          // TODO: Parse asm for rep instructions or call sites?
+          // For now, let's play it safe and emit a cld instruction
+          // just in case.
+          NeedsCLD = true;
+          break;
+        }
+      }
+    }
+
+    if (NeedsCLD) {
+      BuildMI(MBB, MBBI, DL, TII.get(X86::CLD))
+          .setMIFlag(MachineInstr::FrameSetup);
+    }
+  }
 
   // At this point we know if the function has WinCFI or not.
   MF.setHasWinCFI(HasWinCFI);