Merge pull request #179 from sx-aurora-dev/feature/merge-upstream-20220329

Feature/merge upstream 20220329

Author: kaz7

bolt/include/bolt/Core/BinaryFunction.h

@@ -172,6 +172,9 @@ class BinaryFunction {
 
     mutable MCSymbol *FunctionConstantIslandLabel{nullptr};
     mutable MCSymbol *FunctionColdConstantIslandLabel{nullptr};
+
+    // Returns constant island alignment
+    uint16_t getAlignment() const { return sizeof(uint64_t); }
   };
 
   static constexpr uint64_t COUNT_NO_PROFILE =
@@ -2047,6 +2050,10 @@ class BinaryFunction {
     return *std::prev(CodeIter) <= *DataIter;
   }
 
+  uint16_t getConstantIslandAlignment() const {
+    return Islands ? Islands->getAlignment() : 1;
+  }
+
   uint64_t
   estimateConstantIslandSize(const BinaryFunction *OnBehalfOf = nullptr) const {
     if (!Islands)
@@ -2074,9 +2081,13 @@ class BinaryFunction {
       Size += NextMarker - *DataIter;
     }
 
-    if (!OnBehalfOf)
-      for (BinaryFunction *ExternalFunc : Islands->Dependency)
+    if (!OnBehalfOf) {
+      for (BinaryFunction *ExternalFunc : Islands->Dependency) {
+        Size = alignTo(Size, ExternalFunc->getConstantIslandAlignment());
         Size += ExternalFunc->estimateConstantIslandSize(this);
+      }
+    }
+
     return Size;
   }
 

bolt/lib/Core/BinaryEmitter.cpp

@@ -500,6 +500,13 @@ void BinaryEmitter::emitConstantIslands(BinaryFunction &BF, bool EmitColdPart,
   if (Islands.DataOffsets.empty() && Islands.Dependency.empty())
     return;
 
+  // AArch64 requires CI to be aligned to 8 bytes due to access instructions
+  // restrictions. E.g. the ldr with imm, where imm must be aligned to 8 bytes.
+  const uint16_t Alignment = OnBehalfOf
+                                 ? OnBehalfOf->getConstantIslandAlignment()
+                                 : BF.getConstantIslandAlignment();
+  Streamer.emitCodeAlignment(Alignment, &*BC.STI);
+
   if (!OnBehalfOf) {
     if (!EmitColdPart)
       Streamer.emitLabel(BF.getFunctionConstantIslandLabel());

bolt/lib/Passes/LongJmp.cpp

@@ -308,6 +308,7 @@ uint64_t LongJmpPass::tentativeLayoutRelocColdPart(
     LLVM_DEBUG(dbgs() << Func->getPrintName() << " cold tentative: "
                       << Twine::utohexstr(DotAddress) << "\n");
     DotAddress += Func->estimateColdSize();
+    DotAddress = alignTo(DotAddress, Func->getConstantIslandAlignment());
     DotAddress += Func->estimateConstantIslandSize();
   }
   return DotAddress;
@@ -364,6 +365,8 @@ uint64_t LongJmpPass::tentativeLayoutRelocMode(
       DotAddress += Func->estimateSize();
     else
       DotAddress += Func->estimateHotSize();
+
+    DotAddress = alignTo(DotAddress, Func->getConstantIslandAlignment());
     DotAddress += Func->estimateConstantIslandSize();
     ++CurrentIndex;
   }

bolt/test/AArch64/constant-island-alignment.s

@@ -0,0 +1,38 @@
+// This test checks that the constant island is aligned after BOLT tool.
+// In case the nop before .Lci will be removed the pointer to exit function
+// won't be alinged and the test will fail.
+
+# RUN: llvm-mc -filetype=obj -triple aarch64-unknown-unknown \
+# RUN:   %s -o %t.o
+# RUN: %clang %cflags -fPIC -pie %t.o -o %t.exe -Wl,-q \
+# RUN:    -nostartfiles -nodefaultlibs -lc
+# RUN: llvm-bolt %t.exe -o %t.bolt -use-old-text=0 -lite=0 -trap-old-code
+# RUN: llvm-objdump -d --disassemble-symbols='$d' %t.bolt | FileCheck %s
+
+.text
+.align 4
+.global
+.type dummy, %function
+dummy:
+  add x0, x0, #1
+  ret
+
+.global
+.type exitOk, %function
+exitOk:
+  mov x0, #0
+  bl exit
+
+.global _start
+.type _start, %function
+_start:
+  adrp x0, .Lci
+  ldr x0, [x0, #:lo12:.Lci]
+  blr x0
+  mov x1, #1
+  bl exit
+  nop
+# CHECK: {{0|8}} <$d>:
+.Lci:
+  .xword exitOk
+  .xword 0

bolt/test/runtime/AArch64/adrrelaxationpass.s

@@ -41,9 +41,9 @@ br:
   .word 0xff
 
 # CHECK: <main>:
-# CHECK-NEXT: adr x0, #28
+# CHECK-NEXT: adr x0, #{{[0-9][0-9]*}}
 # CHECK-NEXT: adrp x1, 0x{{[1-8a-f][0-9a-f]*}}
 # CHECK-NEXT: add x1, x1, #{{[1-8a-f][0-9a-f]*}}
 # CHECK-NEXT: adrp x2, 0x{{[1-8a-f][0-9a-f]*}}
 # CHECK-NEXT: add x2, x2, #{{[1-8a-f][0-9a-f]*}}
-# CHECK-NEXT: adr x3, #4
+# CHECK-NEXT: adr x3, #{{[0-9][0-9]*}}

clang-tools-extra/clangd/test/CMakeLists.txt

@@ -1,13 +1,3 @@
-# Set CLANG_TOOLS_DIR to buildtree/bin, or buildtree/%(build_mode)s/bin if the
-# location is dynamic. The latter must be interpolated by lit configs.
-# FIXME: this is duplicated in many places.
-if (CMAKE_CFG_INTDIR STREQUAL ".")
-  set(LLVM_BUILD_MODE ".")
-else ()
-  set(LLVM_BUILD_MODE "%(build_mode)s")
-endif ()
-string(REPLACE ${CMAKE_CFG_INTDIR} ${LLVM_BUILD_MODE} CLANG_TOOLS_DIR ${LLVM_RUNTIME_OUTPUT_INTDIR})
-
 set(CLANGD_TEST_DEPS
   clangd
   ClangdTests

clang-tools-extra/clangd/test/lit.site.cfg.py.in

@@ -7,7 +7,7 @@ config.host_triple = "@LLVM_HOST_TRIPLE@"
 config.python_executable = "@Python3_EXECUTABLE@"
 # Support substitution of the tools and libs dirs with user parameters. This is
 # used when we can't determine the tool dir at configuration time.
-config.clang_tools_dir = lit_config.substitute("@CLANG_TOOLS_DIR@")
+config.clang_tools_dir = lit_config.substitute("@CURRENT_TOOLS_DIR@")
 config.llvm_tools_dir = lit_config.substitute("@LLVM_TOOLS_DIR@")
 config.llvm_libs_dir = lit_config.substitute("@LLVM_LIBS_DIR@")
 

clang-tools-extra/docs/ReleaseNotes.rst

@@ -96,6 +96,9 @@ The improvements are...
 Improvements to clang-tidy
 --------------------------
 
+- Added trace code to help narrow down any checks and the relevant source code
+  that result in crashes.
+
 New checks
 ^^^^^^^^^^
 

clang-tools-extra/pseudo/include/clang-pseudo/Forest.h

@@ -17,6 +17,9 @@
 //
 //===----------------------------------------------------------------------===//
 
+#ifndef CLANG_PSEUDO_FOREST_H
+#define CLANG_PSEUDO_FOREST_H
+
 #include "clang-pseudo/Grammar.h"
 #include "clang-pseudo/Token.h"
 #include "llvm/ADT/ArrayRef.h"
@@ -176,3 +179,5 @@ class ForestArena {
 
 } // namespace pseudo
 } // namespace clang
+
+#endif // CLANG_PSEUDO_FOREST_H

clang-tools-extra/pseudo/lib/Forest.cpp

@@ -46,15 +46,23 @@ std::string ForestNode::dumpRecursive(const Grammar &G,
       };
   CountVisits(this);
 
+  // The box-drawing characters that should be added as a child is rendered.
+  struct LineDecoration {
+    std::string Prefix;         // Prepended to every line.
+    llvm::StringRef First;      // added to the child's line.
+    llvm::StringRef Subsequent; // added to descendants' lines.
+  };
+
   // We print a "#<id>" for nonterminal forest nodes that are being dumped
   // multiple times.
   llvm::DenseMap<const ForestNode *, size_t> ReferenceIds;
   std::string Result;
   constexpr Token::Index KEnd = std::numeric_limits<Token::Index>::max();
-  std::function<void(const ForestNode *, unsigned, Token::Index,
-                     llvm::Optional<SymbolID>)>
-      Dump = [&](const ForestNode *P, unsigned Level, Token::Index End,
-                 llvm::Optional<SymbolID> ElidedParent) {
+  std::function<void(const ForestNode *, Token::Index, llvm::Optional<SymbolID>,
+                     LineDecoration &LineDec)>
+      Dump = [&](const ForestNode *P, Token::Index End,
+                 llvm::Optional<SymbolID> ElidedParent,
+                 LineDecoration LineDec) {
         llvm::ArrayRef<const ForestNode *> Children;
         auto EndOfElement = [&](size_t ChildIndex) {
           return ChildIndex + 1 == Children.size()
@@ -72,18 +80,19 @@ std::string ForestNode::dumpRecursive(const Grammar &G,
               if (Children[I]->startTokenIndex() == P->startTokenIndex() &&
                   EndOfElement(I) == End) {
                 return Dump(
-                    Children[I], Level, End,
-                    /*ElidedParent=*/ElidedParent.getValueOr(P->symbol()));
+                    Children[I], End,
+                    /*ElidedParent=*/ElidedParent.getValueOr(P->symbol()),
+                    LineDec);
               }
           }
         }
 
-        // FIXME: pretty ascii trees
         if (End == KEnd)
           Result += llvm::formatv("[{0,3}, end) ", P->startTokenIndex());
         else
           Result += llvm::formatv("[{0,3}, {1,3}) ", P->startTokenIndex(), End);
-        Result.append(2 * Level, ' ');
+        Result += LineDec.Prefix;
+        Result += LineDec.First;
         if (ElidedParent.hasValue()) {
           Result += G.symbolName(*ElidedParent);
           Result += "~";
@@ -99,12 +108,23 @@ std::string ForestNode::dumpRecursive(const Grammar &G,
         }
         Result.push_back('\n');
 
-        ++Level;
-        for (size_t I = 0; I < Children.size(); ++I)
-          Dump(Children[I], Level,
-               P->kind() == Sequence ? EndOfElement(I) : End, llvm::None);
+        auto OldPrefixSize = LineDec.Prefix.size();
+        LineDec.Prefix += LineDec.Subsequent;
+        for (size_t I = 0; I < Children.size(); ++I) {
+          if (I == Children.size() - 1) {
+            LineDec.First = "└─";
+            LineDec.Subsequent = "  ";
+          } else {
+            LineDec.First = "├─";
+            LineDec.Subsequent = "│ ";
+          }
+          Dump(Children[I], P->kind() == Sequence ? EndOfElement(I) : End,
+               llvm::None, LineDec);
+        }
+        LineDec.Prefix.resize(OldPrefixSize);
       };
-  Dump(this, 0, KEnd, llvm::None);
+  LineDecoration LineDec;
+  Dump(this, KEnd, llvm::None, LineDec);
   return Result;
 }