am 31675153: Merge branch \'upstream\' into merge_2

* commit '31675153bd2d7617db8cb6aeb58054934c7b9f73': (542 commits)
  MaximumSpanningTree::EdgeWeightCompare: Make this comparator actually be a strict weak ordering, and don't pass possibly-null pointers to dyn_cast.
  Fix misaligned access in MachO object file reader: despite containing an int64_t, Symbol64TableEntry is actually only stored with 4-byte alignment within the file.
  Fix unaligned memory accesses when performing relocations in X86 JIT. There's no cost to using memcpy here: the fixed code is optimized by LLVM to perfect machine code.
  Don't pass a null pointer to cast<> in its unit tests.
  Don't bind a reference to a dereferenced null pointer (for return value of WeakVH::operator*).
  [ms-inline asm] Do not report a Parser error when matching inline assembly.
  Ignore the documentation-suggested location for compile_commands.json
  The presence of the empty file "foo" unfortunately does not improve LLVM in any way.
  Remove unnecessary cast that was also unnecessarily casting away constness.
  Provide a portability macro for __builtin_trap.
  Fix macros arguments with an underscore, dot or dollar in them. This is based on a patch by Andy/PaX. I added the support for dot and dollar.
  [ms-inline asm] Expose the ErrorInfo from the MatchInstructionImpl.  In general, this is the index of the operand that failed to match.
  Formatting.  No functional change.
  Make the wording in of the "expected identifier" error in the .macro directive consistent with the other "expected identifier" errors. Extracted from the Andy/PaX patch. I added the test.
  Pacify PVS-Studio by changing the type rather than doing a cast, a tweak suggested by David Blaikie.
  Add support for the --param ssp-buffer-size= driver option. PR9673
  Use typedefs. Fix indentation. Extracted from the Andy/PaX patch.
  Remove unused variable. Extracted from the Andy/PaX patch.
  Fix typo. Extracted from the Andy/PaX patch.
  MCJIT: Tidy up the constructor.
  ...

1 files changed, 115 insertions, 122 deletions

diff --git a/utils/TableGen/X86DisassemblerTables.cpp b/utils/TableGen/X86DisassemblerTables.cpp
index 2875168..f3bd373 100644
--- a/utils/TableGen/X86DisassemblerTables.cpp
+++ b/utils/TableGen/X86DisassemblerTables.cpp
@@ -21,10 +21,11 @@
 #include "llvm/ADT/STLExtras.h"
 #include "llvm/Support/ErrorHandling.h"
 #include "llvm/Support/Format.h"
+#include <map>
 
 using namespace llvm;
 using namespace X86Disassembler;
-  
+
 /// inheritsFrom - Indicates whether all instructions in one class also belong
 ///   to another class.
 ///
@@ -36,7 +37,7 @@ static inline bool inheritsFrom(InstructionContext child,
                                 bool VEX_LIG = false) {
   if (child == parent)
     return true;
-  
+
   switch (parent) {
   case IC:
     return(inheritsFrom(child, IC_64BIT) ||
@@ -117,17 +118,17 @@ static inline bool inheritsFrom(InstructionContext child,
 /// @param upper  - The class that may be preferable
 /// @param lower  - The class that may be less preferable
 /// @return       - True if upper is to be preferred, false otherwise.
-static inline bool outranks(InstructionContext upper, 
+static inline bool outranks(InstructionContext upper,
                             InstructionContext lower) {
   assert(upper < IC_max);
   assert(lower < IC_max);
-  
+
 #define ENUM_ENTRY(n, r, d) r,
   static int ranks[IC_max] = {
     INSTRUCTION_CONTEXTS
   };
 #undef ENUM_ENTRY
-  
+
   return (ranks[upper] > ranks[lower]);
 }
 
@@ -170,24 +171,22 @@ static inline const char* stringForOperandEncoding(OperandEncoding encoding) {
   }
 }
 
-void DisassemblerTables::emitOneID(raw_ostream &o,
-                                   uint32_t &i,
-                                   InstrUID id,
+void DisassemblerTables::emitOneID(raw_ostream &o, unsigned &i, InstrUID id,
                                    bool addComma) const {
   if (id)
     o.indent(i * 2) << format("0x%hx", id);
   else
     o.indent(i * 2) << 0;
-  
+
   if (addComma)
     o << ", ";
   else
     o << "  ";
-  
+
   o << "/* ";
   o << InstructionSpecifiers[id].name;
   o << "*/";
-  
+
   o << "\n";
 }
 
@@ -197,8 +196,7 @@ void DisassemblerTables::emitOneID(raw_ostream &o,
 ///
 /// @param o        - The output stream on which to emit the table.
 /// @param i        - The indentation level for that output stream.
-static void emitEmptyTable(raw_ostream &o, uint32_t &i)
-{
+static void emitEmptyTable(raw_ostream &o, unsigned &i) {
   o.indent(i * 2) << "0x0, /* EmptyTable */\n";
 }
 
@@ -207,15 +205,12 @@ static void emitEmptyTable(raw_ostream &o, uint32_t &i)
 ///
 /// @param decision - The decision to be compacted.
 /// @return         - The compactest available representation for the decision.
-static ModRMDecisionType getDecisionType(ModRMDecision &decision)
-{
+static ModRMDecisionType getDecisionType(ModRMDecision &decision) {
   bool satisfiesOneEntry = true;
   bool satisfiesSplitRM = true;
   bool satisfiesSplitReg = true;
 
-  uint16_t index;
-
-  for (index = 0; index < 256; ++index) {
+  for (unsigned index = 0; index < 256; ++index) {
     if (decision.instructionIDs[index] != decision.instructionIDs[0])
       satisfiesOneEntry = false;
 
@@ -252,27 +247,25 @@ static ModRMDecisionType getDecisionType(ModRMDecision &decision)
 ///   to a particular decision type.
 ///
 /// @param dt - The decision type.
-/// @return   - A pointer to the statically-allocated string (e.g., 
+/// @return   - A pointer to the statically-allocated string (e.g.,
 ///             "MODRM_ONEENTRY" for MODRM_ONEENTRY).
-static const char* stringForDecisionType(ModRMDecisionType dt)
-{
+static const char* stringForDecisionType(ModRMDecisionType dt) {
 #define ENUM_ENTRY(n) case n: return #n;
   switch (dt) {
     default:
-      llvm_unreachable("Unknown decision type");  
+      llvm_unreachable("Unknown decision type");
     MODRMTYPES
-  };  
+  };
 #undef ENUM_ENTRY
 }
-  
+
 /// stringForModifierType - Returns a statically-allocated string corresponding
 ///   to an opcode modifier type.
 ///
 /// @param mt - The modifier type.
 /// @return   - A pointer to the statically-allocated string (e.g.,
 ///             "MODIFIER_NONE" for MODIFIER_NONE).
-static const char* stringForModifierType(ModifierType mt)
-{
+static const char* stringForModifierType(ModifierType mt) {
 #define ENUM_ENTRY(n) case n: return #n;
   switch(mt) {
     default:
@@ -281,35 +274,31 @@ static const char* stringForModifierType(ModifierType mt)
   };
 #undef ENUM_ENTRY
 }
-  
+
 DisassemblerTables::DisassemblerTables() {
   unsigned i;
-  
+
   for (i = 0; i < array_lengthof(Tables); i++) {
     Tables[i] = new ContextDecision;
     memset(Tables[i], 0, sizeof(ContextDecision));
   }
-  
+
   HasConflicts = false;
 }
-  
+
 DisassemblerTables::~DisassemblerTables() {
   unsigned i;
-  
+
   for (i = 0; i < array_lengthof(Tables); i++)
     delete Tables[i];
 }
-  
-void DisassemblerTables::emitModRMDecision(raw_ostream &o1,
-                                           raw_ostream &o2,
-                                           uint32_t &i1,
-                                           uint32_t &i2,
-                                           ModRMDecision &decision)
-  const {
-  static uint64_t sTableNumber = 0;
-  static uint64_t sEntryNumber = 1;
+
+void DisassemblerTables::emitModRMDecision(raw_ostream &o1, raw_ostream &o2,
+                                           unsigned &i1, unsigned &i2,
+                                           ModRMDecision &decision) const {
+  static uint32_t sTableNumber = 0;
+  static uint32_t sEntryNumber = 1;
   ModRMDecisionType dt = getDecisionType(decision);
-  uint16_t index;
 
   if (dt == MODRM_ONEENTRY && decision.instructionIDs[0] == 0)
   {
@@ -338,13 +327,13 @@ void DisassemblerTables::emitModRMDecision(raw_ostream &o1,
       emitOneID(o1, i1, decision.instructionIDs[0xc0], true); // mod = 0b11
       break;
     case MODRM_SPLITREG:
-      for (index = 0; index < 64; index += 8)
+      for (unsigned index = 0; index < 64; index += 8)
         emitOneID(o1, i1, decision.instructionIDs[index], true);
-      for (index = 0xc0; index < 256; index += 8)
+      for (unsigned index = 0xc0; index < 256; index += 8)
         emitOneID(o1, i1, decision.instructionIDs[index], true);
       break;
     case MODRM_FULL:
-      for (index = 0; index < 256; ++index)
+      for (unsigned index = 0; index < 256; ++index)
         emitOneID(o1, i1, decision.instructionIDs[index], true);
       break;
   }
@@ -380,20 +369,15 @@ void DisassemblerTables::emitModRMDecision(raw_ostream &o1,
   ++sTableNumber;
 }
 
-void DisassemblerTables::emitOpcodeDecision(
-  raw_ostream &o1,
-  raw_ostream &o2,
-  uint32_t &i1,
-  uint32_t &i2,
-  OpcodeDecision &decision) const {
-  uint16_t index;
-
+void DisassemblerTables::emitOpcodeDecision(raw_ostream &o1, raw_ostream &o2,
+                                            unsigned &i1, unsigned &i2,
+                                            OpcodeDecision &decision) const {
   o2.indent(i2) << "{ /* struct OpcodeDecision */" << "\n";
   i2++;
   o2.indent(i2) << "{" << "\n";
   i2++;
 
-  for (index = 0; index < 256; ++index) {
+  for (unsigned index = 0; index < 256; ++index) {
     o2.indent(i2);
 
     o2 << "/* 0x" << format("%02hhx", index) << " */" << "\n";
@@ -412,21 +396,16 @@ void DisassemblerTables::emitOpcodeDecision(
   o2.indent(i2) << "}" << "\n";
 }
 
-void DisassemblerTables::emitContextDecision(
-  raw_ostream &o1,
-  raw_ostream &o2,
-  uint32_t &i1,
-  uint32_t &i2,
-  ContextDecision &decision,
-  const char* name) const {
+void DisassemblerTables::emitContextDecision(raw_ostream &o1, raw_ostream &o2,
+                                             unsigned &i1, unsigned &i2,
+                                             ContextDecision &decision,
+                                             const char* name) const {
   o2.indent(i2) << "static const struct ContextDecision " << name << " = {\n";
   i2++;
   o2.indent(i2) << "{ /* opcodeDecisions */" << "\n";
   i2++;
 
-  unsigned index;
-
-  for (index = 0; index < IC_max; ++index) {
+  for (unsigned index = 0; index < IC_max; ++index) {
     o2.indent(i2) << "/* ";
     o2 << stringForContext((InstructionContext)index);
     o2 << " */";
@@ -444,58 +423,81 @@ void DisassemblerTables::emitContextDecision(
   o2.indent(i2) << "};" << "\n";
 }
 
-void DisassemblerTables::emitInstructionInfo(raw_ostream &o, uint32_t &i) 
-  const {
+void DisassemblerTables::emitInstructionInfo(raw_ostream &o,
+                                             unsigned &i) const {
+  unsigned NumInstructions = InstructionSpecifiers.size();
+
+  o << "static const struct OperandSpecifier x86OperandSets[]["
+    << X86_MAX_OPERANDS << "] = {\n";
+
+  typedef std::vector<std::pair<const char *, const char *> > OperandListTy;
+  std::map<OperandListTy, unsigned> OperandSets;
+
+  unsigned OperandSetNum = 0;
+  for (unsigned Index = 0; Index < NumInstructions; ++Index) {
+    OperandListTy OperandList;
+
+    for (unsigned OperandIndex = 0; OperandIndex < X86_MAX_OPERANDS;
+         ++OperandIndex) {
+      const char *Encoding =
+        stringForOperandEncoding((OperandEncoding)InstructionSpecifiers[Index]
+                                 .operands[OperandIndex].encoding);
+      const char *Type =
+        stringForOperandType((OperandType)InstructionSpecifiers[Index]
+                             .operands[OperandIndex].type);
+      OperandList.push_back(std::make_pair(Encoding, Type));
+    }
+    unsigned &N = OperandSets[OperandList];
+    if (N != 0) continue;
+
+    N = ++OperandSetNum;
+
+    o << "  { /* " << (OperandSetNum - 1) << " */\n";
+    for (unsigned i = 0, e = OperandList.size(); i != e; ++i) {
+      o << "    { " << OperandList[i].first << ", "
+        << OperandList[i].second << " },\n";
+    }
+    o << "  },\n";
+  }
+  o << "};" << "\n\n";
+
   o.indent(i * 2) << "static const struct InstructionSpecifier ";
   o << INSTRUCTIONS_STR "[" << InstructionSpecifiers.size() << "] = {\n";
-  
-  i++;
 
-  uint16_t numInstructions = InstructionSpecifiers.size();
-  uint16_t index, operandIndex;
+  i++;
 
-  for (index = 0; index < numInstructions; ++index) {
+  for (unsigned index = 0; index < NumInstructions; ++index) {
     o.indent(i * 2) << "{ /* " << index << " */" << "\n";
     i++;
 
     o.indent(i * 2) << stringForModifierType(
                        (ModifierType)InstructionSpecifiers[index].modifierType);
-    o << "," << "\n";
+    o << ",\n";
 
     o.indent(i * 2) << "0x";
     o << format("%02hhx", (uint16_t)InstructionSpecifiers[index].modifierBase);
-    o << "," << "\n";
-
-    o.indent(i * 2) << "{" << "\n";
-    i++;
-
-    for (operandIndex = 0; operandIndex < X86_MAX_OPERANDS; ++operandIndex) {
-      o.indent(i * 2) << "{ ";
-      o <<stringForOperandEncoding((OperandEncoding)InstructionSpecifiers[index]
-                                   .operands[operandIndex]
-                                   .encoding);
-      o << ", ";
-      o << stringForOperandType((OperandType)InstructionSpecifiers[index]
-                                .operands[operandIndex]
-                                .type);
-      o << " }";
-
-      if (operandIndex < X86_MAX_OPERANDS - 1)
-        o << ",";
-
-      o << "\n";
+    o << ",\n";
+
+    OperandListTy OperandList;
+    for (unsigned OperandIndex = 0; OperandIndex < X86_MAX_OPERANDS;
+         ++OperandIndex) {
+      const char *Encoding =
+        stringForOperandEncoding((OperandEncoding)InstructionSpecifiers[index]
+                                 .operands[OperandIndex].encoding);
+      const char *Type =
+        stringForOperandType((OperandType)InstructionSpecifiers[index]
+                             .operands[OperandIndex].type);
+      OperandList.push_back(std::make_pair(Encoding, Type));
     }
+    o.indent(i * 2) << (OperandSets[OperandList] - 1) << ",\n";
 
-    i--;
-    o.indent(i * 2) << "}," << "\n";
-    
     o.indent(i * 2) << "/* " << InstructionSpecifiers[index].name << " */";
     o << "\n";
 
     i--;
     o.indent(i * 2) << "}";
 
-    if (index + 1 < numInstructions)
+    if (index + 1 < NumInstructions)
       o << ",";
 
     o << "\n";
@@ -505,14 +507,12 @@ void DisassemblerTables::emitInstructionInfo(raw_ostream &o, uint32_t &i)
   o.indent(i * 2) << "};" << "\n";
 }
 
-void DisassemblerTables::emitContextTable(raw_ostream &o, uint32_t &i) const {
-  uint16_t index;
-
-  o.indent(i * 2) << "static const InstructionContext " CONTEXTS_STR
+void DisassemblerTables::emitContextTable(raw_ostream &o, unsigned &i) const {
+  o.indent(i * 2) << "static const uint8_t " CONTEXTS_STR
                      "[256] = {\n";
   i++;
 
-  for (index = 0; index < 256; ++index) {
+  for (unsigned index = 0; index < 256; ++index) {
     o.indent(i * 2);
 
     if ((index & ATTR_VEXL) && (index & ATTR_REXW) && (index & ATTR_OPSIZE))
@@ -545,7 +545,7 @@ void DisassemblerTables::emitContextTable(raw_ostream &o, uint32_t &i) const {
       o << "IC_64BIT_REXW_XS";
     else if ((index & ATTR_64BIT) && (index & ATTR_REXW) && (index & ATTR_XD))
       o << "IC_64BIT_REXW_XD";
-    else if ((index & ATTR_64BIT) && (index & ATTR_REXW) && 
+    else if ((index & ATTR_64BIT) && (index & ATTR_REXW) &&
              (index & ATTR_OPSIZE))
       o << "IC_64BIT_REXW_OPSIZE";
     else if ((index & ATTR_64BIT) && (index & ATTR_XD) && (index & ATTR_OPSIZE))
@@ -593,11 +593,8 @@ void DisassemblerTables::emitContextTable(raw_ostream &o, uint32_t &i) const {
   o.indent(i * 2) << "};" << "\n";
 }
 
-void DisassemblerTables::emitContextDecisions(raw_ostream &o1,
-                                            raw_ostream &o2,
-                                            uint32_t &i1,
-                                            uint32_t &i2)
-  const {
+void DisassemblerTables::emitContextDecisions(raw_ostream &o1, raw_ostream &o2,
+                                             unsigned &i1, unsigned &i2) const {
   emitContextDecision(o1, o2, i1, i2, *Tables[0], ONEBYTE_STR);
   emitContextDecision(o1, o2, i1, i2, *Tables[1], TWOBYTE_STR);
   emitContextDecision(o1, o2, i1, i2, *Tables[2], THREEBYTE38_STR);
@@ -607,15 +604,15 @@ void DisassemblerTables::emitContextDecisions(raw_ostream &o1,
 }
 
 void DisassemblerTables::emit(raw_ostream &o) const {
-  uint32_t i1 = 0;
-  uint32_t i2 = 0;
-  
+  unsigned i1 = 0;
+  unsigned i2 = 0;
+
   std::string s1;
   std::string s2;
-  
+
   raw_string_ostream o1(s1);
   raw_string_ostream o2(s2);
-  
+
   emitInstructionInfo(o, i2);
   o << "\n";
 
@@ -641,9 +638,7 @@ void DisassemblerTables::setTableFields(ModRMDecision     &decision,
                                         const ModRMFilter &filter,
                                         InstrUID          uid,
                                         uint8_t           opcode) {
-  unsigned index;
-
-  for (index = 0; index < 256; ++index) {
+  for (unsigned index = 0; index < 256; ++index) {
     if (filter.accepts(index)) {
       if (decision.instructionIDs[index] == uid)
         continue;
@@ -653,10 +648,10 @@ void DisassemblerTables::setTableFields(ModRMDecision     &decision,
           InstructionSpecifiers[uid];
         InstructionSpecifier &previousInfo =
           InstructionSpecifiers[decision.instructionIDs[index]];
-        
+
         if(newInfo.filtered)
           continue; // filtered instructions get lowest priority
-        
+
         if(previousInfo.name == "NOOP" && (newInfo.name == "XCHG16ar" ||
                                            newInfo.name == "XCHG32ar" ||
                                            newInfo.name == "XCHG32ar64" ||
@@ -665,7 +660,7 @@ void DisassemblerTables::setTableFields(ModRMDecision     &decision,
 
         if (outranks(previousInfo.insnContext, newInfo.insnContext))
           continue;
-        
+
         if (previousInfo.insnContext == newInfo.insnContext &&
             !previousInfo.filtered) {
           errs() << "Error: Primary decode conflict: ";
@@ -690,17 +685,15 @@ void DisassemblerTables::setTableFields(OpcodeType          type,
                                         InstrUID            uid,
                                         bool                is32bit,
                                         bool                ignoresVEX_L) {
-  unsigned index;
-  
   ContextDecision &decision = *Tables[type];
 
-  for (index = 0; index < IC_max; ++index) {
+  for (unsigned index = 0; index < IC_max; ++index) {
     if (is32bit && inheritsFrom((InstructionContext)index, IC_64BIT))
       continue;
 
-    if (inheritsFrom((InstructionContext)index, 
+    if (inheritsFrom((InstructionContext)index,
                      InstructionSpecifiers[uid].insnContext, ignoresVEX_L))
-      setTableFields(decision.opcodeDecisions[index].modRMDecisions[opcode], 
+      setTableFields(decision.opcodeDecisions[index].modRMDecisions[opcode],
                      filter,
                      uid,
                      opcode);