[llvm] f0c5044 - [X86InstPrinter] Introduce markup tags emission

[Fangrui Song via llvm-commits]

Author: Antonio Frighetto
Date: 2022-09-01T21:04:35-07:00
New Revision: f0c50447f6d3da6c170c7ccf790af37611456b4c

URL: https://github.com/llvm/llvm-project/commit/f0c50447f6d3da6c170c7ccf790af37611456b4c
DIFF: https://github.com/llvm/llvm-project/commit/f0c50447f6d3da6c170c7ccf790af37611456b4c.diff

LOG: [X86InstPrinter] Introduce markup tags emission

x86 assembly syntax emission now leverages markup tags, if enabled.

Differential Revision: https://reviews.llvm.org/D129869

Added: 
    llvm/test/MC/Disassembler/X86/marked-up-32.txt

Modified: 
    llvm/lib/Target/X86/MCTargetDesc/X86InstPrinterCommon.cpp
    llvm/lib/Target/X86/MCTargetDesc/X86IntelInstPrinter.cpp

Removed: 
    


################################################################################
diff  --git a/llvm/lib/Target/X86/MCTargetDesc/X86InstPrinterCommon.cpp b/llvm/lib/Target/X86/MCTargetDesc/X86InstPrinterCommon.cpp
index e78e98cfc09e7..fc6dc291724aa 100644
--- a/llvm/lib/Target/X86/MCTargetDesc/X86InstPrinterCommon.cpp
+++ b/llvm/lib/Target/X86/MCTargetDesc/X86InstPrinterCommon.cpp
@@ -320,6 +320,7 @@ void X86InstPrinterCommon::printPCRelImm(const MCInst *MI, uint64_t Address,
 
   const MCOperand &Op = MI->getOperand(OpNo);
   if (Op.isImm()) {
+    O << markup("<imm:");
     if (PrintBranchImmAsAddress) {
       uint64_t Target = Address + Op.getImm();
       if (MAI.getCodePointerSize() == 4)
@@ -327,6 +328,7 @@ void X86InstPrinterCommon::printPCRelImm(const MCInst *MI, uint64_t Address,
       O << formatHex(Target);
     } else
       O << formatImm(Op.getImm());
+    O << markup(">");
   } else {
     assert(Op.isExpr() && "unknown pcrel immediate operand");
     // If a symbolic branch target was added as a constant expression then print
@@ -334,7 +336,7 @@ void X86InstPrinterCommon::printPCRelImm(const MCInst *MI, uint64_t Address,
     const MCConstantExpr *BranchTarget = dyn_cast<MCConstantExpr>(Op.getExpr());
     int64_t Address;
     if (BranchTarget && BranchTarget->evaluateAsAbsolute(Address)) {
-      O << formatHex((uint64_t)Address);
+      O << markup("<imm:") << formatHex((uint64_t)Address) << markup(">");
     } else {
       // Otherwise, just print the expression.
       Op.getExpr()->print(O, &MAI);

diff  --git a/llvm/lib/Target/X86/MCTargetDesc/X86IntelInstPrinter.cpp b/llvm/lib/Target/X86/MCTargetDesc/X86IntelInstPrinter.cpp
index 2a2afa925a9c4..d45ec6cc8dfa3 100644
--- a/llvm/lib/Target/X86/MCTargetDesc/X86IntelInstPrinter.cpp
+++ b/llvm/lib/Target/X86/MCTargetDesc/X86IntelInstPrinter.cpp
@@ -34,7 +34,7 @@ using namespace llvm;
 #include "X86GenAsmWriter1.inc"
 
 void X86IntelInstPrinter::printRegName(raw_ostream &OS, unsigned RegNo) const {
-  OS << getRegisterName(RegNo);
+  OS << markup("<reg:") << getRegisterName(RegNo) << markup(">");
 }
 
 void X86IntelInstPrinter::printInst(const MCInst *MI, uint64_t Address,
@@ -361,7 +361,7 @@ void X86IntelInstPrinter::printOperand(const MCInst *MI, unsigned OpNo,
   if (Op.isReg()) {
     printRegName(O, Op.getReg());
   } else if (Op.isImm()) {
-    O << formatImm((int64_t)Op.getImm());
+    O << markup("<imm:") << formatImm((int64_t)Op.getImm()) << markup(">");
   } else {
     assert(Op.isExpr() && "unknown operand kind in printOperand");
     O << "offset ";
@@ -388,7 +388,7 @@ void X86IntelInstPrinter::printMemReference(const MCInst *MI, unsigned Op,
   // If this has a segment register, print it.
   printOptionalSegReg(MI, Op + X86::AddrSegmentReg, O);
 
-  O << '[';
+  O << markup("<mem:") << '[';
 
   bool NeedPlus = false;
   if (BaseReg.getReg()) {
@@ -419,28 +419,28 @@ void X86IntelInstPrinter::printMemReference(const MCInst *MI, unsigned Op,
           DispVal = -DispVal;
         }
       }
-      O << formatImm(DispVal);
+      O << markup("<imm:") << formatImm(DispVal) << markup(">");
     }
   }
 
-  O << ']';
+  O << ']' << markup(">");
 }
 
 void X86IntelInstPrinter::printSrcIdx(const MCInst *MI, unsigned Op,
                                       raw_ostream &O) {
   // If this has a segment register, print it.
   printOptionalSegReg(MI, Op + 1, O);
-  O << '[';
+  O << markup("<mem:") << '[';
   printOperand(MI, Op, O);
-  O << ']';
+  O << ']' << markup(">");
 }
 
 void X86IntelInstPrinter::printDstIdx(const MCInst *MI, unsigned Op,
                                       raw_ostream &O) {
   // DI accesses are always ES-based.
-  O << "es:[";
+  O << "es:" << markup("<mem:") << '[';
   printOperand(MI, Op, O);
-  O << ']';
+  O << ']' << markup(">");
 }
 
 void X86IntelInstPrinter::printMemOffset(const MCInst *MI, unsigned Op,
@@ -450,16 +450,16 @@ void X86IntelInstPrinter::printMemOffset(const MCInst *MI, unsigned Op,
   // If this has a segment register, print it.
   printOptionalSegReg(MI, Op + 1, O);
 
-  O << '[';
+  O << markup("<mem:") << '[';
 
   if (DispSpec.isImm()) {
-    O << formatImm(DispSpec.getImm());
+    O << markup("<imm:") << formatImm(DispSpec.getImm()) << markup(">");
   } else {
     assert(DispSpec.isExpr() && "non-immediate displacement?");
     DispSpec.getExpr()->print(O, &MAI);
   }
 
-  O << ']';
+  O << ']' << markup(">");
 }
 
 void X86IntelInstPrinter::printU8Imm(const MCInst *MI, unsigned Op,
@@ -467,7 +467,8 @@ void X86IntelInstPrinter::printU8Imm(const MCInst *MI, unsigned Op,
   if (MI->getOperand(Op).isExpr())
     return MI->getOperand(Op).getExpr()->print(O, &MAI);
 
-  O << formatImm(MI->getOperand(Op).getImm() & 0xff);
+  O << markup("<imm:") << formatImm(MI->getOperand(Op).getImm() & 0xff)
+    << markup(">");
 }
 
 void X86IntelInstPrinter::printSTiRegOperand(const MCInst *MI, unsigned OpNo,

diff  --git a/llvm/test/MC/Disassembler/X86/marked-up-32.txt b/llvm/test/MC/Disassembler/X86/marked-up-32.txt
new file mode 100644
index 0000000000000..20fe31538848c
--- /dev/null
+++ b/llvm/test/MC/Disassembler/X86/marked-up-32.txt
@@ -0,0 +1,6 @@
+# RUN: llvm-mc -triple=i386 --mdis %s | FileCheck %s
+
+# CHECK: je	<imm:60>
+0x0f 0x84 0x3c 0x00 0x00 0x00
+# CHECK: jecxz	<imm:-127>
+0xe3 0x81