[ms-inline asm] Add support for lexing binary integers with a [bB] suffix.

This is complicated by backward labels (e.g., 0b can be both a backward label
and a binary zero).  The current implementation assumes [0-9]b is always a
label and thus it's possible for 0b and 1b to not be interpreted correctly for
ms-style inline assembly.  However, this is relatively simple to fix in the
inline assembly (i.e., drop the [bB]).

This patch also limits backward labels to [0-9]b, so that only 0b and 1b are
ambiguous.
Part of rdar://12470373

llvm-svn: 174983

llvm/lib/MC/MCParser/AsmLexer.cpp

@@ -179,26 +179,48 @@ static unsigned doLookAhead(const char *&CurPtr, unsigned DefaultRadix) {
     }
   }
   bool isHex = *LookAhead == 'h' || *LookAhead == 'H';
-  CurPtr = isHex || !FirstHex ? LookAhead : FirstHex;
+  bool isBinary = LookAhead[-1] == 'b' || LookAhead[-1] == 'B';
+  CurPtr = (isBinary || isHex || !FirstHex) ? LookAhead : FirstHex;
   if (isHex)
     return 16;
+  if (isBinary) {
+    --CurPtr;
+    return 2;
+  }
   return DefaultRadix;
 }
 
 /// LexDigit: First character is [0-9].
 ///   Local Label: [0-9][:]
-///   Forward/Backward Label: [0-9][fb]
-///   Binary integer: 0b[01]+
+///   Forward/Backward Label: [0-9]+f or [0-9]b
+///   Binary integer: 0b[01]+ or [01][bB]
 ///   Octal integer: 0[0-7]+
 ///   Hex integer: 0x[0-9a-fA-F]+ or [0x]?[0-9][0-9a-fA-F]*[hH]
 ///   Decimal integer: [1-9][0-9]*
 AsmToken AsmLexer::LexDigit() {
+
+  // Backward Label: [0-9]b
+  if (*CurPtr == 'b') {
+    // See if we actually have "0b" as part of something like "jmp 0b\n"
+    if (!isdigit(CurPtr[1])) {
+      long long Value;
+      StringRef Result(TokStart, CurPtr - TokStart);
+      if (Result.getAsInteger(10, Value))
+        return ReturnError(TokStart, "invalid backward label");
+
+      return AsmToken(AsmToken::Integer, Result, Value);
+    }
+  }
+
+  // Binary integer: 1[01]*[bB]
   // Decimal integer: [1-9][0-9]*
+  // Hexidecimal integer: [1-9][0-9a-fA-F]*[hH]
   if (CurPtr[-1] != '0' || CurPtr[0] == '.') {
     unsigned Radix = doLookAhead(CurPtr, 10);
-    bool isHex = Radix == 16;
+    bool isDecimal = Radix == 10;
+
     // Check for floating point literals.
-    if (!isHex && (*CurPtr == '.' || *CurPtr == 'e')) {
+    if (isDecimal && (*CurPtr == '.' || *CurPtr == 'e')) {
       ++CurPtr;
       return LexFloatLiteral();
     }
@@ -211,7 +233,7 @@ AsmToken AsmLexer::LexDigit() {
       // integer, but that do fit in an unsigned one, we just convert them over.
       unsigned long long UValue;
       if (Result.getAsInteger(Radix, UValue))
-        return ReturnError(TokStart, !isHex ? "invalid decimal number" :
+        return ReturnError(TokStart, isDecimal ? "invalid decimal number" :
                            "invalid hexdecimal number");
       Value = (long long)UValue;
     }
@@ -227,15 +249,9 @@ AsmToken AsmLexer::LexDigit() {
     return AsmToken(AsmToken::Integer, Result, Value);
   }
 
+  // Binary integer: 0b[01]+
   if (*CurPtr == 'b') {
-    ++CurPtr;
-    // See if we actually have "0b" as part of something like "jmp 0b\n"
-    if (!isdigit(CurPtr[0])) {
-      --CurPtr;
-      StringRef Result(TokStart, CurPtr - TokStart);
-      return AsmToken(AsmToken::Integer, Result, 0);
-    }
-    const char *NumStart = CurPtr;
+    const char *NumStart = ++CurPtr;
     while (CurPtr[0] == '0' || CurPtr[0] == '1')
       ++CurPtr;
 
@@ -256,6 +272,7 @@ AsmToken AsmLexer::LexDigit() {
     return AsmToken(AsmToken::Integer, Result, Value);
   }
 
+  // Hex integer: 0x[0-9a-fA-F]+
   if (*CurPtr == 'x') {
     ++CurPtr;
     const char *NumStart = CurPtr;
@@ -282,17 +299,21 @@ AsmToken AsmLexer::LexDigit() {
                     (int64_t)Result);
   }
 
-  // Either octal or hexidecimal.
+  // Binary: 0[01]*[Bb], but not 0b.
+  // Octal: 0[0-7]*
+  // Hexidecimal: [0][0-9a-fA-F]*[hH]
   long long Value;
   unsigned Radix = doLookAhead(CurPtr, 8);
-  bool isHex = Radix == 16;
+  bool isBinary = Radix == 2;
+  bool isOctal = Radix == 8;
   StringRef Result(TokStart, CurPtr - TokStart);
   if (Result.getAsInteger(Radix, Value))
-    return ReturnError(TokStart, !isHex ? "invalid octal number" :
+    return ReturnError(TokStart, isOctal ? "invalid octal number" :
+                       isBinary ? "invalid binary number" :
                        "invalid hexdecimal number");
 
-  // Consume the [hH].
-  if (Radix == 16)
+  // Consume the [bB][hH].
+  if (Radix == 2 || Radix == 16)
     ++CurPtr;
 
   // The darwin/x86 (and x86-64) assembler accepts and ignores ULL and LL

llvm/test/MC/MachO/direction_labels.s

@@ -1,15 +1,15 @@
 // RUN: llvm-mc -triple i386-apple-darwin9 %s -filetype=obj -o - | macho-dump --dump-section-data | FileCheck %s
 
 direction_labels:
-10:     nop
-        jmp 10b
-        nop
+8:  nop
+    jmp 8b
+    nop
 	jne 0f
 0:	nop
 	jne 0b
-        jmp 11f
-11:     nop
-        ret
+    jmp 9f
+9:  nop
+    ret
 
 // CHECK: ('cputype', 7)
 // CHECK: ('cpusubtype', 3)

llvm/test/MC/X86/intel-syntax-binary.s

@@ -0,0 +1,14 @@
+// RUN: llvm-mc -triple x86_64-unknown-unknown -x86-asm-syntax=intel %s | FileCheck %s
+// rdar://12470373
+
+// Checks to make sure we parse the binary suffix properly.
+// CHECK: movl $1, %eax
+  mov eax, 01b
+// CHECK: movl $2, %eax
+  mov eax, 10b
+// CHECK: movl $3, %eax
+  mov eax, 11b
+// CHECK: movl $3, %eax
+  mov eax, 11B
+// CHECK: movl $2711, %eax
+  mov eax, 101010010111B