Change TableGen so that binary literals such as 0b001 are now sized.

Instead of these becoming an integer literal internally, they now become bits<n> values.

Prior to this change, 0b001 was 1 bit long.  This is confusing as clearly the user gave 3 bits.
This new type holds both the literal value and the size, and so can ensure sizes match on initializers.

For example, this used to be legal

bits<1> x = 0b00;

but now it must be written as

bits<2> x = 0b00;

llvm-svn: 215084

4 files changed, 56 insertions, 1 deletions

diff --git a/llvm/lib/TableGen/TGLexer.cpp b/llvm/lib/TableGen/TGLexer.cpp
index cfcc119f82d..63b85842d6a 100644
--- a/llvm/lib/TableGen/TGLexer.cpp
+++ b/llvm/lib/TableGen/TGLexer.cpp
@@ -411,7 +411,7 @@ tgtok::TokKind TGLexer::LexNumber() {
       if (CurPtr == NumStart)
         return ReturnError(CurPtr-2, "Invalid binary number");
       CurIntVal = strtoll(NumStart, nullptr, 2);
-      return tgtok::IntVal;
+      return tgtok::BinaryIntVal;
     }
   }
 
diff --git a/llvm/lib/TableGen/TGLexer.h b/llvm/lib/TableGen/TGLexer.h
index c7f7306a149..9fb2e3536e0 100644
--- a/llvm/lib/TableGen/TGLexer.h
+++ b/llvm/lib/TableGen/TGLexer.h
@@ -52,6 +52,10 @@ namespace tgtok {
 
     // Integer value.
     IntVal,
+
+    // Binary constant.  Note that these are sized according to the number of
+    // bits given.
+    BinaryIntVal,
     
     // String valued tokens.
     Id, StrVal, VarName, CodeFragment
@@ -105,6 +109,11 @@ public:
     assert(CurCode == tgtok::IntVal && "This token isn't an integer");
     return CurIntVal;
   }
+  std::pair<int64_t, unsigned> getCurBinaryIntVal() const {
+    assert(CurCode == tgtok::BinaryIntVal &&
+           "This token isn't a binary integer");
+    return std::make_pair(CurIntVal, (CurPtr - TokStart)-2);
+  }
 
   SMLoc getLoc() const;
   
diff --git a/llvm/lib/TableGen/TGParser.cpp b/llvm/lib/TableGen/TGParser.cpp
index d020a8bdc1a..709c7a52b9f 100644
--- a/llvm/lib/TableGen/TGParser.cpp
+++ b/llvm/lib/TableGen/TGParser.cpp
@@ -1182,6 +1182,15 @@ Init *TGParser::ParseSimpleValue(Record *CurRec, RecTy *ItemType,
     Lex.Lex();  // Skip '#'.
     return ParseSimpleValue(CurRec, ItemType, Mode);
   case tgtok::IntVal: R = IntInit::get(Lex.getCurIntVal()); Lex.Lex(); break;
+  case tgtok::BinaryIntVal: {
+    auto BinaryVal = Lex.getCurBinaryIntVal();
+    SmallVector<Init*, 16> Bits(BinaryVal.second);
+    for (unsigned i = 0, e = BinaryVal.second; i != e; ++i)
+      Bits[i] = BitInit::get(BinaryVal.first & (1 << i));
+    R = BitsInit::get(Bits);
+    Lex.Lex();
+    break;
+  }
   case tgtok::StrVal: {
     std::string Val = Lex.getCurStrVal();
     Lex.Lex();
diff --git a/llvm/test/TableGen/BitsInit.td b/llvm/test/TableGen/BitsInit.td
index 28fd3196fd0..c3d5113ebdb 100644
--- a/llvm/test/TableGen/BitsInit.td
+++ b/llvm/test/TableGen/BitsInit.td
@@ -20,3 +20,40 @@ def a {
 // CHECK:   bits<2> b = { 1, 0 };
 // CHECK:   bits<2> c = { 1, 1 };
 // CHECK: }
+
+def {
+  bits<2> B1 = 0b011;  // bitfield is too small, reject
+  bits<3> B2 = 0b011;  // ok
+
+  bits<2> C1 = 0b111;  // bitfield is too small, reject
+  bits<3> C2 = 0b111;  // ok
+
+  bits<2> D1 = { 0, 0 }; // ok
+  bits<2> D2 = { 0b00 }; // ok
+  bits<3> D3 = { 0, 0 }; // type mismatch.  RHS doesn't have enough bits
+  bits<3> D4 = { 0b00 }; // type mismatch.  RHS doesn't have enough bits
+  bits<1> D5 = { 0 };    // ok
+  bits<1> D6 = { 1 };    // ok
+  bits<1> D7 = { 3 };    // type mismatch.  LHS doesn't have enough bits
+  bits<2> D8 = { 0 };    // type mismatch.  RHS doesn't have enough bits
+
+  bits<8> E;
+  let E{7-0} = {0,0,1,?,?,?,?,?};
+  let E{3-0} = 0b0010;
+}
+
+// CHECK: def {{.*}} {
+// CHECK: bits<2> B1;
+// CHECK: bits<3> B2 = { 0, 1, 1 };
+// CHECK: bits<2> C1;
+// CHECK: bits<3> C2 = { 1, 1, 1 };
+// CHECK: bits<2> D1 = { 0, 0 };
+// CHECK: bits<2> D2 = { 0, 0 };
+// CHECK: bits<3> D3;
+// CHECK: bits<3> D4;
+// CHECK: bits<1> D5 = { 0 };
+// CHECK: bits<1> D6 = { 1 };
+// CHECK: bits<1> D7 = { ? };
+// CHECK: bits<2> D8;
+// CHECK: bits<8> E = { 0, 0, 1, ?, 0, 0, 1, 0 };
+// CHECK: }