IR: Support parsing numeric block ids, and emit them in textual output.

Just as as llvm IR supports explicitly specifying numeric value ids
for instructions, and emits them by default in textual output, now do
the same for blocks.

This is a slightly incompatible change in the textual IR format.

Previously, llvm would parse numeric labels as string names. E.g.
  define void @f() {
    br label %"55"
  55:
    ret void
  }
defined a label *named* "55", even without needing to be quoted, while
the reference required quoting. Now, if you intend a block label which
looks like a value number to be a name, you must quote it in the
definition too (e.g. `"55":`).

Previously, llvm would print nameless blocks only as a comment, and
would omit it if there was no predecessor. This could cause confusion
for readers of the IR, just as unnamed instructions did prior to the
addition of "%5 = " syntax, back in 2008 (PR2480).

Now, it will always print a label for an unnamed block, with the
exception of the entry block. (IMO it may be better to print it for
the entry-block as well. However, that requires updating many more
tests.)

Thus, the following is supported, and is the canonical printing:
  define i32 @f(i32, i32) {
    %3 = add i32 %0, %1
    br label %4

  4:
    ret i32 %3
  }

New test cases covering this behavior are added, and other tests
updated as required.

Differential Revision: https://reviews.llvm.org/D58548

llvm-svn: 356789

Author: jyknight

clang/test/CodeGenCXX/discard-name-values.cpp

@@ -10,7 +10,7 @@ bool test(bool pred) {
   // CHECK: br i1 %pred, label %if.then, label %if.end
 
   if (pred) {
-    // DISCARDVALUE: ; <label>:2:
+    // DISCARDVALUE: 2:
     // DISCARDVALUE-NEXT: tail call void @branch()
     // DISCARDVALUE-NEXT: br label %3
 
@@ -20,7 +20,7 @@ bool test(bool pred) {
     branch();
   }
 
-  // DISCARDVALUE: ; <label>:3:
+  // DISCARDVALUE: 3:
   // DISCARDVALUE-NEXT: ret i1 %0
 
   // CHECK: if.end:

llgo/test/irgen/imports.go

@@ -13,7 +13,7 @@ var X interface{}
 // CHECK-NEXT: %[[N:.*]] = load i1, i1* @"init$guard"
 // CHECK-NEXT: br i1 %[[N]], label %{{.*}}, label %[[L:.*]]
 
-// CHECK: ; <label>:[[L]]
+// CHECK: [[L]]:
 // CHECK-NEXT: call void @__go_register_gc_roots
 // CHECK-NEXT: store i1 true, i1* @"init$guard"
 // CHECK-NEXT: call void @fmt..import(i8* undef)

llvm/docs/LangRef.rst

@@ -741,11 +741,13 @@ A function definition contains a list of basic blocks, forming the CFG (Control
 Flow Graph) for the function. Each basic block may optionally start with a label
 (giving the basic block a symbol table entry), contains a list of instructions,
 and ends with a :ref:`terminator <terminators>` instruction (such as a branch or
-function return). If an explicit label is not provided, a block is assigned an
-implicit numbered label, using the next value from the same counter as used for
-unnamed temporaries (:ref:`see above<identifiers>`). For example, if a function
-entry block does not have an explicit label, it will be assigned label "%0",
-then the first unnamed temporary in that block will be "%1", etc.
+function return). If an explicit label name is not provided, a block is assigned
+an implicit numbered label, using the next value from the same counter as used
+for unnamed temporaries (:ref:`see above<identifiers>`). For example, if a
+function entry block does not have an explicit label, it will be assigned label
+"%0", then the first unnamed temporary in that block will be "%1", etc. If a
+numeric label is explicitly specified, it must match the numeric label that
+would be used implicitly.
 
 The first basic block in a function is special in two ways: it is
 immediately executed on entrance to the function, and it is not allowed

llvm/lib/AsmParser/LLLexer.cpp

@@ -1048,7 +1048,17 @@ lltok::Kind LLLexer::LexDigitOrNegative() {
   for (; isdigit(static_cast<unsigned char>(CurPtr[0])); ++CurPtr)
     /*empty*/;
 
-  // Check to see if this really is a label afterall, e.g. "-1:".
+  // Check if this is a fully-numeric label:
+  if (isdigit(TokStart[0]) && CurPtr[0] == ':') {
+    uint64_t Val = atoull(TokStart, CurPtr);
+    ++CurPtr; // Skip the colon.
+    if ((unsigned)Val != Val)
+      Error("invalid value number (too large)!");
+    UIntVal = unsigned(Val);
+    return lltok::LabelID;
+  }
+
+  // Check to see if this really is a string label, e.g. "-1:".
   if (isLabelChar(CurPtr[0]) || CurPtr[0] == ':') {
     if (const char *End = isLabelTail(CurPtr)) {
       StrVal.assign(TokStart, End-1);

llvm/lib/AsmParser/LLParser.cpp

@@ -2926,13 +2926,27 @@ BasicBlock *LLParser::PerFunctionState::GetBB(unsigned ID, LocTy Loc) {
 /// unnamed.  If there is an error, this returns null otherwise it returns
 /// the block being defined.
 BasicBlock *LLParser::PerFunctionState::DefineBB(const std::string &Name,
-                                                 LocTy Loc) {
+                                                 int NameID, LocTy Loc) {
   BasicBlock *BB;
-  if (Name.empty())
+  if (Name.empty()) {
+    if (NameID != -1 && unsigned(NameID) != NumberedVals.size()) {
+      P.Error(Loc, "label expected to be numbered '" +
+                       Twine(NumberedVals.size()) + "'");
+      return nullptr;
+    }
     BB = GetBB(NumberedVals.size(), Loc);
-  else
+    if (!BB) {
+      P.Error(Loc, "unable to create block numbered '" +
+                       Twine(NumberedVals.size()) + "'");
+      return nullptr;
+    }
+  } else {
     BB = GetBB(Name, Loc);
-  if (!BB) return nullptr; // Already diagnosed error.
+    if (!BB) {
+      P.Error(Loc, "unable to create block named '" + Name + "'");
+      return nullptr;
+    }
+  }
 
   // Move the block to the end of the function.  Forward ref'd blocks are
   // inserted wherever they happen to be referenced.
@@ -5489,20 +5503,23 @@ bool LLParser::ParseFunctionBody(Function &Fn) {
 }
 
 /// ParseBasicBlock
-///   ::= LabelStr? Instruction*
+///   ::= (LabelStr|LabelID)? Instruction*
 bool LLParser::ParseBasicBlock(PerFunctionState &PFS) {
   // If this basic block starts out with a name, remember it.
   std::string Name;
+  int NameID = -1;
   LocTy NameLoc = Lex.getLoc();
   if (Lex.getKind() == lltok::LabelStr) {
     Name = Lex.getStrVal();
     Lex.Lex();
+  } else if (Lex.getKind() == lltok::LabelID) {
+    NameID = Lex.getUIntVal();
+    Lex.Lex();
   }
 
-  BasicBlock *BB = PFS.DefineBB(Name, NameLoc);
+  BasicBlock *BB = PFS.DefineBB(Name, NameID, NameLoc);
   if (!BB)
-    return Error(NameLoc,
-                 "unable to create block named '" + Name + "'");
+    return true;
 
   std::string NameStr;
 

llvm/lib/AsmParser/LLParser.h

@@ -445,7 +445,7 @@ namespace llvm {
       /// DefineBB - Define the specified basic block, which is either named or
       /// unnamed.  If there is an error, this returns null otherwise it returns
       /// the block being defined.
-      BasicBlock *DefineBB(const std::string &Name, LocTy Loc);
+      BasicBlock *DefineBB(const std::string &Name, int NameID, LocTy Loc);
 
       bool resolveForwardRefBlockAddresses();
     };

llvm/lib/AsmParser/LLToken.h

@@ -422,6 +422,7 @@ enum Kind {
   kw_varFlags,
 
   // Unsigned Valued tokens (UIntVal).
+  LabelID,    // 42:
   GlobalID,   // @42
   LocalVarID, // %42
   AttrGrpID,  // #42

llvm/lib/IR/AsmWriter.cpp

@@ -3482,23 +3482,24 @@ void AssemblyWriter::printArgument(const Argument *Arg, AttributeSet Attrs) {
 
 /// printBasicBlock - This member is called for each basic block in a method.
 void AssemblyWriter::printBasicBlock(const BasicBlock *BB) {
+  bool IsEntryBlock = BB == &BB->getParent()->getEntryBlock();
   if (BB->hasName()) {              // Print out the label if it exists...
     Out << "\n";
     PrintLLVMName(Out, BB->getName(), LabelPrefix);
     Out << ':';
-  } else if (!BB->use_empty()) {      // Don't print block # of no uses...
-    Out << "\n; <label>:";
+  } else if (!IsEntryBlock) {
+    Out << "\n";
     int Slot = Machine.getLocalSlot(BB);
     if (Slot != -1)
       Out << Slot << ":";
     else
-      Out << "<badref>";
+      Out << "<badref>:";
   }
 
   if (!BB->getParent()) {
     Out.PadToColumn(50);
     Out << "; Error: Block without parent!";
-  } else if (BB != &BB->getParent()->getEntryBlock()) {  // Not the entry block?
+  } else if (!IsEntryBlock) {
     // Output predecessors for the block.
     Out.PadToColumn(50);
     Out << ";";

llvm/test/Analysis/DominanceFrontier/new_pm_test.ll

@@ -3,48 +3,48 @@
 
 define void @a_linear_impl_fig_1() nounwind {
 0:
-  br label %"1"
+  br label %1
 1:
-  br label %"2"
+  br label %2
 2:
-  br label %"3"
+  br label %3
 3:
-  br i1 1, label %"13", label %"4"
+  br i1 1, label %12, label %4
 4:
-  br i1 1, label %"5", label %"1"
+  br i1 1, label %5, label %1
 5:
-  br i1 1, label %"8", label %"6"
+  br i1 1, label %8, label %6
 6:
-  br i1 1, label %"7", label %"4"
+  br i1 1, label %7, label %4
 7:
   ret void
 8:
-  br i1 1, label %"9", label %"1"
+  br i1 1, label %9, label %1
 9:
-  br label %"10"
+  br label %10
 10:
-  br i1 1, label %"12", label %"11"
+  br i1 1, label %13, label %11
 11:
-  br i1 1, label %"9", label %"8"
-13:
-  br i1 1, label %"2", label %"1"
+  br i1 1, label %9, label %8
 12:
-   switch i32 0, label %"1" [ i32 0, label %"9"
-                              i32 1, label %"8"]
+  br i1 1, label %2, label %1
+13:
+   switch i32 0, label %1 [ i32 0, label %9
+                              i32 1, label %8]
 }
 
 ; CHECK: DominanceFrontier for function: a_linear_impl_fig_1
-; CHECK-DAG:  DomFrontier for BB %"0" is:
-; CHECK-DAG:  DomFrontier for BB %"11" is:   %"{{[8|9]}}" %"{{[8|9]}}"
-; CHECK-DAG:  DomFrontier for BB %"1" is:    %"1"
-; CHECK-DAG:  DomFrontier for BB %"2" is:    %"{{[1|2]}}" %"{{[1|2]}}"
-; CHECK-DAG:  DomFrontier for BB %"3" is:    %"{{[1|2]}}" %"{{[1|2]}}"
-; CHECK-DAG:  DomFrontier for BB %"13" is:   %"{{[1|2]}}" %"{{[1|2]}}"
-; CHECK-DAG:  DomFrontier for BB %"4" is:    %"{{[1|4]}}" %"{{[1|4]}}"
-; CHECK-DAG:  DomFrontier for BB %"5" is:    %"{{[1|4]}}" %"{{[1|4]}}"
-; CHECK-DAG:  DomFrontier for BB %"8" is:    %"{{[1|8]}}" %"{{[1|8]}}"
-; CHECK-DAG:  DomFrontier for BB %"6" is:    %"4"
-; CHECK-DAG:  DomFrontier for BB %"7" is:
-; CHECK-DAG:  DomFrontier for BB %"9" is:    %"{{[1|8|9]}}" %"{{[1|8|9]}}" %"{{[1|8|9]}}"
-; CHECK-DAG:  DomFrontier for BB %"10" is:   %"{{[1|8|9]}}" %"{{[1|8|9]}}" %"{{[1|8|9]}}"
-; CHECK-DAG:  DomFrontier for BB %"12" is:   %"{{[1|8|9]}}" %"{{[1|8|9]}}" %"{{[1|8|9]}}"
+; CHECK-DAG:  DomFrontier for BB %0 is:
+; CHECK-DAG:  DomFrontier for BB %11 is:   %{{[8|9]}} %{{[8|9]}}
+; CHECK-DAG:  DomFrontier for BB %1 is:    %1
+; CHECK-DAG:  DomFrontier for BB %2 is:    %{{[1|2]}} %{{[1|2]}}
+; CHECK-DAG:  DomFrontier for BB %3 is:    %{{[1|2]}} %{{[1|2]}}
+; CHECK-DAG:  DomFrontier for BB %12 is:   %{{[1|2]}} %{{[1|2]}}
+; CHECK-DAG:  DomFrontier for BB %4 is:    %{{[1|4]}} %{{[1|4]}}
+; CHECK-DAG:  DomFrontier for BB %5 is:    %{{[1|4]}} %{{[1|4]}}
+; CHECK-DAG:  DomFrontier for BB %8 is:    %{{[1|8]}} %{{[1|8]}}
+; CHECK-DAG:  DomFrontier for BB %6 is:    %4
+; CHECK-DAG:  DomFrontier for BB %7 is:
+; CHECK-DAG:  DomFrontier for BB %9 is:    %{{[1|8|9]}} %{{[1|8|9]}} %{{[1|8|9]}}
+; CHECK-DAG:  DomFrontier for BB %10 is:   %{{[1|8|9]}} %{{[1|8|9]}} %{{[1|8|9]}}
+; CHECK-DAG:  DomFrontier for BB %13 is:   %{{[1|8|9]}} %{{[1|8|9]}} %{{[1|8|9]}}

llvm/test/Analysis/RegionInfo/cond_loop.ll

@@ -7,18 +7,18 @@
 ; RUN: opt < %s -passes='print<regions>' 2>&1 | FileCheck %s
 
 define void @normal_condition() nounwind {
-5:
+"5":
         br label %"0"
 
-0:
+"0":
 	br label %"1"
-1:
+"1":
 	br i1 1, label %"2", label %"3"
-2:
+"2":
 	ret void
-3:
+"3":
 	br i1 1, label %"1", label %"4"
-4:
+"4":
         br label %"0"
 }