[LLVM][inline-asm] Altmacro absolute expression '%' feature

[m_zuckerman]

Summary

In this patch, I introduce a new alt macro feature.
This feature adds meaning for the % when using it as a prefix to the calling macro arguments.

In the altmacro mode, the percent sign '%' before an absolute expression convert the expression first to a string.
As described in the https://sourceware.org/binutils/docs-2.27/as/Altmacro.html
"Expression results as strings
You can write `%expr' to evaluate the expression expr and use the result as a string."

expression assumptions:

1. '%' can only evaluate an absolute expression.
2. Altmacro '%' must be the first character of the evaluated expression.
3. If no '%' is located before the expression, a regular module operation is expected.
4. The result of Absolute Expressions can be only integer.

[rengolin]

I think this patch and D32523 are small and simple enough that it could be merged into one...

MC/MCParser/AsmParser.cpp

488

These functions seem redundant... You could have used getLexer() directly.

[rengolin]

MC/AsmParser/AltMacroExpression.s

65 ↗

(On Diff #96740)

Can you add .noaltmacro and have some cases where it breaks?

This could be a separate file, so that you can add // RUN: not instead.

Make sure you cover before .altmacro is set as well as after it's unset.

[m_zuckerman]

Ping

[rengolin]

Can you merge D32523 with this one?

We don't usually add functionality without purpose/tests, and this is the functionality (and tests) D32523 needs. :)

Otherwise, looks good. Are you going to implement the remaining functionality?

[m_zuckerman]

In D32526#742126, @rengolin wrote:

Can you merge D32523 with this one?

We don't usually add functionality without purpose/tests, and this is the functionality (and tests) D32523 needs. :)

Otherwise, looks good. Are you going to implement the remaining functionality?

Yes,
Thanks, I will merge the two together.

[m_zuckerman]

merge two reviews to one D32526 and D32523

[m_zuckerman]

If it's ok with you and you think the review is ready, please mark LGTM

[rengolin]

Thanks! LGTM with the test change.

MC/AsmParser/negativ_altmacro_expression.s

12

Better to move this test above .altmacro to make sure the default behaviour is .noaltmacro.

MC/AsmParser/altmacro_expression.s

+# RUN: llvm-mc  %s| FileCheck %s
+
+# Checking that the '%' was evaluated as a string first
+# In a fail scenario: The asmprint will print: addl $%(1+4), %eax
+
+# CHECK:  addl $5, %eax
+.altmacro
+.macro percent_expr arg
+    addl $\arg, %eax
+.endm
+
+percent_expr %(1+4)
+
+
+# Checking that the second '%' acts as modulo operator
+# The altmacro percent '%' must be located before the first argument
+# If a percent is located in the middle of the estimated argument without any
+# '%' in the beginning , error will be generated.
+# The second percent '%' after the first altmacro percent '%' is a regular operator.
+
+#CHECK:  addl $1, %eax
+.macro inner_percent arg
+    addl $\arg, %eax
+.endm
+
+inner_percent %(1%4)
+
+
+# Checking for nested macro
+# The first argument use is for the calling function and the second use is for the evaluation.
+
+#CHECK:  addl    $1, %eax
+.macro macro_call_0 number
+    addl $\number, %eax
+.endm
+
+.macro macro_call_1 number
+    macro_call_\number %(\number + 1)
+.endm
+
+macro_call_1 %(1-1)
+
+
+# Checking the ability to pass a number of arguments.
+# The arguments can be separated by ',' or not.
+
+#CHECK: label013:
+#CHECK:  addl $0, %eax
+#CHECK:  addl $1, %eax
+#CHECK:  addl $3, %eax
+
+#CHECK: label014:
+#CHECK:  addl $0, %eax
+#CHECK:  addl $1, %eax
+#CHECK:  addl $4, %eax
+
+.macro multi_args_macro arg1 arg2 arg3
+    label\arg1\arg2\arg3:
+  addl $\arg1, %eax
+  addl $\arg2, %eax
+  addl $\arg3, %eax
+.endm
+
+multi_args_macro %(1+4-5) 1 %2+1
+multi_args_macro %(1+4-5),1,%4%10

MC/AsmParser/negativ_altmacro_expression.s

+# RUN: not llvm-mc  %s 2>&1 | FileCheck %s
+
+# This test is a negative test for the altmacro expression.
+# In this test we check the '.noaltmacro' directive.
+# We expect that '.altmacro' and '.noaltmacro' will act as a switch on/off directives to the alternate macro mode.
+# .noaltmacro returns the format into a regular macro handling.
+
+.altmacro
+.noaltmacro
+# CHECK:  error: unknown token in expression
+# CHECK-NEXT: addl $%(1%4), %eax
+.macro inner_percent arg
+    addl $\arg, %eax
+.endm
+
+inner_percent %(1%4)
+
+ # CHECK: multi_args_macro %(1+4-5) 1 %2+1
+ # CHECK: error: unknown token in expression
+ # CHECK-NEXT: addl $%(1+4-5), %eax
+
+
+ # CHECK: multi_args_macro %(1+4-5),1,%4%10
+ # CHECK: error: unknown token in expression
+ # CHECK-NEXT: addl $%(1+4-5), %eax
+ .macro multi_args_macro arg1 arg2 arg3
+    label\arg1\arg2\arg3:
+    addl $\arg1, %eax
+.endm
+
+multi_args_macro %(1+4-5) 1 %2+1
+multi_args_macro %(1+4-5),1,%4%10

MC/MCParser/AsmParser.cpp

enum DirectiveKind {
     DK_CV_INLINE_LINETABLE, DK_CV_DEF_RANGE, DK_CV_STRINGTABLE,
     DK_CV_FILECHECKSUMS,
     DK_CFI_SECTIONS, DK_CFI_STARTPROC, DK_CFI_ENDPROC, DK_CFI_DEF_CFA,
     DK_CFI_DEF_CFA_OFFSET, DK_CFI_ADJUST_CFA_OFFSET, DK_CFI_DEF_CFA_REGISTER,
     DK_CFI_OFFSET, DK_CFI_REL_OFFSET, DK_CFI_PERSONALITY, DK_CFI_LSDA,
     DK_CFI_REMEMBER_STATE, DK_CFI_RESTORE_STATE, DK_CFI_SAME_VALUE,
     DK_CFI_RESTORE, DK_CFI_ESCAPE, DK_CFI_SIGNAL_FRAME, DK_CFI_UNDEFINED,
     DK_CFI_REGISTER, DK_CFI_WINDOW_SAVE,
-    DK_MACROS_ON, DK_MACROS_OFF,
+    DK_MACROS_ON, DK_MACROS_OFF, DK_ALTMACRO, DK_NOALTMACRO,
     DK_MACRO, DK_EXITM, DK_ENDM, DK_ENDMACRO, DK_PURGEM,
     DK_SLEB128, DK_ULEB128,
     DK_ERR, DK_ERROR, DK_WARNING,
     DK_END
   };
 
   /// \brief Maps directive name --> DirectiveKind enum, for
   /// directives parsed by this class.
private:
   bool parseDirectiveCFIUndefined(SMLoc DirectiveLoc);
 
   // macro directives
   bool parseDirectivePurgeMacro(SMLoc DirectiveLoc);
   bool parseDirectiveExitMacro(StringRef Directive);
   bool parseDirectiveEndMacro(StringRef Directive);
   bool parseDirectiveMacro(SMLoc DirectiveLoc);
   bool parseDirectiveMacrosOnOff(StringRef Directive);
-
+  // alternate macro mode directives
+  bool parseDirectiveAltmacro(StringRef Directive);
   // ".bundle_align_mode"
   bool parseDirectiveBundleAlignMode();
   // ".bundle_lock"
   bool parseDirectiveBundleLock();
   // ".bundle_unlock"
   bool parseDirectiveBundleUnlock();
 
   // ".space", ".skip"
case DK_CFI_REGISTER:
       return parseDirectiveCFIRegister(IDLoc);
     case DK_CFI_WINDOW_SAVE:
       return parseDirectiveCFIWindowSave();
     case DK_MACROS_ON:
     case DK_MACROS_OFF:
       return parseDirectiveMacrosOnOff(IDVal);
     case DK_MACRO:
       return parseDirectiveMacro(IDLoc);
+    case DK_ALTMACRO:
+    case DK_NOALTMACRO:
+      return parseDirectiveAltmacro(IDVal);
     case DK_EXITM:
       return parseDirectiveExitMacro(IDVal);
     case DK_ENDM:
     case DK_ENDMACRO:
       return parseDirectiveEndMacro(IDVal);
     case DK_PURGEM:
       return parseDirectivePurgeMacro(IDLoc);
     case DK_END:
if (IsDarwin && !NParameters) {
             Pos += 3;
           else {
             OS << '\\' << Argument;
             Pos = I;
           }
         } else {
           bool VarargParameter = HasVararg && Index == (NParameters - 1);
           for (const AsmToken &Token : A[Index])
+            // For altmacro mode, you can write '%expr'.
+            // The prefix '%' evaluates the expression 'expr'
+            // and uses the result as a string (e.g. replace %(1+2) with the string "3").
+            // Here, we identify the integer token which is the result of the
+            // absolute expression evaluation and replace it with its string representation.
+            if ((Lexer.IsaAltMacroMode()) &&
+                 (*(Token.getString().begin()) == '%') && Token.is(AsmToken::Integer))
+              // Emit an integer value to the buffer.
+              OS << Token.getIntVal();
             // We expect no quotes around the string's contents when
             // parsing for varargs.
-            if (Token.getKind() != AsmToken::String || VarargParameter)
+            else if (Token.isNot(AsmToken::String) || VarargParameter)
               OS << Token.getString();
             else
               OS << Token.getStringContents();
 
           Pos += 1 + Argument.size();
         }
       }
     }
if (Lexer.is(AsmToken::Identifier) && Lexer.peekTok().is(AsmToken::Equal)) {
 
       if (Lexer.isNot(AsmToken::Equal))
         return TokError("expected '=' after formal parameter identifier");
 
       Lex();
 
       NamedParametersFound = true;
     }
+    bool Vararg = HasVararg && Parameter == (NParameters - 1);
 
     if (NamedParametersFound && FA.Name.empty())
       return Error(IDLoc, "cannot mix positional and keyword arguments");
 
-    bool Vararg = HasVararg && Parameter == (NParameters - 1);
-    if (parseMacroArgument(FA.Value, Vararg))
+    if (Lexer.IsaAltMacroMode() && Lexer.is(AsmToken::Percent)) {
+        SMLoc StrLoc = Lexer.getLoc();
+        SMLoc EndLoc;
+        const MCExpr *AbsoluteExp;
+        int64_t Value;
+        /// Eat '%'
+        Lex();
+        if (parseExpression(AbsoluteExp, EndLoc))
+          return false;
+        if (!AbsoluteExp->evaluateAsAbsolute(Value))
+          return Error(StrLoc, "expected absolute expression");
+        const char *StrChar = StrLoc.getPointer();
+        const char *EndChar = EndLoc.getPointer();
+        AsmToken newToken(AsmToken::Integer, StringRef(StrChar , EndChar - StrChar), Value);
+        FA.Value.push_back(newToken);
+    }
+    else if(parseMacroArgument(FA.Value, Vararg))
         return true;
 
     unsigned PI = Parameter;
     if (!FA.Name.empty()) {
       unsigned FAI = 0;
       for (FAI = 0; FAI < NParameters; ++FAI)
         if (M->Parameters[FAI].Name == FA.Name)
           break;
 
bool AsmParser::parseDirectiveCFIUndefined(SMLoc DirectiveLoc) {
 
   if (parseRegisterOrRegisterNumber(Register, DirectiveLoc))
     return true;
 
   getStreamer().EmitCFIUndefined(Register);
   return false;
 }
 
+/// parseDirectiveAltmacro
+/// ::= .altmacro
+/// ::= .noaltmacro
+bool AsmParser::parseDirectiveAltmacro(StringRef Directive) {
+  if (getLexer().isNot(AsmToken::EndOfStatement))
+    return TokError("unexpected token in '" + Directive + "' directive");
+  if (Directive == ".altmacro")
+    getLexer().SetAltMacroMode(true);
+  else
+    getLexer().SetAltMacroMode(false);
+  return false;
+}
+
 /// parseDirectiveMacrosOnOff
 /// ::= .macros_on
 /// ::= .macros_off
 bool AsmParser::parseDirectiveMacrosOnOff(StringRef Directive) {
   if (parseToken(AsmToken::EndOfStatement,
                  "unexpected token in '" + Directive + "' directive"))
     return true;
 
void AsmParser::initializeDirectiveKindMap() {
   DirectiveKindMap[".macro"] = DK_MACRO;
   DirectiveKindMap[".exitm"] = DK_EXITM;
   DirectiveKindMap[".endm"] = DK_ENDM;
   DirectiveKindMap[".endmacro"] = DK_ENDMACRO;
   DirectiveKindMap[".purgem"] = DK_PURGEM;
   DirectiveKindMap[".err"] = DK_ERR;
   DirectiveKindMap[".error"] = DK_ERROR;
   DirectiveKindMap[".warning"] = DK_WARNING;
+  DirectiveKindMap[".altmacro"] = DK_ALTMACRO;
+  DirectiveKindMap[".noaltmacro"] = DK_NOALTMACRO;
   DirectiveKindMap[".reloc"] = DK_RELOC;
   DirectiveKindMap[".dc"] = DK_DC;
   DirectiveKindMap[".dc.a"] = DK_DC_A;
   DirectiveKindMap[".dc.b"] = DK_DC_B;
   DirectiveKindMap[".dc.d"] = DK_DC_D;
   DirectiveKindMap[".dc.l"] = DK_DC_L;
   DirectiveKindMap[".dc.s"] = DK_DC_S;
   DirectiveKindMap[".dc.w"] = DK_DC_W;

MC/MCParser/MCAsmLexer.cpp

 //===- MCAsmLexer.cpp - Abstract Asm Lexer Interface ----------------------===//
 //
 //                     The LLVM Compiler Infrastructure
 //
 // This file is distributed under the University of Illinois Open Source
 // License. See LICENSE.TXT for details.
 //
 //===----------------------------------------------------------------------===//
 
 #include "llvm/ADT/StringRef.h"
 #include "llvm/MC/MCParser/MCAsmLexer.h"
 #include "llvm/Support/SMLoc.h"
 
 using namespace llvm;
 
-MCAsmLexer::MCAsmLexer() {
+MCAsmLexer::MCAsmLexer() : AltMacroMode(false) {
   CurTok.emplace_back(AsmToken::Space, StringRef());
 }
 
 MCAsmLexer::~MCAsmLexer() = default;
 
 SMLoc MCAsmLexer::getLoc() const {
   return SMLoc::getFromPointer(TokStart);
 }

llvm/MC/MCParser/MCAsmLexer.h

 
 protected: // Can only create subclasses.
   const char *TokStart = nullptr;
   bool SkipSpace = true;
   bool AllowAtInIdentifier;
   bool IsAtStartOfStatement = true;
   AsmCommentConsumer *CommentConsumer = nullptr;
 
+  bool AltMacroMode;
   MCAsmLexer();
 
   virtual AsmToken LexToken() = 0;
 
   void SetError(SMLoc errLoc, const std::string &err) {
     ErrLoc = errLoc;
     Err = err;
   }
 
 public:
   MCAsmLexer(const MCAsmLexer &) = delete;
   MCAsmLexer &operator=(const MCAsmLexer &) = delete;
   virtual ~MCAsmLexer();
 
+  bool IsaAltMacroMode() {
+    return AltMacroMode;
+  }
+
+  void SetAltMacroMode(bool AltMacroSet) {
+    AltMacroMode = AltMacroSet;
+  }
+
   /// Consume the next token from the input stream and return it.
   ///
   /// The lexer will continuosly return the end-of-file token once the end of
   /// the main input file has been reached.
   const AsmToken &Lex() {
     assert(!CurTok.empty());
     // Mark if we parsing out a EndOfStatement.
     IsAtStartOfStatement = CurTok.front().getKind() == AsmToken::EndOfStatement;