C++: Support constexpr strings for asm statements

Some programing styles use a lot of inline assembler, and it is common
to use very complex preprocessor macros to generate the assembler
strings for the asm statements. In C++ there would be a typesafe alternative
using templates and constexpr to generate the assembler strings, but
unfortunately the asm statement requires plain string literals, so this
doesn't work.

This patch modifies the C++ parser to accept strings generated by
constexpr instead of just plain strings. This requires new syntax
because e.g. asm("..." : "r" (expr)) would be ambigious with a function
call. I chose () to make it unique. For example now you can write

constexpr const char *genasm() { return "insn"; }
constexpr const char *genconstraint() { return "r"; }

	asm(genasm() :: (genconstraint()) (input));

The constexpr strings are allowed for the asm template, the
constraints and the clobbers (every time current asm accepts a string)

This version allows the same constexprs as C++26 static_assert,
following Jakub's suggestion.

The drawback of this scheme is that the constexpr doesn't have
full control over the input/output/clobber lists, but that can be
usually handled with a switch statement.  One could imagine
more flexible ways to handle that, for example supporting constexpr
vectors for the clobber list, or similar. But even without
that it is already useful.

Bootstrapped and full test on x86_64-linux.

gcc/c-family/ChangeLog:

	* c-cppbuiltin.cc (c_cpp_builtins): Define __GXX_CONSTEXPR_ASM__

gcc/cp/ChangeLog:

	* parser.cc (cp_parser_asm_string_expression): New function
	to handle constexpr strings for asm.
	(cp_parser_asm_definition): Use cp_parser_asm_string_expression.
	(cp_parser_yield_expression): Dito.
	(cp_parser_asm_specification_opt): Dito.
	(cp_parser_asm_operand_list): Dito.
	(cp_parser_asm_clobber_list): Dito.

gcc/ChangeLog:

	* doc/extend.texi: Document constexpr asm.

gcc/testsuite/ChangeLog:

	* g++.dg/ext/asm11.C: Adjust to new error message.
	* g++.dg/ext/asm9.C: Dito.
	* g++.dg/parse/asm1.C: Dito.
	* g++.dg/parse/asm2.C: Dito.
	* g++.dg/parse/asm3.C: Dito.
	* g++.dg/cpp1z/constexpr-asm-1.C: New test.
	* g++.dg/cpp1z/constexpr-asm-2.C: New test.
	* g++.dg/cpp1z/constexpr-asm-3.C: New test.

11 files changed, 193 insertions, 42 deletions

diff --git a/gcc/c-family/c-cppbuiltin.cc b/gcc/c-family/c-cppbuiltin.cc
index d9b84a0..dfd8f6f 100644
--- a/gcc/c-family/c-cppbuiltin.cc
+++ b/gcc/c-family/c-cppbuiltin.cc
@@ -954,7 +954,10 @@ c_cpp_builtins (cpp_reader *pfile)
 	}
 
       if (cxx_dialect >= cxx11)
-        cpp_define (pfile, "__GXX_EXPERIMENTAL_CXX0X__");
+	{
+	  cpp_define (pfile, "__GXX_EXPERIMENTAL_CXX0X__");
+	  cpp_define (pfile, "__GXX_CONSTEXPR_ASM__");
+	}
 
       /* Binary literals have been allowed in g++ before C++11
 	 and were standardized for C++14.  */
diff --git a/gcc/cp/parser.cc b/gcc/cp/parser.cc
index 9f43a77..6cd7274 100644
--- a/gcc/cp/parser.cc
+++ b/gcc/cp/parser.cc
@@ -22833,6 +22833,52 @@ cp_parser_using_directive (cp_parser* parser)
   cp_parser_require (parser, CPP_SEMICOLON, RT_SEMICOLON);
 }
 
+/* Parse a string literal or constant expression yielding a string.
+   The constant expression uses extra parens to avoid ambiguity with "x" (expr).
+
+   asm-string-expr:
+     string-literal
+     ( constant-expr ) */
+
+static tree
+cp_parser_asm_string_expression (cp_parser *parser)
+{
+  cp_token *tok = cp_lexer_peek_token (parser->lexer);
+
+  if (tok->type == CPP_OPEN_PAREN)
+    {
+      matching_parens parens;
+      parens.consume_open (parser);
+      tree string = cp_parser_constant_expression (parser);
+      if (string != error_mark_node)
+	string = cxx_constant_value (string, tf_error);
+      if (TREE_CODE (string) == NOP_EXPR)
+	string = TREE_OPERAND (string, 0);
+      if (TREE_CODE (string) == ADDR_EXPR
+	  && TREE_CODE (TREE_OPERAND (string, 0)) == STRING_CST)
+	string = TREE_OPERAND (string, 0);
+      if (TREE_CODE (string) == VIEW_CONVERT_EXPR)
+	string = TREE_OPERAND (string, 0);
+      cexpr_str cstr (string);
+      if (!cstr.type_check (tok->location))
+	return error_mark_node;
+      const char *msg;
+      int len;
+      if (!cstr.extract (tok->location, msg, len))
+	return error_mark_node;
+      parens.require_close (parser);
+      string = build_string (len, msg);
+      return string;
+    }
+  else if (!cp_parser_is_string_literal (tok))
+    {
+      error_at (tok->location,
+		"expected string-literal or constexpr in brackets");
+      return error_mark_node;
+    }
+  return cp_parser_string_literal (parser, false, false);
+}
+
 /* Parse an asm-definition.
 
   asm-qualifier:
@@ -22845,19 +22891,19 @@ cp_parser_using_directive (cp_parser* parser)
     asm-qualifier-list asm-qualifier
 
    asm-definition:
-     asm ( string-literal ) ;
+     asm ( constant-expr ) ;
 
    GNU Extension:
 
    asm-definition:
-     asm asm-qualifier-list [opt] ( string-literal ) ;
-     asm asm-qualifier-list [opt] ( string-literal : asm-operand-list [opt] ) ;
-     asm asm-qualifier-list [opt] ( string-literal : asm-operand-list [opt]
+     asm asm-qualifier-list [opt] ( asm-string-expr ) ;
+     asm asm-qualifier-list [opt] ( asm-string-expr : asm-operand-list [opt] ) ;
+     asm asm-qualifier-list [opt] ( asm-string-expr : asm-operand-list [opt]
 				    : asm-operand-list [opt] ) ;
-     asm asm-qualifier-list [opt] ( string-literal : asm-operand-list [opt]
+     asm asm-qualifier-list [opt] ( asm-string-expr : asm-operand-list [opt]
 				    : asm-operand-list [opt]
 			  : asm-clobber-list [opt] ) ;
-     asm asm-qualifier-list [opt] ( string-literal : : asm-operand-list [opt]
+     asm asm-qualifier-list [opt] ( asm-string-expr : : asm-operand-list [opt]
 				    : asm-clobber-list [opt]
 				    : asm-goto-list ) ;
 
@@ -22976,8 +23022,7 @@ cp_parser_asm_definition (cp_parser* parser)
   if (!cp_parser_require (parser, CPP_OPEN_PAREN, RT_OPEN_PAREN))
     return;
   /* Look for the string.  */
-  tree string = cp_parser_string_literal (parser, /*translate=*/false,
-					  /*wide_ok=*/false);
+  tree string = cp_parser_asm_string_expression (parser);
   if (string == error_mark_node)
     {
       cp_parser_skip_to_closing_parenthesis (parser, true, false,
@@ -29647,7 +29692,7 @@ cp_parser_yield_expression (cp_parser* parser)
 /* Parse an (optional) asm-specification.
 
    asm-specification:
-     asm ( string-literal )
+     asm ( asm-string-expr )
 
    If the asm-specification is present, returns a STRING_CST
    corresponding to the string-literal.  Otherwise, returns
@@ -29670,9 +29715,7 @@ cp_parser_asm_specification_opt (cp_parser* parser)
   parens.require_open (parser);
 
   /* Look for the string-literal.  */
-  tree asm_specification = cp_parser_string_literal (parser,
-						     /*translate=*/false,
-						     /*wide_ok=*/false);
+  tree asm_specification = cp_parser_asm_string_expression (parser);
 
   /* Look for the `)'.  */
   parens.require_close (parser);
@@ -29687,8 +29730,8 @@ cp_parser_asm_specification_opt (cp_parser* parser)
      asm-operand-list , asm-operand
 
    asm-operand:
-     string-literal ( expression )
-     [ string-literal ] string-literal ( expression )
+     asm-string-expr ( expression )
+     [ asm-string-expr ] asm-string-expr ( expression )
 
    Returns a TREE_LIST representing the operands.  The TREE_VALUE of
    each node is the expression.  The TREE_PURPOSE is itself a
@@ -29721,10 +29764,8 @@ cp_parser_asm_operand_list (cp_parser* parser)
 	}
       else
 	name = NULL_TREE;
-      /* Look for the string-literal.  */
-      tree string_literal = cp_parser_string_literal (parser,
-						      /*translate=*/false,
-						      /*wide_ok=*/false);
+      /* Look for the string.  */
+      tree string_literal = cp_parser_asm_string_expression (parser);
 
       /* Look for the `('.  */
       matching_parens parens;
@@ -29757,8 +29798,8 @@ cp_parser_asm_operand_list (cp_parser* parser)
 /* Parse an asm-clobber-list.
 
    asm-clobber-list:
-     string-literal
-     asm-clobber-list , string-literal
+     const-expression
+     asm-clobber-list , const-expression
 
    Returns a TREE_LIST, indicating the clobbers in the order that they
    appeared.  The TREE_VALUE of each node is a STRING_CST.  */
@@ -29771,9 +29812,7 @@ cp_parser_asm_clobber_list (cp_parser* parser)
   while (true)
     {
       /* Look for the string literal.  */
-      tree string_literal = cp_parser_string_literal (parser,
-						      /*translate=*/false,
-						      /*wide_ok=*/false);
+      tree string_literal = cp_parser_asm_string_expression (parser);
       /* Add it to the list.  */
       clobbers = tree_cons (NULL_TREE, string_literal, clobbers);
       /* If the next token is not a `,', then the list is
diff --git a/gcc/doc/extend.texi b/gcc/doc/extend.texi
index 799a365..17e26c5 100644
--- a/gcc/doc/extend.texi
+++ b/gcc/doc/extend.texi
@@ -10700,14 +10700,30 @@ contain any instructions recognized by the assembler, including directives.
 GCC does not parse the assembler instructions themselves and 
 does not know what they mean or even whether they are valid assembler input. 
 
-You may place multiple assembler instructions together in a single @code{asm} 
-string, separated by the characters normally used in assembly code for the 
-system. A combination that works in most places is a newline to break the 
+You may place multiple assembler instructions together in a single @code{asm}
+string, separated by the characters normally used in assembly code for the
+system. A combination that works in most places is a newline to break the
 line, plus a tab character (written as @samp{\n\t}).
-Some assemblers allow semicolons as a line separator. However, 
-note that some assembler dialects use semicolons to start a comment. 
+Some assemblers allow semicolons as a line separator. However,
+note that some assembler dialects use semicolons to start a comment.
 @end table
 
+@node asm constexprs
+With gnu++11 or later the string can also be a compile time constant expression
+inside parens.  The constant expression can return a string or a container
+with data and size members, following similar rules as C++26 @code{static_assert}
+message. Any string is converted to the character set of the source code.
+When this feature is available the @code{__GXX_CONSTEXPR_ASM__} cpp symbol is defined.
+
+@example
+constexpr const char *genfoo() @{ return "foo"; @}
+
+void function()
+@{
+  asm((genfoo()));
+@}
+@end example
+
 @subsubheading Remarks
 Using extended @code{asm} (@pxref{Extended Asm}) typically produces
 smaller, safer, and more efficient code, and in most cases it is a
@@ -10850,20 +10866,27 @@ perform a jump to one of the labels listed in the @var{GotoLabels}.
 @item AssemblerTemplate
 This is a literal string that is the template for the assembler code. It is a 
 combination of fixed text and tokens that refer to the input, output, 
-and goto parameters. @xref{AssemblerTemplate}.
+and goto parameters. @xref{AssemblerTemplate}. With gnu++11 or later it can
+also be a constant expression inside parens (see @ref{asm constexprs}).
 
 @item OutputOperands
 A comma-separated list of the C variables modified by the instructions in the 
 @var{AssemblerTemplate}.  An empty list is permitted.  @xref{OutputOperands}.
+With gnu++11 or later the strings can also be constant expressions inside parens
+(see @ref{asm constexprs})
 
 @item InputOperands
 A comma-separated list of C expressions read by the instructions in the 
 @var{AssemblerTemplate}.  An empty list is permitted.  @xref{InputOperands}.
+With gnu++11 or later the strings can also be constant expressions inside parens
+(see @ref{asm constexprs})
 
 @item Clobbers
 A comma-separated list of registers or other values changed by the 
 @var{AssemblerTemplate}, beyond those listed as outputs.
 An empty list is permitted.  @xref{Clobbers and Scratch Registers}.
+With gnu++11 or later the strings can also be constant expressions inside parens
+(see @ref{asm constexprs})
 
 @item GotoLabels
 When you are using the @code{goto} form of @code{asm}, this section contains 
diff --git a/gcc/testsuite/g++.dg/cpp1z/constexpr-asm-1.C b/gcc/testsuite/g++.dg/cpp1z/constexpr-asm-1.C
new file mode 100644
index 0000000..7cc6b37
--- /dev/null
+++ b/gcc/testsuite/g++.dg/cpp1z/constexpr-asm-1.C
@@ -0,0 +1,30 @@
+/* { dg-do compile } */
+/* { dg-options "-std=gnu++11" } */
+
+constexpr const char *genfoo ()
+{
+  return "foo %1,%0";
+}
+
+constexpr const char *genoutput ()
+{
+  return "=r";
+}
+
+constexpr const char *geninput ()
+{
+  return "r";
+}
+
+constexpr const char *genclobber ()
+{
+  return "memory";
+}
+
+void f()
+{
+  int a;
+  asm((genfoo ()) : (genoutput ()) (a) : (geninput ()) (1) : (genclobber ()));
+}
+
+/* { dg-final { scan-assembler "foo" } } */
diff --git a/gcc/testsuite/g++.dg/cpp1z/constexpr-asm-2.C b/gcc/testsuite/g++.dg/cpp1z/constexpr-asm-2.C
new file mode 100644
index 0000000..7d0eb59
--- /dev/null
+++ b/gcc/testsuite/g++.dg/cpp1z/constexpr-asm-2.C
@@ -0,0 +1,21 @@
+/* { dg-do compile } */
+/* { dg-options "-std=gnu++11" } */
+
+using size_t = typeof(sizeof(0));
+template <typename T, size_t N>
+struct array {
+  constexpr size_t size () const { return N; }
+  constexpr const T *data () const { return a; }
+  const T a[N];
+};
+
+void f()
+{
+  int a;
+  asm((array<char, 3> {'f','o','o'}) :
+      (array<char, 2>{'=','r'})  (a) :
+      (array<char, 1>{'r'}) (1) :
+      (array<char, 6>{'m','e','m','o','r','y'}));
+}
+
+/* { dg-final { scan-assembler "foo" } } */
diff --git a/gcc/testsuite/g++.dg/cpp1z/constexpr-asm-3.C b/gcc/testsuite/g++.dg/cpp1z/constexpr-asm-3.C
new file mode 100644
index 0000000..d336318
--- /dev/null
+++ b/gcc/testsuite/g++.dg/cpp1z/constexpr-asm-3.C
@@ -0,0 +1,31 @@
+/* { dg-do compile } */
+/* { dg-options "-std=gnu++11" } */
+
+constexpr const char *genfoo ()
+{
+  return "foo %1,%0";
+}
+
+constexpr const char *genoutput ()
+{
+  return "=r";
+}
+
+constexpr const char *geninput ()
+{
+  return "r";
+}
+
+constexpr const char *genclobber ()
+{
+  return "memory";
+}
+
+void f()
+{
+  int a;
+  asm(genfoo () : /* { dg-error "expected string-literal or constexpr in brackets" } */
+      genoutput() (a) :
+      geninput() (1) :
+      genclobber());
+}
diff --git a/gcc/testsuite/g++.dg/ext/asm11.C b/gcc/testsuite/g++.dg/ext/asm11.C
index 7939aac..97a299a 100644
--- a/gcc/testsuite/g++.dg/ext/asm11.C
+++ b/gcc/testsuite/g++.dg/ext/asm11.C
@@ -6,15 +6,15 @@ void
 foo ()
 {
   int i;
-  asm ();		  // { dg-error "expected string-literal before" }
-  asm (1);		  // { dg-error "expected string-literal before" }
-  asm (int);		  // { dg-error "expected string-literal before" }
-  asm (: "=r" (i));	  // { dg-error "expected string-literal before" }
-  asm (1 : "=r" (i));	  // { dg-error "expected string-literal before" }
-  asm (int : "=r" (i));	  // { dg-error "expected string-literal before" }
-  asm (: : "r" (i));	  // { dg-error "expected string-literal before" }
-  asm (1 : : "r" (i));	  // { dg-error "expected string-literal before" }
-  asm (int : : "r" (i));  // { dg-error "expected string-literal before" }
-  asm (: : : "memory");	  // { dg-error "expected string-literal before" }
-  asm (1 : : : "memory"); // { dg-error "expected string-literal before" }
+  asm ();		  // { dg-error "expected string-literal" }
+  asm (1);		  // { dg-error "expected string-literal" }
+  asm (int);		  // { dg-error "expected string-literal" }
+  asm (: "=r" (i));	  // { dg-error "expected string-literal" }
+  asm (1 : "=r" (i));	  // { dg-error "expected string-literal" }
+  asm (int : "=r" (i));	  // { dg-error "expected string-literal" }
+  asm (: : "r" (i));	  // { dg-error "expected string-literal" }
+  asm (1 : : "r" (i));	  // { dg-error "expected string-literal" }
+  asm (int : : "r" (i));  // { dg-error "expected string-literal" }
+  asm (: : : "memory");	  // { dg-error "expected string-literal" }
+  asm (1 : : : "memory"); // { dg-error "expected string-literal" }
 }
diff --git a/gcc/testsuite/g++.dg/ext/asm9.C b/gcc/testsuite/g++.dg/ext/asm9.C
index 9daa01b..3bce845 100644
--- a/gcc/testsuite/g++.dg/ext/asm9.C
+++ b/gcc/testsuite/g++.dg/ext/asm9.C
@@ -3,5 +3,6 @@
 
 void foo()
 {
-  asm("" ::: X); // { dg-error "before" }
+  asm("" ::: X); // { dg-error "string-literal" }
+		 // { dg-error "before" "" { target *-*-* } .-1 }
 }
diff --git a/gcc/testsuite/g++.dg/parse/asm1.C b/gcc/testsuite/g++.dg/parse/asm1.C
index 009ffbd..9ce24b5 100644
--- a/gcc/testsuite/g++.dg/parse/asm1.C
+++ b/gcc/testsuite/g++.dg/parse/asm1.C
@@ -3,4 +3,5 @@
 void foo()
 {
   asm("" : 0);  // { dg-error "numeric constant|token" }
+		// { dg-error "string-literal" "" { target *-*-* } .-1 }
 }
diff --git a/gcc/testsuite/g++.dg/parse/asm2.C b/gcc/testsuite/g++.dg/parse/asm2.C
index 09924c9..d4e1c6e 100644
--- a/gcc/testsuite/g++.dg/parse/asm2.C
+++ b/gcc/testsuite/g++.dg/parse/asm2.C
@@ -3,4 +3,5 @@
 void foo()
 {
   asm("" :: 0);  // { dg-error "numeric constant|token" }
+		 // { dg-error "string-literal" "" { target *-*-* } .-1 }
 }
diff --git a/gcc/testsuite/g++.dg/parse/asm3.C b/gcc/testsuite/g++.dg/parse/asm3.C
index 91f19e4..dccee5a 100644
--- a/gcc/testsuite/g++.dg/parse/asm3.C
+++ b/gcc/testsuite/g++.dg/parse/asm3.C
@@ -3,4 +3,5 @@
 void foo()
 {
   asm ("%[x]" : [0](x));  // { dg-error "numeric constant|token" }
+			  // { dg-error "string-literal" "" { target *-*-* } .-1 }
 }