gccrs: Check function names with no_mangle.

gcc/rust/ChangeLog: * backend/rust-compile-base.cc (HIRCompileBase::setup_fndecl): Add parameter. * backend/rust-compile-base.h: Likewise. * backend/rust-compile-implitem.cc (CompileTraitItem::visit): Change type of parameter. * backend/rust-compile-item.cc (CompileItem::visit): Likewise. * lex/rust-input-source.h: Move constants to rust-codepoint.h * util/rust-codepoint.h (struct Codepoint): Add is_ascii method. * util/rust-punycode.cc (extract_basic_string): Use it. * lex/rust-lex.cc (Lexer::parse_byte_char): Likewise. * backend/rust-mangle.cc (legacy_mangle_name): Likewise. * util/rust-unicode.cc (is_ascii_only): New function. * util/rust-unicode.h (is_ascii_only): Likewise. Signed-off-by: Raiki Tamura <[email protected]>
Rust-GCC · Aug 13, 2023 · fa611cb · fa611cb
1 parent 87c5cc1
commit fa611cb
Show file tree

Hide file tree

Showing 11 changed files with 50 additions and 23 deletions.
diff --git a/gcc/rust/backend/rust-compile-base.cc b/gcc/rust/backend/rust-compile-base.cc
@@ -30,6 +30,7 @@
 #include "rust-hir-path-probe.h"
 #include "rust-type-util.h"
 #include "rust-compile-implitem.h"
+#include "rust-unicode.h"
 
 #include "fold-const.h"
 #include "stringpool.h"
@@ -46,8 +47,9 @@ bool inline should_mangle_item (const tree fndecl)
 }
 
 void
-HIRCompileBase::setup_fndecl (tree fndecl, bool is_main_entry_point,
-			      bool is_generic_fn, HIR::Visibility &visibility,
+HIRCompileBase::setup_fndecl (tree fndecl, const Rust::Identifier &fn_name,
+			      bool is_main_entry_point, bool is_generic_fn,
+			      HIR::Visibility &visibility,
 			      const HIR::FunctionQualifiers &qualifiers,
 			      const AST::AttrVec &attrs)
 {
@@ -101,7 +103,7 @@ HIRCompileBase::setup_fndecl (tree fndecl, bool is_main_entry_point,
 	}
       else if (no_mangle)
 	{
-	  handle_no_mangle_attribute_on_fndecl (fndecl, attr);
+	  handle_no_mangle_attribute_on_fndecl (fndecl, attr, fn_name);
 	}
     }
 }
@@ -151,7 +153,7 @@ HIRCompileBase::handle_link_section_attribute_on_fndecl (
 
 void
 HIRCompileBase::handle_no_mangle_attribute_on_fndecl (
-  tree fndecl, const AST::Attribute &attr)
+  tree fndecl, const AST::Attribute &attr, const Rust::Identifier &fn_name)
 {
   if (attr.has_attr_input ())
     {
@@ -160,6 +162,13 @@ HIRCompileBase::handle_no_mangle_attribute_on_fndecl (
       return;
     }
 
+  if (!is_ascii_only (fn_name.as_string ()))
+    {
+      rust_error_at (fn_name.get_locus (), ErrorCode::E0754,
+		     "attribute %<no_mangle%> requires ASCII identifier");
+      return;
+    }
+
   DECL_ATTRIBUTES (fndecl) = tree_cons (get_identifier ("no_mangle"), NULL_TREE,
 					DECL_ATTRIBUTES (fndecl));
 }
@@ -534,7 +543,7 @@ HIRCompileBase::compile_function_body (tree fndecl,
 
 tree
 HIRCompileBase::compile_function (
-  const std::string &fn_name, HIR::SelfParam &self_param,
+  const Rust::Identifier &fn_name, HIR::SelfParam &self_param,
   std::vector<HIR::FunctionParam> &function_params,
   const HIR::FunctionQualifiers &qualifiers, HIR::Visibility &visibility,
   AST::AttrVec &outer_attrs, location_t locus, HIR::BlockExpr *function_body,
@@ -545,15 +554,16 @@ HIRCompileBase::compile_function (
     = canonical_path->get () + fntype->subst_as_string ();
 
   // we don't mangle the main fn since we haven't implemented the main shim
-  bool is_main_fn = fn_name.compare ("main") == 0;
-  std::string asm_name = fn_name;
+  bool is_main_fn = fn_name.as_string ().compare ("main") == 0;
+  std::string asm_name = fn_name.as_string ();
 
   unsigned int flags = 0;
   tree fndecl = ctx->get_backend ()->function (compiled_fn_type, ir_symbol_name,
 					       "" /* asm_name */, flags, locus);
 
-  setup_fndecl (fndecl, is_main_fn, fntype->has_subsititions_defined (),
-		visibility, qualifiers, outer_attrs);
+  setup_fndecl (fndecl, fn_name, is_main_fn,
+		fntype->has_subsititions_defined (), visibility, qualifiers,
+		outer_attrs);
   setup_abi_options (fndecl, qualifiers.get_abi ());
 
   // conditionally mangle the function name

diff --git a/gcc/rust/backend/rust-compile-base.h b/gcc/rust/backend/rust-compile-base.h
@@ -92,7 +92,8 @@ class HIRCompileBase
 			      const Resolver::CanonicalPath *canonical_path,
 			      HIR::Expr *const_value_expr, location_t locus);
 
-  tree compile_function (const std::string &fn_name, HIR::SelfParam &self_param,
+  tree compile_function (const Rust::Identifier &fn_name,
+			 HIR::SelfParam &self_param,
 			 std::vector<HIR::FunctionParam> &function_params,
 			 const HIR::FunctionQualifiers &qualifiers,
 			 HIR::Visibility &visibility, AST::AttrVec &outer_attrs,
@@ -102,8 +103,9 @@ class HIRCompileBase
 
   static tree unit_expression (Context *ctx, location_t locus);
 
-  static void setup_fndecl (tree fndecl, bool is_main_entry_point,
-			    bool is_generic_fn, HIR::Visibility &visibility,
+  static void setup_fndecl (tree fndecl, const Rust::Identifier &fn_name,
+			    bool is_main_entry_point, bool is_generic_fn,
+			    HIR::Visibility &visibility,
 			    const HIR::FunctionQualifiers &qualifiers,
 			    const AST::AttrVec &attrs);
 
@@ -123,8 +125,9 @@ class HIRCompileBase
   handle_deprecated_attribute_on_fndecl (tree fndecl,
 					 const AST::Attribute &attr);
 
-  static void handle_no_mangle_attribute_on_fndecl (tree fndecl,
-						    const AST::Attribute &attr);
+  static void
+  handle_no_mangle_attribute_on_fndecl (tree fndecl, const AST::Attribute &attr,
+					const Rust::Identifier &fn_name);
 
   static void setup_abi_options (tree fndecl, ABI abi);
 

diff --git a/gcc/rust/backend/rust-compile-implitem.cc b/gcc/rust/backend/rust-compile-implitem.cc
@@ -86,8 +86,8 @@ CompileTraitItem::visit (HIR::TraitItemFunc &func)
   auto vis = HIR::Visibility (HIR::Visibility::VisType::PUBLIC);
   HIR::TraitFunctionDecl &function = func.get_decl ();
   tree fndecl
-    = compile_function (function.get_function_name ().as_string (),
-			function.get_self (), function.get_function_params (),
+    = compile_function (function.get_function_name (), function.get_self (),
+			function.get_function_params (),
 			function.get_qualifiers (), vis,
 			func.get_outer_attrs (), func.get_locus (),
 			func.get_block_expr ().get (), canonical_path, fntype);

diff --git a/gcc/rust/backend/rust-compile-item.cc b/gcc/rust/backend/rust-compile-item.cc
@@ -178,7 +178,7 @@ CompileItem::visit (HIR::Function &function)
     ctx->push_const_context ();
 
   tree fndecl
-    = compile_function (function.get_function_name ().as_string (),
+    = compile_function (function.get_function_name (),
 			function.get_self_param (),
 			function.get_function_params (),
 			function.get_qualifiers (), function.get_visibility (),

diff --git a/gcc/rust/backend/rust-mangle.cc b/gcc/rust/backend/rust-mangle.cc
@@ -87,7 +87,7 @@ legacy_mangle_name (const std::string &name)
 	  i++;
 	  m = "..";
 	}
-      else if (c.value < 0x80)
+      else if (c.is_ascii ())
 	// ASCII
 	m.push_back (c.value);
       else

diff --git a/gcc/rust/lex/rust-input-source.h b/gcc/rust/lex/rust-input-source.h
@@ -28,9 +28,6 @@ constexpr uint8_t UTF8_BOM1 = 0xEF;
 constexpr uint8_t UTF8_BOM2 = 0xBB;
 constexpr uint8_t UTF8_BOM3 = 0xBF;
 
-constexpr uint32_t MAX_ASCII_CODEPOINT = 0x7F;
-constexpr uint32_t CODEPOINT_INVALID = 0xFFFE;
-
 // Input source wrapper thing.
 class InputSource
 {

diff --git a/gcc/rust/lex/rust-lex.cc b/gcc/rust/lex/rust-lex.cc
@@ -1734,7 +1734,7 @@ Lexer::parse_byte_char (location_t loc)
       // otherwise, get character from direct input character
       byte_char = current_char;
 
-      if (byte_char.value > 0x7f)
+      if (!byte_char.is_ascii ())
 	{
 	  rust_error_at (get_current_location (),
 			 "non-ASCII character in %<byte char%>");

diff --git a/gcc/rust/util/rust-codepoint.h b/gcc/rust/util/rust-codepoint.h
@@ -23,6 +23,9 @@
 
 namespace Rust {
 
+constexpr uint32_t MAX_ASCII_CODEPOINT = 0x7F;
+constexpr uint32_t CODEPOINT_INVALID = 0xFFFE;
+
 // FIXME: move this to rust-unicode.h?
 struct Codepoint
 {
@@ -36,6 +39,7 @@ struct Codepoint
 
   static Codepoint eof () { return Codepoint (UINT32_MAX); }
   bool is_eof () const { return value == UINT32_MAX; }
+  bool is_ascii () const { return value <= MAX_ASCII_CODEPOINT; }
 
   // Returns a C++ string containing string value of codepoint.
   std::string as_string ();

diff --git a/gcc/rust/util/rust-punycode.cc b/gcc/rust/util/rust-punycode.cc
@@ -42,7 +42,7 @@ extract_basic_string (const std::vector<Codepoint> &src)
   std::string basic_string;
   for (auto c : src)
     {
-      if (c.value <= 0x7F)
+      if (c.is_ascii ())
 	basic_string += c.as_string ();
     }
   return basic_string;

diff --git a/gcc/rust/util/rust-unicode.cc b/gcc/rust/util/rust-unicode.cc
@@ -16,6 +16,7 @@
 // along with GCC; see the file COPYING3.  If not see
 // <http://www.gnu.org/licenses/>.
 
+#include "rust-input-source.h"
 #include "rust-system.h"
 #include "optional.h"
 #include "selftest.h"
@@ -328,6 +329,15 @@ is_numeric (uint32_t codepoint)
     return true;
 }
 
+bool
+is_ascii_only (const std::string &str)
+{
+  for (char c : str)
+    if (static_cast<uint32_t> (c) > MAX_ASCII_CODEPOINT)
+      return false;
+  return true;
+}
+
 } // namespace Rust
 
 #if CHECKING_P

diff --git a/gcc/rust/util/rust-unicode.h b/gcc/rust/util/rust-unicode.h
@@ -62,6 +62,9 @@ class Utf8String
 bool
 is_alphabetic (uint32_t codepoint);
 
+bool
+is_ascii_only (const std::string &str);
+
 bool
 is_numeric (uint32_t codepoint);