[PATCH] Fix memory leak in Fortran FE with init_emit

Jakub Jelinek jakub@redhat.com
Wed Aug 22 13:09:00 GMT 2012


Hi!

init_function_start is these days supposed to be called only at the start of
RTL expansion, it shouldn't be called much earlier, because then we leak
e.g. the memory allocated by init_emit.

Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?

2012-08-22  Jakub Jelinek  <jakub@redhat.com>

	* trans-decl.c (trans_function_start, generate_coarray_init,
	create_main_function, gfc_generate_constructors): Call
	allocate_struct_function instead of init_function_start.

--- gcc/fortran/trans-decl.c.jj	2012-08-13 14:36:51.000000000 +0200
+++ gcc/fortran/trans-decl.c	2012-08-22 12:18:29.963953958 +0200
@@ -2265,7 +2265,8 @@ trans_function_start (gfc_symbol * sym)
   /* Create RTL for function definition.  */
   make_decl_rtl (fndecl);
 
-  init_function_start (fndecl);
+  gcc_checking_assert (DECL_STRUCT_FUNCTION (fndecl) == NULL);
+  allocate_struct_function (fndecl, false);
 
   /* function.c requires a push at the start of the function.  */
   pushlevel ();
@@ -4408,7 +4409,8 @@ generate_coarray_init (gfc_namespace * n
 
   rest_of_decl_compilation (fndecl, 0, 0);
   make_decl_rtl (fndecl);
-  init_function_start (fndecl);
+  gcc_checking_assert (DECL_STRUCT_FUNCTION (fndecl) == NULL);
+  allocate_struct_function (fndecl, false);
 
   pushlevel ();
   gfc_init_block (&caf_init_block);
@@ -4982,7 +4984,8 @@ create_main_function (tree fndecl)
 
   rest_of_decl_compilation (ftn_main, 1, 0);
   make_decl_rtl (ftn_main);
-  init_function_start (ftn_main);
+  gcc_checking_assert (DECL_STRUCT_FUNCTION (ftn_main) == NULL);
+  allocate_struct_function (ftn_main, false);
   pushlevel ();
 
   gfc_init_block (&body);
@@ -5537,7 +5540,8 @@ gfc_generate_constructors (void)
 
   make_decl_rtl (fndecl);
 
-  init_function_start (fndecl);
+  gcc_checking_assert (DECL_STRUCT_FUNCTION (fndecl) == NULL);
+  allocate_struct_function (fndecl, false);
 
   pushlevel ();
 

	Jakub



More information about the Fortran mailing list