This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Revert my change freeing lto_global_var_decls


Hi,
I broke LTO fortran common blocks with this change.  The problem is that lto_global_var_decls
are actually used once again from wrapup_global_vars langhook and correctly freed afterwards.
This should be cleaned up, sicne all global vars should be wrapped up before lto streaming IMO,
but for a moment I am reverting the chagne, it makes no sense.

Bootstrapped/regtested x86_64-linux, committed.
Honza

Index: ChangeLog
===================================================================
--- ChangeLog	(revision 159304)
+++ ChangeLog	(working copy)
@@ -1,3 +1,7 @@
+2010-05-11  Jan Hubicka  <jh@suse.cz>
+
+	* lto.c (materialize_cgraph): Revert my previous patch.
+
 2010-05-11  Kai Tietz  <kai.tietz@onevision.com>
 
 	* lto-coff.c (IMAGE_FILE_MACHINE_ADM64): Rename to
Index: lto.c
===================================================================
--- lto.c	(revision 159304)
+++ lto.c	(working copy)
@@ -1708,9 +1708,6 @@ materialize_cgraph (void)
   for (i = 0; VEC_iterate (tree, lto_global_var_decls, i, decl); i++)
     rest_of_decl_compilation (decl, 1, 0);
 
-  VEC_free (tree, gc, lto_global_var_decls);
-  lto_global_var_decls = NULL;
-
   if (!quiet_flag)
     fprintf (stderr, "\n");
 


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]