This is the mail archive of the gcc-bugs@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]

[Bug c/12281] [3.4 Regression]not emiting function which is static and marked as inline


PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=12281



------- Additional Comments From pinskia at gcc dot gnu dot org  2003-09-22 02:29 -------
Here this the reason why f is not being marked in the first place:
Breakpoint 5, mark_referenced (id=0x40f95600) at /Volumes/UFS_Partition/pinskia/src/
fsf/gcc-clean/src/gcc/varasm.c:1656
1656              if (node)
(gdb) l
1651          struct cgraph_varpool_node *vnode;
1652
1653          if (!cgraph_global_info_ready)
1654            {
1655              node = cgraph_node_for_identifier (id);
1656              if (node)
1657                cgraph_mark_needed_node (node);
1658            }
1659
1660          vnode = cgraph_varpool_node_for_identifier (id);
(gdb) p debug_tree(id)
 <identifier_node 0x40f95600 &L_f$non_lazy_ptr>

See that in mark_referenced the identifer which it gets is the non-lazy pointer instead of 
the function it self.  There should be a way to get the function it self  from this non-lazy-
pointer identifier.


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