--- Additional Comments From amodra at bigpond dot net dot au 2008-02-16
00:08 ---
http://sourceware.org/ml/binutils/2008-02/msg00158.html
--
What|Removed |Added
--
What|Removed |Added
AssignedTo|unassigned at sources dot |amodra at bigpond dot net
|redhat dot com |dot au
Status|ASSIGNED
--
What|Removed |Added
Status|NEW |ASSIGNED
Last reconfirmed|-00-00 00:00:00 |2008-02-16 00:06:46
date|
--- Additional Comments From osv at javad dot com 2008-02-15 12:01 ---
Fortunately the problem has nothing to do with object files. It's the linker
command file that does matter, so here is a simple way to reproduce the
segfault:
$ cat linker.cmd
MEMORY
{
RAM : ORIGIN = 0, LENGT
--- Additional Comments From amodra at bigpond dot net dot au 2008-02-15
10:39 ---
http://sourceware.org/ml/binutils-cvs/2008-02/msg00085.html
http://sourceware.org/ml/binutils/2008-02/msg00155.html
--
What|Removed |Added
-
--- Additional Comments From nickc at redhat dot com 2008-02-15 10:25
---
Hi Chu,
Thanks for pointing that out. In fact the integer_arg() function is
completely redundant as the C library function strtoul() does everything that is
needed. So I have applied a revised version of the p
--- Additional Comments From nickc at redhat dot com 2008-02-15 10:23
---
Created an attachment (id=2269)
--> (http://sourceware.org/bugzilla/attachment.cgi?id=2269&action=view)
Revised patch which removes integer_arg()
--
What|Removed |Added
--
--- Additional Comments From amodra at bigpond dot net dot au 2008-02-15
10:13 ---
Indeed. Nicely spotted. It's been that way since 2005-02-21
--
What|Removed |Added
--- Additional Comments From chul at cn dot fujitsu dot com 2008-02-15
09:44 ---
Created an attachment (id=2268)
--> (http://sourceware.org/bugzilla/attachment.cgi?id=2268&action=view)
Used for "-n num" when num is hexadecimal or octal
--
http://sourceware.org/bugzilla/show_bug.cgi
--- Additional Comments From chul at cn dot fujitsu dot com 2008-02-15
09:41 ---
Hi,
I have tried it out. It' Ok now. And I found if the "num" is the hex number
in "-n num" option , the result is a little strange and 0xA can't be
recognized.
So I made a patch for it. Hope it is us
Hi!
Consider this code:
bfd/elflink.c:
. . .
void
_bfd_elf_section_already_linked (bfd *abfd, struct bfd_section *sec,
struct bfd_link_info *info)
{
. . .
switch (flags & SEC_LINK_DUPLICATES)
{
. . .
case SEC_LINK_DUPLICATES_DISC
11 matches
Mail list logo
