--- Additional Comments From nickc at redhat dot com 2008-05-30 16:15
---
Hi Frediano,
Well the patch makes sense - we are creating symbols for shared libraries, and
so the default ought to be to place the symbol in the dynamic symbol table
rather than the static symbol table. So I h
--- Additional Comments From frediano dot ziglio at vodafone dot com
2008-05-22 15:24 ---
M... in elf64-hppa.c I replaced line
owner = (h ? h->root.u.def.section->owner : dyn_h->owner);
(line 1148) with
owner = (!dyn_h->owner ? h->root.u.def.section->owner : dyn_h->owner);
and no
--- Additional Comments From frediano dot ziglio at vodafone dot com
2008-05-22 11:29 ---
>
> It does not core dump for me Can you send me the entire linker
> command line that you are using when the seg-fault occurs ?
>
Oh... I changed code adding a core if "invalid string of
--- Additional Comments From nickc at redhat dot com 2008-05-22 11:05
---
Subject: Re: segmentation fault building final object
Hi Frediano,
> Thanks for the patches. However ld still core :(
It does not core dump for me Can you send me the entire linker
command line that you
Hi Frediano,
Thanks for the patches. However ld still core :(
It does not core dump for me Can you send me the entire linker
command line that you are using when the seg-fault occurs ?
Cheers
Nick
___
bug-binutils mailing list
bug-binuti
--- Additional Comments From frediano dot ziglio at vodafone dot com
2008-05-22 10:05 ---
Thanks for the patches. However ld still core :(
I noted trying to debug for symoffset in bfd_elf_string_from_elf_section is
passed in a strange way... having ct.o and ctutil.o as input ld parse fir
--- Additional Comments From nickc at redhat dot com 2008-05-22 09:56
---
Hi Frediano,
Thanks for the object files. With them I was able to reproduce the behaviour
that you describe. I have uploaded two patches. The first, to bfd/elf.c, fixes
the problem with strange warning messag
--- Additional Comments From nickc at redhat dot com 2008-05-22 09:53
---
Created an attachment (id=2754)
--> (http://sourceware.org/bugzilla/attachment.cgi?id=2754&action=view)
Create the new dynamic symbol if it does not already exist
--
What|Removed
--- Additional Comments From nickc at redhat dot com 2008-05-22 09:52
---
Created an attachment (id=2753)
--> (http://sourceware.org/bugzilla/attachment.cgi?id=2753&action=view)
Stop bfd_elf_get_elf_syms from reading beyond the end of the section
--
http://sourceware.org/bugzilla/s
--- Additional Comments From frediano dot ziglio at vodafone dot com
2008-05-21 15:26 ---
Created an attachment (id=2749)
--> (http://sourceware.org/bugzilla/attachment.cgi?id=2749&action=view)
object files from my gcc version
result from hppa64-hp-hpux11.00-gcc -O2 -c ctutil.c && hppa
--- Additional Comments From nickc at redhat dot com 2008-05-21 15:17
---
Subject: Re: segmentation fault building final object
Hi Frediano,
> Can I send binary object files ?? They are quite small (about 1 kb)
Sure - put them in a tarball and I'll see what I can do.
Cheers
Nick
Hi Frediano,
Can I send binary object files ?? They are quite small (about 1 kb)
Sure - put them in a tarball and I'll see what I can do.
Cheers
Nick
___
bug-binutils mailing list
bug-binutils@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-
--- Additional Comments From frediano dot ziglio at vodafone dot com
2008-05-21 14:23 ---
I tried the patch. It just move the problem but it core after some while...
I tried to produce a test case but after three hours I got two c files that make
ld core only if compiled with some parame
--- Additional Comments From nickc at redhat dot com 2008-05-21 09:10
---
Hi Frediano,
It would be very helpful if you could submit a small test case that reproduces
the problem for us.
In the meantime I have uploaded a small patch which adds a check for a NULL
return from elf_link
--- Additional Comments From nickc at redhat dot com 2008-05-21 09:09
---
Created an attachment (id=2746)
--> (http://sourceware.org/bugzilla/attachment.cgi?id=2746&action=view)
Check for a NULL return from elf_link_hash_lookup
--
http://sourceware.org/bugzilla/show_bug.cgi?id=6511
15 matches
Mail list logo
