When binutils is configured with --target=i386-aesec-coff
--enabled-target=i386-elf (or i586-linux or variations thereon), and
you run "gas/as-new -v"
you get the version string, followed by a segmentation fault
Configuration command:
../binutils-2.16.1/configure --target=i386-coff --enable-targets=i586-linux
and build the system, then run the new gas with -v:
[EMAIL PROTECTED]:/d1/box/obj-binutils-current> gas/as-new -v
GNU assembler version 2.16.1 (i386-coff) using BFD version 2.16.1
Segmentation fault
Looking through gdb:
[EMAIL PROTECTED]:/d1/box/obj-binutils-current> gdb gas/as-new
GNU gdb 6.3
Copyright 2004 Free Software Foundation, Inc.
GDB is free software, covered by the GNU General Public License, and you are
welcome to change it and/or distribute copies of it under certain conditions.
Type "show copying" to see the conditions.
There is absolutely no warranty for GDB. Type "show warranty" for details.
This GDB was configured as "i586-suse-linux"...Using host libthread_db library
"/lib/tls/libthread_db.so.1".
(gdb) run -v
Starting program: /d1/box/obj-binutils-2.16.1/gas/as-new -v
GNU assembler version 2.16.1 (i386-coff) using BFD version 2.16.1
Program received signal SIGSEGV, Segmentation fault.
coff_new_section_hook (abfd=0x80ed3e8, section=0x80d56a0) at coffcode.h:1602
1602 coffsymbol (section->symbol)->native = native;
(gdb) print (section->symbol)
$1 = (struct bfd_symbol *) 0x80ce550
(gdb) print *(section->symbol)
$2 = {the_bfd = 0x0, name = 0x80b5370 "*ABS*", value = 0, flags = 256, section =
0x80d56a0, udata = {p = 0x0, i = 0}}
(gdb)
By stepping through the debugger prior to the segfault, I can see the .text and
.data sections passing this line of code with no issues. But the second time it
is called with *ABS* section, it horks.
I'm guessing that as is trying to setup section headers and symbol tables
getting ready for source code input, and trying to set up ELF-things, which is
causing some consternation in coffcode.h.
I haven't a clue as to where to tell as to not bother trying to set up ELF
section headers for the default COFF output target.
I've tried this on binutils versions 2.15, 2.16 and 2.16.1
It's annoying. I'd like to have binutils assume COFF, but be able to handle
ELF, and this is the only glitch I've found, so far.
Thanks,
Ed
--
Summary: GAS segmentation fault when target=coff and enabled-
target=elf
Product: binutils
Version: 2.16
Status: NEW
Severity: normal
Priority: P2
Component: gas
AssignedTo: unassigned at sources dot redhat dot com
ReportedBy: ed dot reed at aesec dot com
CC: bug-binutils at gnu dot org,ed dot reed at aesec dot com
GCC build triplet: i586-suse-linux
GCC host triplet: i586-suse-linux
GCC target triplet: i386-coff
http://sourceware.org/bugzilla/show_bug.cgi?id=2724
------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.
_______________________________________________
bug-binutils mailing list
bug-binutils@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-binutils
