--- Additional Comments From vincent dot riviere at freesbee dot fr
2010-06-22 18:47 ---
Hello, Nick.
I don't really use the m68k-elf target myself, I discovered this bug by
running the testsuite. I often use an a.out toolchain to produce srec binaries
directly from .o files, it
--- Additional Comments From vincent dot riviere at freesbee dot fr
2010-06-17 18:08 ---
Thanks, it works.
Of course, I had included my.exp by mistake.
--
http://sourceware.org/bugzilla/show_bug.cgi?id=11673
--- You are receiving this mail because: ---
You are on the CC
--- Additional Comments From vincent dot riviere at freesbee dot fr
2010-06-17 18:04 ---
Thanks Nick, it works.
--
http://sourceware.org/bugzilla/show_bug.cgi?id=11676
--- You are receiving this mail because: ---
You are on the CC list for the bug, or are watching someone
--- Additional Comments From vincent dot riviere at freesbee dot fr
2010-06-17 18:01 ---
I can reproduce the bug again very easily.
My host is Cygwin, I get explicitly a Segmentation fault when linking.
Here is a very simple testcase:
$ cat a.s
nop
$ gas/as-new a.s -o a.o
--- Additional Comments From vincent dot riviere at freesbee dot fr
2010-06-07 22:46 ---
Created an attachment (id=4831)
--> (http://sourceware.org/bugzilla/attachment.cgi?id=4831&action=view)
The fix
This patch fixes the problem.
I have not included a testcase because
b
oduct: binutils
Version: 2.21 (HEAD)
Status: NEW
Severity: minor
Priority: P3
Component: binutils
AssignedTo: unassigned at sources dot redhat dot com
ReportedBy: vincent dot riviere at freesbee dot fr
CC: bug-binuti
ummary: -- oformat binary or srec broken
Product: binutils
Version: 2.21 (HEAD)
Status: NEW
Severity: critical
Priority: P1
Component: ld
AssignedTo: unassigned at sources dot redhat dot com
ReportedBy: vincent dot rivi
--- Additional Comments From vincent dot riviere at freesbee dot fr
2010-06-07 21:18 ---
This is not a bug, you used the wrong syntax.
When you write:
fmove.s #1,fp0
the argument is the binary pattern of the single precision value, and it is
not what you expect.
For example
--- Additional Comments From vincent dot riviere at freesbee dot fr
2010-06-07 20:41 ---
Created an attachment (id=4830)
--> (http://sourceware.org/bugzilla/attachment.cgi?id=4830&action=view)
Testcase and proposed patch
Here is a new testcase, as well as a proposed patch.
I
: unassigned at sources dot redhat dot com
ReportedBy: vincent dot riviere at freesbee dot fr
CC: bug-binutils at gnu dot org
GCC target triplet: m68k-*-*
http://sourceware.org/bugzilla/show_bug.cgi?id=11673
--- You are receiving this mail because: ---
You are on the CC
--- Additional Comments From vincent dot riviere at freesbee dot fr
2010-05-13 16:38 ---
The code for checking this error condition is in gas/config/tc-m68k.c, at the
end of md_apply_fix(). The error is not detected on a.out targets because:
fixP->fx_pcrel == 0 and fixP->fx_
Priority: P3
Component: gas
AssignedTo: unassigned at sources dot redhat dot com
ReportedBy: vincent dot riviere at freesbee dot fr
CC: bug-binutils at gnu dot org
GCC target triplet: m68k-netbsd
http://sourceware.org/bugzilla/show_bug.cgi?id=11597
--- You
--
What|Removed |Added
CC||vincent dot riviere at
||freesbee dot fr
http://sourceware.org
--- Additional Comments From vincent dot riviere at freesbee dot fr
2009-10-16 20:58 ---
The new testcase copy-4 fails on target m68k-netbsd (using a.out).
The test seems to be valid, maybe the fix didn't work for a.out objects.
--
http://sourceware.org/bugzilla/show_bug.cgi?id
--- Additional Comments From vincent dot riviere at freesbee dot fr
2009-10-13 23:54 ---
Yes, I believe other targets using a.out for other processors are affected,
too, but I cowardly fixed the bug only for m68k.
It would probably be more clean to fix bfd_install_relocation(), but
--- Additional Comments From vincent dot riviere at freesbee dot fr
2009-10-13 19:41 ---
Many thanks, Nick.
I see it is too late for 2.20.
--
http://sourceware.org/bugzilla/show_bug.cgi?id=3041
--- You are receiving this mail because: ---
You are on the CC list for the bug
--- Additional Comments From vincent dot riviere at freesbee dot fr
2009-10-07 20:11 ---
Here are proposals for the ChangeLog.
gas:
2009-10-07 Vincent Riviere
PR gas/3041
* config/tc-m68k.c (tc_gen_reloc): Fix addend for relocations
located in data
--- Additional Comments From vincent dot riviere at freesbee dot fr
2009-10-03 16:38 ---
Created an attachment (id=4250)
--> (http://sourceware.org/bugzilla/attachment.cgi?id=4250&action=view)
Fix addend for weak references to data symbols
The attached patch fixes this pro
--- Additional Comments From vincent dot riviere at freesbee dot fr
2009-10-03 16:31 ---
The resulting addend is still wrong when the referenced symbol is defined as
weak in the .data or .bss section.
This happens in binutils 2.19.1 and the latest 2.20 from CVS.
--
What
--- Additional Comments From vincent dot riviere at freesbee dot fr
2009-03-18 21:55 ---
This bug may be related the the already fixed Bug #3041 about weak symbols on
a.out targets.
--
http://sourceware.org/bugzilla/show_bug.cgi?id=977
--- You are receiving this mail because
--- Additional Comments From vincent dot riviere at freesbee dot fr
2008-08-17 12:28 ---
I confirm everything is fixed in the current CVS version.
Many thanks for your great job, Alan.
--
http://sourceware.org/bugzilla/show_bug.cgi?id=6478
--- You are receiving this mail
--- Additional Comments From vincent dot riviere at freesbee dot fr
2008-08-15 12:37 ---
I confirm the previous patch applied to binutils 2.18 fixes the testcase, as
well as my original real-world problem. Furthermore, the linker warnings
present in a.out object files are displayed as
--- Additional Comments From vincent dot riviere at freesbee dot fr
2008-05-06 20:52 ---
Hi, Nick.
I agree with you on the 2 points.
1) Shame on me, there is effectively a typo near the end, it shoud be read:
Further strangeness comes if we convert b.o into ELF, too.
$ objcopy b.o b
t: binutils
Version: 2.18
Status: NEW
Severity: normal
Priority: P2
Component: ld
AssignedTo: unassigned at sources dot redhat dot com
ReportedBy: vincent dot riviere at freesbee dot fr
CC: bug-binutils at gn
--- Additional Comments From vincent dot riviere at freesbee dot fr
2008-01-26 19:28 ---
Sorry, there is a typo in the comments:
accruate instead of accurate
--
http://sourceware.org/bugzilla/show_bug.cgi?id=5652
--- You are receiving this mail because: ---
You are on the
--- Additional Comments From vincent dot riviere at freesbee dot fr
2008-01-25 22:55 ---
Furthermore:
- SVR2 (1984):
- shell functions (sh)
http://www.faqs.org/faqs/unix-faq/faq/part6/
I didn't find any documentation about POSIX. But bash --posix knows about
functions
--- Additional Comments From vincent dot riviere at freesbee dot fr
2008-01-21 18:39 ---
Created an attachment (id=2208)
--> (http://sourceware.org/bugzilla/attachment.cgi?id=2208&action=view)
genscripts fix
This patch fixes the bug. The availability of the BASH_LINENO variable
with BASH_LINENO
Product: binutils
Version: 2.19 (HEAD)
Status: NEW
Severity: normal
Priority: P2
Component: ld
AssignedTo: unassigned at sources dot redhat dot com
ReportedBy: vincent dot riviere at freesbee dot fr
--- Additional Comments From vincent dot riviere at freesbee dot fr
2007-08-01 19:44 ---
Great, it works !
Thank you, Alan.
--
http://sourceware.org/bugzilla/show_bug.cgi?id=4694
--- You are receiving this mail because: ---
You are on the CC list for the bug, or are
(HEAD)
Status: NEW
Severity: normal
Priority: P2
Component: ld
AssignedTo: unassigned at sources dot redhat dot com
ReportedBy: vincent dot riviere at freesbee dot fr
CC: bug-binutils at gnu dot org
GCC target triplet: m68k-netbsd
--- Additional Comments From vincent dot riviere at freesbee dot fr
2007-06-05 19:27 ---
Hi, Nick !
> "igncr trick" ? What "igncr trick" ?
> So we would have to be a bit more explicit.
Of course, you're right. My little script was just an example t
--- Additional Comments From vincent dot riviere at freesbee dot fr
2007-05-30 21:41 ---
My previous solution is not so good, because the main configure is run without
the option igncr. And configure.in contains the line :
baseargs=`cat conftestsed.out`
Thus the generated Makefile
--- Additional Comments From vincent dot riviere at freesbee dot fr
2007-05-28 18:24 ---
More ideas (again) :
6)
libbfd_la_LIBADD = $(shell cat ofiles) @WIN32LIBADD@
This one works with every shell, but only with GNU Make.
http://www.gnu.org/software/make/manual/html_node/Shell
--- Additional Comments From vincent dot riviere at freesbee dot fr
2007-05-23 19:57 ---
Sorry, it doesn't work. The filesystem is always considered in binmode.
This is because you forgot the brackets in the awk regular expression !
Here is the fixed line :
build_dir_mount_poin
--- Additional Comments From vincent dot riviere at freesbee dot fr
2007-05-22 20:57 ---
Fixing this "bug" may not be a good idea, because the build fails later for
another reason :
gcc ... readelf.c
../bfd/bfd.h:103:3: #error No 64 bit integer type available
There is also a
--- Additional Comments From vincent dot riviere at freesbee dot fr
2007-05-22 20:46 ---
Hi.
Yes, the "expand" trick solves the problem described here.
The last object file listed in ofiles is successfully added to libbfd.a
Here is a small testcase showing the proble
--
What|Removed |Added
CC||vincent dot riviere at
||freesbee dot fr
http://sourceware.org
--- Additional Comments From vincent dot riviere at freesbee dot fr
2007-05-19 16:12 ---
More ideas :
4)
libbfd_la_LIBADD = `more ofiles` @WIN32LIBADD@
5)
libbfd_la_LIBADD = `expand ofiles` @WIN32LIBADD@
--
http://sourceware.org/bugzilla/show_bug.cgi?id=4334
--- You are
--- Additional Comments From vincent dot riviere at freesbee dot fr
2007-05-19 14:02 ---
The problem is in the file bfd/Makefile.am, line 748.
libbfd_la_LIBADD = `cat ofiles` @WIN32LIBADD@
By default, cat opens the files in binary mode, which is not acceptable here
because of the
--- Additional Comments From vincent dot riviere at freesbee dot fr
2007-05-17 16:34 ---
After looking at this a bit closer, I think these 4 failures are unrelated to
my patch. Furthermore, I think the actual result is good - atleast for m68k-*-
netbsd. But it should be checked by a
--- Additional Comments From vincent dot riviere at freesbee dot fr
2007-05-15 21:44 ---
Great !
Thank you, Nick.
My testcase passes.
But some tests fails :
1) gas/all/weakref1u.d
It seems that this test should not be run on a.out targets.
I propose to not-target it for m68k-*-netbsd
--- Additional Comments From vincent dot riviere at freesbee dot fr
2007-05-08 14:37 ---
Reopen the bug due to the previous comment.
--
What|Removed |Added
--- Additional Comments From vincent dot riviere at freesbee dot fr
2007-05-08 14:35 ---
Created an attachment (id=1778)
--> (http://sourceware.org/bugzilla/attachment.cgi?id=1778&action=view)
Write correct offsets into a.out object files
There is still a bug when the reference
--- Additional Comments From vincent dot riviere at freesbee dot fr
2007-05-03 20:04 ---
Great !
It works as expected.
Even the linker does its job when linking with another object containing a
strong symbol with the same name.
Thank you Nick !
--
http://sourceware.org/bugzilla
--- Additional Comments From vincent dot riviere at freesbee dot fr
2007-04-28 21:38 ---
Created an attachment (id=1739)
--> (http://sourceware.org/bugzilla/attachment.cgi?id=1739&action=view)
Write correct values and relocs into a.out object files
After spending *weeks*
--- Additional Comments From vincent dot riviere at freesbee dot fr
2007-04-09 21:59 ---
I found an allusion to a weak-bug in bfd/aout-cris.c
In MY (swap_ext_reloc_out) :
...
if (bfd_is_und_section (bfd_get_section (sym))
/* Remember to check for weak symbols; they count
--- Additional Comments From vincent dot riviere at freesbee dot fr
2006-09-09 10:31 ---
A lot better this time !
But half-fixed.
$ as bug.s -o bug.o
$ objdump -d bug.o
bug.o: file format a.out-m68k-netbsd
Disassembly of section .text:
:
0: 4ef9 0010 jmp 10
--- Additional Comments From vincent dot riviere at freesbee dot fr
2006-09-08 20:46 ---
Sorry, your fix does not work because... it is not compiled !!!
It is inside a #ifdef OBJ_ELF block, but the m68k-unknown-netbsd target uses
the a.out format.
Furthermore, I think you forgot to
s dot redhat dot com
ReportedBy: vincent dot riviere at freesbee dot fr
CC: bug-binutils at gnu dot org
GCC build triplet: i686-pc-linux-gnu
GCC host triplet: i686-pc-linux-gnu
GCC target triplet: m68k-unknown-netbsd
http://sourceware.org/bugzilla/show_bug.cgi?id=3041
--- Yo
--- Additional Comments From vincent dot riviere at freesbee dot fr
2006-08-03 18:57 ---
Great !
It works perfectly !
Many thanks !
--
What|Removed |Added
--- Additional Comments From vincent dot riviere at freesbee dot fr
2006-08-02 21:03 ---
Sorry, it doesn't work.
Did you forget autoheader ?
--
What|Removed |
signed at sources dot redhat dot com
ReportedBy: vincent dot riviere at freesbee dot fr
CC: bug-binutils at gnu dot org
GCC build triplet: i686-pc-cygwin
GCC host triplet: i686-pc-cygwin
GCC target triplet: i686-pc-cygwin
http://sourceware.org/bugzilla/show_bug.c
--- Additional Comments From vincent dot riviere at freesbee dot fr
2006-07-22 15:00 ---
Here are detailed tests.
# target i686-pc-cygwin
# gas 2.16.91 20050610 on host Windows XP SP2
--- Additional Comments From vincent dot riviere at freesbee dot fr
2006-07-22 12:52 ---
Sorry, you're right.
The jump is relocated.
But I'm very surprised.
Without the "weak" keyword, gas generates a relative jump and no relocation
entry. Good.
With the
urces dot redhat dot com
ReportedBy: vincent dot riviere at freesbee dot fr
CC: bug-binutils at gnu dot org
GCC build triplet: all
GCC host triplet: all
GCC target triplet: i?86-linux, m68k-*-*
http://sourceware.org/bugzilla/show_bug.cgi?id=2946
--- You are receiving
55 matches
Mail list logo
