--- Additional Comments From amodra at bigpond dot net dot au 2009-12-22
02:10 ---
This must be a host issue because I can certainly cross-compile a
powerpc-wrs-vxworks binutils-2.20. I suspect your shell may not like the shell
commands near the end of vxworks.em. Building under bash
--- Additional Comments From amodra at bigpond dot net dot au 2009-12-22
02:13 ---
Created an attachment (id=4479)
-- (http://sourceware.org/bugzilla/attachment.cgi?id=4479action=view)
eelf32ppcwxworks.c as it should be
--
http://sourceware.org/bugzilla/show_bug.cgi?id=11107
--- Additional Comments From amodra at bigpond dot net dot au 2009-12-19
10:58 ---
This has already been fixed. Update to 2.20 or a recent snapshot.
--
What|Removed |Added
--- Additional Comments From amodra at bigpond dot net dot au 2009-12-16
23:34 ---
Found it. The symbol in question is hidden. check_relocs records a need for a
plt entry ok, but then we hit code after this comment in elflink.c
/* If the symbol already has a dynamic index
--- Additional Comments From amodra at bigpond dot net dot au 2009-12-16
00:12 ---
HJ, please apply your patch to the 2.20 branch so the bug can be closed.
--
http://sourceware.org/bugzilla/show_bug.cgi?id=11037
--- You are receiving this mail because: ---
You are on the CC
--- Additional Comments From amodra at bigpond dot net dot au 2009-12-14
23:40 ---
I'd like to know why we have the failure before you apply your patch, Nick.
Especially since there's a similar ppc32 report but no testcase. In looking
over the ppc32 code I found a number of problems
--
What|Removed |Added
AssignedTo|unassigned at sources dot |amodra at bigpond dot net
|redhat dot com |dot au
Status|NEW
--- Additional Comments From amodra at bigpond dot net dot au 2009-12-09
01:29 ---
This fixes the ifunc prelink test failures. Two prelink tests still fail
FAIL: deps1.sh
FAIL: deps2.sh
I haven't yet investigated them.
--
What|Removed |Added
--- Additional Comments From amodra at bigpond dot net dot au 2009-12-08
01:23 ---
Ah ha! I wanted to get rid of seg_left == undefined_section too, but hadn't
spotted the problem with local symbols. Your patch looks good except for the
calls to local_symbol_convert, which
--- Additional Comments From amodra at bigpond dot net dot au 2009-12-06
02:50 ---
Please look at your config.log, recreate conftest.c from the log, and run the
command that failed but add -v to the command line. This will tell us the
version of the assembler being run.
--
http
--- Additional Comments From amodra at bigpond dot net dot au 2009-12-03
08:33 ---
.
--
What|Removed |Added
Status|ASSIGNED|RESOLVED
--- Additional Comments From amodra at bigpond dot net dot au 2009-12-03
23:39 ---
Thanks for reporting this but the bug has already been fixed.
--
What|Removed |Added
--
What|Removed |Added
GCC target triplet||i486-*
http://sourceware.org/bugzilla/show_bug.cgi?id=11037
--- You are receiving this mail because:
--
What|Removed |Added
Status|NEW |ASSIGNED
Last reconfirmed|-00-00 00:00:00 |2009-12-03 05:15:40
date|
--
What|Removed |Added
AssignedTo|unassigned at sources dot |amodra at bigpond dot net
|redhat dot com |dot au
Status|NEW
--- Additional Comments From amodra at bigpond dot net dot au 2009-11-30
00:33 ---
The patch I've just committed should allow you to implement this behaviour with
a simple wrapper script, something like the following:
#! /bin/sh
exec real_ld --start-group $@ --end-group
You could have
--- Additional Comments From amodra at bigpond dot net dot au 2009-11-26
06:35 ---
No there isn't a chance of reverting this fix. Having ld search in -L paths for
scripts was a bad idea. It has resulted in a number of very confused users over
the years, who installed a new version
--
What|Removed |Added
Component|binutils|gold
http://sourceware.org/bugzilla/show_bug.cgi?id=10916
--- You are receiving this mail because:
--- Additional Comments From amodra at bigpond dot net dot au 2009-11-04
02:06 ---
Created an attachment (id=4356)
-- (http://sourceware.org/bugzilla/attachment.cgi?id=4356action=view)
likely fix
I believe this patch will fix your problem, but I don't have mips machines and
toolchains
--- Additional Comments From amodra at bigpond dot net dot au 2009-11-01
12:25 ---
Only if you pick that patch without the one that goes with it
http://sources.redhat.com/cgi-bin/cvsweb.cgi/src/gas/as.h.diff?cvsroot=srcr1=1.64r2=1.65
--
What|Removed
--- Additional Comments From amodra at bigpond dot net dot au 2009-11-01
13:19 ---
Yes, binutile-2.20 went out with this bug. It has already been fixed on the
2.20 branch and on mainline. (The patch you referenced on mainline was part of
the *fix* for the bug, so was a little
--- Additional Comments From amodra at bigpond dot net dot au 2009-10-28
10:08 ---
.
--
What|Removed |Added
Status|ASSIGNED|RESOLVED
--- Additional Comments From amodra at bigpond dot net dot au 2009-10-20
00:52 ---
I'll put this on the branch after verifying that testsuite results are not
perturbed.
--
What|Removed |Added
--- Additional Comments From amodra at bigpond dot net dot au 2009-10-18
23:15 ---
.
--
What|Removed |Added
Status|NEW |RESOLVED
--- Additional Comments From amodra at bigpond dot net dot au 2009-10-15
11:04 ---
Fixed, almost exactly four years from when the bug was opened.
--
What|Removed |Added
--
What|Removed |Added
Component|admin |binutils
http://sourceware.org/bugzilla/show_bug.cgi?id=10767
--- You are receiving this mail because:
--- Additional Comments From amodra at bigpond dot net dot au 2009-10-16
01:52 ---
I could not reproduce this problem. I tried with many different versions of ar.
--
http://sourceware.org/bugzilla/show_bug.cgi?id=10773
--- You are receiving this mail because: ---
You
--- Additional Comments From amodra at bigpond dot net dot au 2009-10-14
11:32 ---
Patch applied. Apologies for the horrendous delay. Ben, how many other bugs
are you sitting on??
--
What|Removed |Added
--
What|Removed |Added
Status|ASSIGNED|NEW
http://sourceware.org/bugzilla/show_bug.cgi?id=2613
--- You are receiving this mail because: ---
--
What|Removed |Added
Status|ASSIGNED|NEW
http://sourceware.org/bugzilla/show_bug.cgi?id=5136
--- You are receiving this mail because: ---
--- Additional Comments From amodra at bigpond dot net dot au 2009-10-14
11:45 ---
Closing as reporter says bug in migw runtime
--
What|Removed |Added
Status
--
What|Removed |Added
Status|ASSIGNED|NEW
http://sourceware.org/bugzilla/show_bug.cgi?id=6906
--- You are receiving this mail because: ---
--
What|Removed |Added
Status|ASSIGNED|NEW
http://sourceware.org/bugzilla/show_bug.cgi?id=10767
--- You are receiving this mail because:
--
What|Removed |Added
Status|ASSIGNED|NEW
http://sourceware.org/bugzilla/show_bug.cgi?id=3990
--- You are receiving this mail because: ---
--
What|Removed |Added
Status|ASSIGNED|NEW
http://sourceware.org/bugzilla/show_bug.cgi?id=1038
--- You are receiving this mail because: ---
--
What|Removed |Added
AssignedTo|ramana dot radhakrishnan at |unassigned at sources dot
|codito dot com |redhat dot com
--- Additional Comments From amodra at bigpond dot net dot au 2009-10-14
14:28 ---
My guess is that this has been fixed in 2.20. Closing anyway as no feedback in
4 years
--
What|Removed |Added
--
What|Removed |Added
Status|WAITING |NEW
Last reconfirmed|-00-00 00:00:00 |2009-10-14 14:30:44
date|
--- Additional Comments From amodra at bigpond dot net dot au 2009-10-15
01:19 ---
Looking into it. Caused by LABELS_WITHOUT_COLONS
--
What|Removed |Added
--- Additional Comments From amodra at bigpond dot net dot au 2009-10-13
23:47 ---
This patch probably is relevant for other targets too. I can't ask you to look
at them as I fully sympathize with your comment #8 regarding time spent in the
debugger.
--
http://sourceware.org
--- Additional Comments From amodra at bigpond dot net dot au 2009-10-12
23:25 ---
Fixed
--
What|Removed |Added
Status|NEW |RESOLVED
--- Additional Comments From amodra at bigpond dot net dot au 2009-10-10
23:12 ---
I guess the hppa test failure is due to hppaelf_after_parse
if (link_info.relocatable)
lang_add_unique (.text);
I'd be inclined just to xfail the test for hppa.
--
What|Removed
--- Additional Comments From amodra at bigpond dot net dot au 2009-10-11
01:55 ---
Exactly. The reason is that .text is made a unique section, so doesn't match
the .text in the script. See unique_section_p, in particular the one in
output_section_callback. So .text is treated
--- Additional Comments From amodra at bigpond dot net dot au 2009-10-11
02:05 ---
Created an attachment (id=4265)
-- (http://sourceware.org/bugzilla/attachment.cgi?id=4265action=view)
ld --unique tweak
It might be reasonable to change the behaviour of unique sections so
--- Additional Comments From amodra at bigpond dot net dot au 2009-10-11
02:26 ---
Sorry, I shouldn't have said .text is treated as an orphan. That isn't quite
correct. .text obviously isn't an orphan since .text is mentioned in the linker
script, so we don't hit the /DISCARD/. .text
--- Additional Comments From amodra at bigpond dot net dot au 2009-10-09
01:28 ---
Do you have a good and a bad kernel image for me to look at?
--
http://sourceware.org/bugzilla/show_bug.cgi?id=10749
--- You are receiving this mail because: ---
You are on the CC list
--- Additional Comments From amodra at bigpond dot net dot au 2009-10-07
11:43 ---
Doesn't segfault for me with current mainline or branch.
--
http://sourceware.org/bugzilla/show_bug.cgi?id=10704
--- You are receiving this mail because: ---
You are on the CC list
--- Additional Comments From amodra at bigpond dot net dot au 2009-10-08
01:22 ---
patch applied
--
What|Removed |Added
Severity|critical
--- Additional Comments From amodra at bigpond dot net dot au 2009-09-30
22:38 ---
Not a bug. ld only extracts an object from an archive if the object satisfies
undefined symbols at the point where ld searches the archive. You linked
test_e.o *after* the archive.
--
What
--- Additional Comments From amodra at bigpond dot net dot au 2009-09-19
02:22 ---
I've committed some patches that may stop the segfault. Please try with the
latest sources. (At the moment that means CVS, the snapshot doesn't yet have my
patch).
You'll get -g by default but also get
--- Additional Comments From amodra at bigpond dot net dot au 2009-09-14
07:12 ---
Yes, the error may well only occur when ld is linking non-gcc object files.
Regardless, ld should not segfault. Can you build binutils with debug info and
provide a backtrace?
--
http
--- Additional Comments From amodra at bigpond dot net dot au 2009-09-01
11:04 ---
I couldn't reproduce this with either 2.19.1 or current cvs binutils. Using
2.19.1 gas gives me an object file with just one byte different, the addend in
the reloc you show as being wrong. I get
--- Additional Comments From amodra at bigpond dot net dot au 2009-08-27
13:58 ---
Fixed. See http://sourceware.org/bugzilla/show_bug.cgi?id=10515 for the patch
commit log
--
What|Removed |Added
--
What|Removed |Added
AssignedTo|unassigned at sources dot |amodra at bigpond dot net
|redhat dot com |dot au
Status|NEW
--- Additional Comments From amodra at bigpond dot net dot au 2009-08-14
03:45 ---
libxul.so has been built from non-pic objects.
--
What|Removed |Added
Status
--- Additional Comments From amodra at bigpond dot net dot au 2009-08-14
04:07 ---
The linker script shown has two global wildcard matches. As the ld info doc
says, this is slightly crazy. They both match test_new, so which one should
ld choose? The most specific match? How do you
--- Additional Comments From amodra at bigpond dot net dot au 2009-08-12
23:32 ---
.
--
What|Removed |Added
Status|ASSIGNED|RESOLVED
--- Additional Comments From amodra at bigpond dot net dot au 2009-08-08
08:06 ---
You are using the old-style plt and got rather than the new secure plt (forced
by old startup files even if all the other code is capable of using the new
secure plt). The old-style plt is a bss section
--- Additional Comments From amodra at bigpond dot net dot au 2009-08-07
09:27 ---
That's one weird looking ELF image. The root of the problem is in
_bfd_elf_make_section_from_shdr, where we set up section LMAs. .whiteboard1 is
incorrectly assumed to belong with the first PT_LOAD
--- Additional Comments From amodra at bigpond dot net dot au 2009-08-07
10:06 ---
Thanks, I'm now convinced you are using a new ld. :) Hmm, the -v output also
shows that gcc is passing --relax to ld anyway, so the automatic --relax code
isn't even a factor. I don't know what else
--- Additional Comments From amodra at bigpond dot net dot au 2009-08-07
23:57 ---
$ wget http://svn.jg555.com/build/alan.tar.bz2
--09:24:01-- http://svn.jg555.com/build/alan.tar.bz2
= `alan.tar.bz2'
Resolving svn.jg555.com... 64.183.102.12
Connecting to svn.jg555.com
--- Additional Comments From amodra at bigpond dot net dot au 2009-08-08
01:38 ---
Thanks, got it. Reproduced here too, and looks like file sizes are such that
second relax pass wants to add more stubs, and something is going wrong.
Looking..
--
What|Removed
--
What|Removed |Added
Status|REOPENED|ASSIGNED
Last reconfirmed|2009-08-08 01:38:01 |2009-08-08 01:38:14
date|
--- Additional Comments From amodra at bigpond dot net dot au 2009-08-06
12:36 ---
sounds like a tweak to sym_is_global is needed
--
http://sourceware.org/bugzilla/show_bug.cgi?id=10492
--- You are receiving this mail because: ---
You are on the CC list for the bug
--- Additional Comments From amodra at bigpond dot net dot au 2009-08-06
12:51 ---
Please commit
--
What|Removed |Added
Status|NEW
--- Additional Comments From amodra at bigpond dot net dot au 2009-08-06
23:17 ---
Are you absolutely certain that gcc is invoking the new ld? (Add -v -Wl,-v to
gcc flags.) I think you are still using an old ld.
--
http://sourceware.org/bugzilla/show_bug.cgi?id=10474
--- You
--- Additional Comments From amodra at bigpond dot net dot au 2009-08-07
03:06 ---
How do you know it is using the new ld (with the patch in comment #4!!)? I
specified -v -Wl,-v for a good reason. Please run (from your build dir) the
single gcc command you show in comment #6
--- Additional Comments From amodra at bigpond dot net dot au 2009-08-04
08:18 ---
.
--
What|Removed |Added
Status|ASSIGNED|RESOLVED
--- Additional Comments From amodra at bigpond dot net dot au 2009-08-05
00:25 ---
This is the reason for the failure:
[13] .exception_ranges PROGBITS000412ec 0412ec 007aac 00 A 65280
0 4
When reading an ELF file BFD does some basic sanity checks in elf_object_p.
Why
--
What|Removed |Added
AssignedTo|unassigned at sources dot |amodra at bigpond dot net
|redhat dot com |dot au
Status|NEW
--- Additional Comments From amodra at bigpond dot net dot au 2009-08-03
23:38 ---
I committed a patch yesterday as part of the powerpc ifunc support that should
have fixed this bug. The ChangeLog entry says
(ppc_elf_relax_section): Relax branches to ifunc plt entries too
--- Additional Comments From amodra at bigpond dot net dot au 2009-08-04
02:52 ---
No cancel that comment. On looking at the code more carefully it looks like it
should have been OK. In fact the patch I committed was buggy..
I can't reproduce this bug on my system, which suggest you
--- Additional Comments From amodra at bigpond dot net dot au 2009-07-31
01:03 ---
bed-need_finish_dynamic_ifunc_symbol is unacceptable in my opinion.
Comment #3 says the only difference between good and bad binaries was a
.rela.plt relocation. That would seem to indicate
--- Additional Comments From amodra at bigpond dot net dot au 2009-07-30
01:26 ---
Your change duplicates the logic in elf_link_output_extsym so that
you can fix this bug. You have to do this for all backends which
support IFUNC symbols.
Rubbish. See for example the powerpc backend
--- Additional Comments From amodra at bigpond dot net dot au 2009-07-22
23:51 ---
However I don't see this checkin mentioned on the ML.
It was on bug-binutils.
http://lists.gnu.org/archive/html/bug-binutils/2009-07/msg00067.html
HJ, please commit your patch.
--
http
--- Additional Comments From amodra at bigpond dot net dot au 2009-07-18
10:39 ---
The default libtool value for $RM is rm -f, so the expectation is that your
$RM deletes files without complaining.
--
What|Removed |Added
--- Additional Comments From amodra at bigpond dot net dot au 2009-07-19
00:10 ---
A revised patch, http://sourceware.org/ml/binutils/2006-12/msg00298.html was
committed 2007-01-11. You might like to increase tries = 10 in
ld/emultempl/elf-generic.em, but I doubt this will cure
--- Additional Comments From amodra at bigpond dot net dot au 2009-07-11
03:15 ---
$RM is used by libtool to remove files, so of course setting it to something
nonsensical breaks the build. So does setting a whole lot of other environment
variables. Not a bug.
--
What
--- Additional Comments From amodra at bigpond dot net dot au 2009-07-05
00:11 ---
Your library isn't missing. ld found it somewhere else on your system.
--
What|Removed |Added
--- Additional Comments From amodra at bigpond dot net dot au 2009-06-24
02:22 ---
archive:file specifiers in a linker script are just a pattern to match against.
They do not cause the corresponding file to be loaded. So you need to arrange
for the archive member to be extracted
--- Additional Comments From amodra at bigpond dot net dot au 2009-06-20
02:53 ---
documentation fixed
--
What|Removed |Added
Status|NEW
--- Additional Comments From amodra at bigpond dot net dot au 2009-06-18
23:20 ---
Testcase fixed.
--
http://sourceware.org/bugzilla/show_bug.cgi?id=10255
--- You are receiving this mail because: ---
You are on the CC list for the bug, or are watching someone who
--- Additional Comments From amodra at bigpond dot net dot au 2009-06-06
11:20 ---
Fixed
--
What|Removed |Added
Status|ASSIGNED|RESOLVED
--- Additional Comments From amodra at bigpond dot net dot au 2009-06-05
06:24 ---
Please read the gas info doc, node Flonums. Your floating point literals are
invalid.
--
What|Removed |Added
--
What|Removed |Added
Severity|critical|normal
http://sourceware.org/bugzilla/show_bug.cgi?id=10223
--- You are receiving this mail because:
--- Additional Comments From amodra at bigpond dot net dot au 2009-06-04
09:31 ---
Most likely you have built oprofile using bfd headers that don't match the
libbfd you are linking against.
--
http://sourceware.org/bugzilla/show_bug.cgi?id=10223
--- You are receiving this mail
--- Additional Comments From amodra at bigpond dot net dot au 2009-06-05
05:40 ---
.
--
What|Removed |Added
Status|NEW |RESOLVED
--- Additional Comments From amodra at bigpond dot net dot au 2009-06-01
01:51 ---
Fixed.
--
What|Removed |Added
Status|REOPENED|RESOLVED
--- Additional Comments From amodra at bigpond dot net dot au 2009-05-26
00:19 ---
Because that would be more work in BFD if we wanted to catch all possible cases
of padding, slowing down all file writes. You could of course patch your own
binutils sources. The place that skipped
Severity: minor
Priority: P2
Component: gold
AssignedTo: ian at airs dot com
ReportedBy: amodra at bigpond dot net dot au
CC: bug-binutils at gnu dot org
GCC target triplet: i686-linux
http://sourceware.org/bugzilla/show_bug.cgi?id=10133
--- You
--- Additional Comments From amodra at bigpond dot net dot au 2009-04-29
12:54 ---
I checked strip -K myself, and, yes, it won't help. You still lose the name on
any section symbols in elf.c:swap_out_syms. So I don't think there is any way
you can use the GNU tools to strip debug
--
What|Removed |Added
AssignedTo|unassigned at sources dot |amodra at bigpond dot net
|redhat dot com |dot au
Status|NEW
--- Additional Comments From amodra at bigpond dot net dot au 2009-04-24
15:11 ---
Please let me know if the patch I've just committed fixes it for you.
--
http://sourceware.org/bugzilla/show_bug.cgi?id=10061
--- You are receiving this mail because: ---
You are on the CC
--- Additional Comments From amodra at bigpond dot net dot au 2009-04-12
05:38 ---
Use of genscrba.sh is protected by a test of a bash specific variable, so that
should not be causing a problem.
--
http://sourceware.org/bugzilla/show_bug.cgi?id=10061
--- You are receiving
--- Additional Comments From amodra at bigpond dot net dot au 2009-03-30
07:47 ---
This particular target uses '/' to start a comment. Unfortunately that means
you can't do division.
--
What|Removed |Added
--- Additional Comments From amodra at bigpond dot net dot au 2009-03-25
13:16 ---
See http://sourceware.org/bugzilla/show_bug.cgi?id=6494
Please close this bug as a duplicate if so.
--
http://sourceware.org/bugzilla/show_bug.cgi?id=10003
--- You are receiving this mail because
--- Additional Comments From amodra at bigpond dot net dot au 2009-03-26
02:55 ---
Fixed. You will of course continue to get error about lack of 16-byte
relocations.
--
What|Removed |Added
--- Additional Comments From amodra at bigpond dot net dot au 2009-03-23
14:22 ---
objdump is a better interpreter of x86 opcodes than most humans
--
What|Removed |Added
--- Additional Comments From amodra at bigpond dot net dot au 2009-03-23
14:23 ---
see objdump --help, -w option
--
What|Removed |Added
Status|NEW
--- Additional Comments From amodra at bigpond dot net dot au 2009-03-23
04:14 ---
This is not a ld bug, it is a glibc one. I believe Adreas when he says it has
been fixed, but to be sure I checked the glibc source and indeed it has been
fixed.
--
What|Removed
1 - 100 of 421 matches
Mail list logo
