https://sourceware.org/bugzilla/show_bug.cgi?id=30578
--- Comment #1 from Mike Hommey ---
Created attachment 14943
--> https://sourceware.org/bugzilla/attachment.cgi?id=14943&action=edit
Preprocessed C source
--
You are receiving this mail because:
You are on the CC list for the bug.
IRMED
Severity: normal
Priority: P2
Component: gas
Assignee: unassigned at sourceware dot org
Reporter: mh-sourceware at glandium dot org
Target Milestone: ---
Created attachment 14942
--> https://sourceware.org/bugzilla/attachment.cgi?id=14942&action=ed
https://sourceware.org/bugzilla/show_bug.cgi?id=30566
--- Comment #2 from Mike Hommey ---
Seems to work better with the patch:
Maximum resident set size (kbytes): 9096288
--
You are receiving this mail because:
You are on the CC list for the bug.
https://sourceware.org/bugzilla/show_bug.cgi?id=30566
Mike Hommey changed:
What|Removed |Added
CC||hjl.tools at gmail dot com
--
You are
Severity: normal
Priority: P2
Component: ld
Assignee: unassigned at sourceware dot org
Reporter: mh-sourceware at glandium dot org
Target Milestone: ---
When linking libxul.so from Firefox, with -z pack-relative-relocs (output from
time -v
: UNCONFIRMED
Severity: normal
Priority: P2
Component: gold
Assignee: ccoutant at gmail dot com
Reporter: mh-sourceware at glandium dot org
CC: ian at airs dot com
Target Milestone: ---
We've recently got this link failure with armv7 an
https://sourceware.org/bugzilla/show_bug.cgi?id=29489
--- Comment #4 from Mike Hommey ---
The patch seems to work as expected.
--
You are receiving this mail because:
You are on the CC list for the bug.
https://sourceware.org/bugzilla/show_bug.cgi?id=29489
--- Comment #2 from Mike Hommey ---
The bfd objects used for dlltool should have the BFD_DETERMINISTIC_OUTPUT flag
set.
--
You are receiving this mail because:
You are on the CC list for the bug.
IRMED
Severity: normal
Priority: P2
Component: binutils
Assignee: unassigned at sourceware dot org
Reporter: mh-sourceware at glandium dot org
Target Milestone: ---
```
$ cat <foo.def
LIBRARY foo.dll
EXPORTS
foo
EOF
$ x86_64-w64-mingw32-dlltool -m
: binutils
Assignee: unassigned at sourceware dot org
Reporter: mh-sourceware at glandium dot org
Target Milestone: ---
Both ar and objcopy default to build deterministic archives, but dlltool
doesn't.
--
You are receiving this mail because:
You are on the CC list for the bug.
Component: ld
Assignee: unassigned at sourceware dot org
Reporter: mh-sourceware at glandium dot org
Target Milestone: ---
Both ld and gold support the option, but their manual page don't mention it
existing.
--
You are receiving this mail because:
You are on the CC
Component: gold
Assignee: ccoutant at gmail dot com
Reporter: mh-sourceware at glandium dot org
CC: ian at airs dot com
Target Milestone: ---
$ cat test.c
__attribute__((weak, visibility("hidden"))) extern void foo();
int main() {
if (foo)
foo();
https://sourceware.org/bugzilla/show_bug.cgi?id=23600
--- Comment #6 from Mike Hommey ---
It works, and it still detects cases where the IR file doesn't have the right
architecture:
$ clang -flto -o main.o -c main.c
$ clang -flto -o main main.ld
$ clang -m32 -flto -o main main.ld
/tmp/bin/ld: i3
https://sourceware.org/bugzilla/show_bug.cgi?id=23600
--- Comment #4 from Mike Hommey ---
AFAICT, bfd/plugin.c is never setting the arch_info, so that would likely be
the problem.
--
You are receiving this mail because:
You are on the CC list for the bug.
___
https://sourceware.org/bugzilla/show_bug.cgi?id=23600
--- Comment #3 from Mike Hommey ---
So the difference when linking a non LTO object file with the same linker
script is that after the call to bfd_check_format, check->arch_info is
bfd_x86_64_arch (it is bfd_default_arch_struct before the call
https://sourceware.org/bugzilla/show_bug.cgi?id=23600
--- Comment #2 from Mike Hommey ---
(entry->the_bfd->arch_info is also bfd_default_arch_struct when not using the
linker script)
--
You are receiving this mail because:
You are on the CC list for the bug.
https://sourceware.org/bugzilla/show_bug.cgi?id=23600
--- Comment #1 from Mike Hommey ---
In ldfile_try_open_bfd, we end up in the error case presumably because:
(rr) print check->filename
$2 = 0x5582b37b57e0 "main.o"
(rr) print check->arch_info
$3 = (const struct bfd_arch_info *) 0x5582b3286fe0
: normal
Priority: P2
Component: ld
Assignee: unassigned at sourceware dot org
Reporter: mh-sourceware at glandium dot org
Target Milestone: ---
$ cat < main.c
int main() { return 0; }
EOF
$ clang -flto -o main.o -c main.c
$ cat < main.ld
INPUT(main.
https://sourceware.org/bugzilla/show_bug.cgi?id=23591
--- Comment #7 from Mike Hommey ---
It worked.
--
You are receiving this mail because:
You are on the CC list for the bug.
___
bug-binutils mailing list
bug-binutils@gnu.org
https://lists.gnu.org/m
https://sourceware.org/bugzilla/show_bug.cgi?id=23591
--- Comment #4 from Mike Hommey ---
It's worth noting that gold doesn't trip on the same object files (that is,
chanfing the linker command line to include -fuse-ld=gold makes it work)
--
You are receiving this mail because:
You are on the C
https://sourceware.org/bugzilla/show_bug.cgi?id=23591
--- Comment #3 from Mike Hommey ---
They are generate by clang from the combination of -fsanitize=address and
-fsanitize=fuzzer-no-link. But using that with some dummy source file doesn't
trigger the problem, so some subtle thing is happening.
https://sourceware.org/bugzilla/show_bug.cgi?id=23591
--- Comment #1 from Mike Hommey ---
It seems to be due to most __sancov_cntrs sections having the SHT_GROUP flag.
This makes _bfd_elf_section_already_linked set the section's output_section to
bfd_abs_section_ptr, which then fulfils the condit
ified
Status: UNCONFIRMED
Severity: normal
Priority: P2
Component: ld
Assignee: unassigned at sourceware dot org
Reporter: mh-sourceware at glandium dot org
Target Milestone: ---
I haven't figured out a small test case, but here's a real worl
https://sourceware.org/bugzilla/show_bug.cgi?id=23268
--- Comment #6 from Mike Hommey ---
Thanks. Is this going to make the cut for 2.31?
--
You are receiving this mail because:
You are on the CC list for the bug.
___
bug-binutils mailing list
bug-bin
https://sourceware.org/bugzilla/show_bug.cgi?id=23268
--- Comment #3 from Mike Hommey ---
I'm using --version-script to hide symbols (because --exclude-symbols is PE
only, and I don't think it takes globs anyways). Of course, when not using one,
it works.
--
You are receiving this mail because:
https://sourceware.org/bugzilla/show_bug.cgi?id=23268
--- Comment #4 from Mike Hommey ---
FWIW, it works with both BFD ld and lld.
--
You are receiving this mail because:
You are on the CC list for the bug.
___
bug-binutils mailing list
bug-binutils@g
https://sourceware.org/bugzilla/show_bug.cgi?id=23268
--- Comment #1 from Mike Hommey ---
As for what kind of subtle problem this can cause:
$ cat foo.c
#include
void* malloc_hook(size_t s) {
abort();
}
void* (*__malloc_hook)(size_t s) = malloc_hook;
int main() { malloc(1);}
$ cat ver
hi
ified
Status: UNCONFIRMED
Severity: normal
Priority: P2
Component: gold
Assignee: ccoutant at gmail dot com
Reporter: mh-sourceware at glandium dot org
CC: ian at airs dot com
Target Milestone: ---
$ cat foo.c
void foo() {}
int
Severity: normal
Priority: P2
Component: binutils
Assignee: unassigned at sourceware dot org
Reporter: mh-sourceware at glandium dot org
Target Milestone: ---
Created attachment 11061
--> https://sourceware.org/bugzilla/attachment.cgi?id=11061&action=e
https://sourceware.org/bugzilla/show_bug.cgi?id=20828
Mike Hommey changed:
What|Removed |Added
CC||mh-sourceware at glandium dot
org
https://sourceware.org/bugzilla/show_bug.cgi?id=19392
--- Comment #1 from Mike Hommey ---
Bisected:
a5cd8f05ca759fdb9b27fc98a08edb5f85369ad9 is the first bad commit
commit a5cd8f05ca759fdb9b27fc98a08edb5f85369ad9
Author: Cary Coutant
Date: Fri Jan 9 15:55:50 2015 -0800
Don't align start
Priority: P2
Component: gold
Assignee: ccoutant at gmail dot com
Reporter: mh-sourceware at glandium dot org
CC: ian at airs dot com
Target Milestone: ---
$ gold --version
GNU gold (GNU Binutils for Debian 2.25.90.20151209) 1.11
Copyright (C
gold
Assignee: ccoutant at google dot com
Reporter: mh-sourceware at glandium dot org
CC: ian at airs dot com
There are many reasons one may want to use a partial linker script, such as
rearranging sections. Going with a complete linker script in those cases is
most of the
https://sourceware.org/bugzilla/show_bug.cgi?id=17451
--- Comment #1 from Mike Hommey ---
I believe https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=624208 is the same
bug.
--
You are receiving this mail because:
You are on the CC list for the bug.
https://sourceware.org/bugzilla/show_bug.cgi?id=13597
Mike Hommey changed:
What|Removed |Added
CC||mh-sourceware at glandium dot
org
http://sourceware.org/bugzilla/show_bug.cgi?id=14356
--- Comment #1 from Mike Hommey 2012-07-12
07:59:25 UTC ---
It also looks like -z max-page-size doesn't work properly when targetting arm:
LOAD 0x00 0x 0x 0x538cc 0x538cc R E 0x1000
LOAD 0x054000 0x
http://sourceware.org/bugzilla/show_bug.cgi?id=14356
Bug #: 14356
Summary: -z max-page-size doesn't change PT_LOAD's p_align
Product: binutils
Version: 2.22
Status: NEW
Severity: normal
Priority: P2
Component: g
http://sourceware.org/bugzilla/show_bug.cgi?id=13809
Bug #: 13809
Summary: --print-icf-sections prints about folding empty
sections
Product: binutils
Version: 2.22
Status: NEW
Severity: normal
Priorit
http://sourceware.org/bugzilla/show_bug.cgi?id=13364
--- Comment #2 from Mike Hommey 2011-10-31
18:23:26 UTC ---
I'll try to see if I can reduce the testcase.
--
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
--- You are receiving this mail because: ---
You a
http://sourceware.org/bugzilla/show_bug.cgi?id=13366
Bug #: 13366
Summary: Allow to "pack" some functions together
Product: binutils
Version: 2.23 (HEAD)
Status: NEW
Severity: normal
Priority: P2
Component: gold
http://sourceware.org/bugzilla/show_bug.cgi?id=13365
Bug #: 13365
Summary: "Fold" functions that only jump to some other one
Product: binutils
Version: 2.23 (HEAD)
Status: NEW
Severity: enhancement
Priority: P2
http://sourceware.org/bugzilla/show_bug.cgi?id=13364
Bug #: 13364
Summary: ICF may not be as thorough as one would expect
Product: binutils
Version: 2.23 (HEAD)
Status: NEW
Severity: normal
Priority: P2
Componen
http://sourceware.org/bugzilla/show_bug.cgi?id=13320
--- Comment #1 from Mike Hommey 2011-10-26
11:41:22 UTC ---
BTW, it should be possible, in binaries that are completely compiled as thumb,
to have a completely thumb PLT code.
--
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cg
http://sourceware.org/bugzilla/show_bug.cgi?id=13314
--- Comment #2 from Mike Hommey 2011-10-26
05:57:06 UTC ---
I'm not very much convinced the keep-unique flag would be very practical. Maybe
having an option to choose the global behavior of ICF vs. reordering would be
good. Something that woul
http://sourceware.org/bugzilla/show_bug.cgi?id=13322
Bug #: 13322
Summary: ARM/Thumb interwork stubs should appear in some way in
the symbol table
Product: binutils
Version: 2.23 (HEAD)
Status: NEW
Severity: no
http://sourceware.org/bugzilla/show_bug.cgi?id=13321
Bug #: 13321
Summary: Gold creates mixed Thumb/ARM files that objdump -d
doesn't understand
Product: binutils
Version: 2.23 (HEAD)
Status: NEW
Severity: norm
http://sourceware.org/bugzilla/show_bug.cgi?id=13320
Bug #: 13320
Summary: Please support thumb entry points in PLT instead of
using (more or less randomly located) stubs
Product: binutils
Version: 2.23 (HEAD)
Status: NE
http://sourceware.org/bugzilla/show_bug.cgi?id=12322
Mike Hommey changed:
What|Removed |Added
CC||mh-sourceware at glandium
http://sourceware.org/bugzilla/show_bug.cgi?id=13314
Bug #: 13314
Summary: ICF clashes with section ordering
Product: binutils
Version: 2.23 (HEAD)
Status: NEW
Severity: normal
Priority: P2
Component: gold
http://sourceware.org/bugzilla/show_bug.cgi?id=13049
Summary: Generates R_ARM_NONE relocations for stub_* symbols
when using -fdata-sections
Product: binutils
Version: 2.22 (HEAD)
Status: NEW
Severity: normal
Prio
http://sourceware.org/bugzilla/show_bug.cgi?id=13025
Mike Hommey changed:
What|Removed |Added
Status|NEW |RESOLVED
Resolution|
http://sourceware.org/bugzilla/show_bug.cgi?id=13025
--- Comment #5 from Mike Hommey 2011-07-25
15:08:48 UTC ---
(In reply to comment #4)
> In the end, both comments still apply, once I build with -ffunction-sections.
*Sigh*, this is getting embarassing. I had -Wl,--icf=safe twice on my command
http://sourceware.org/bugzilla/show_bug.cgi?id=13025
--- Comment #4 from Mike Hommey 2011-07-25
15:03:14 UTC ---
In the end, both comments still apply, once I build with -ffunction-sections.
--
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
--- You are receiving
http://sourceware.org/bugzilla/show_bug.cgi?id=13025
--- Comment #3 from Mike Hommey 2011-07-25
14:32:43 UTC ---
(In reply to comment #2)
> (In reply to comment #1)
> > Whatever Debian provides as version 2.21.53.20110720 doesn't have this
> > problem... (with or without the debian patches) so i
http://sourceware.org/bugzilla/show_bug.cgi?id=13025
--- Comment #2 from Mike Hommey 2011-07-25
14:21:50 UTC ---
(In reply to comment #1)
> Whatever Debian provides as version 2.21.53.20110720 doesn't have this
> problem... (with or without the debian patches) so it might actually be fixed
> on
http://sourceware.org/bugzilla/show_bug.cgi?id=13025
--- Comment #1 from Mike Hommey 2011-07-25
14:13:15 UTC ---
Whatever Debian provides as version 2.21.53.20110720 doesn't have this
problem... (with or without the debian patches) so it might actually be fixed
on trunk.
--
Configure bugmail:
http://sourceware.org/bugzilla/show_bug.cgi?id=13025
Summary: Segmentation fault when linking mozilla-central with
icf on android target
Product: binutils
Version: 2.21
Status: NEW
Severity: normal
Priority: P2
http://sourceware.org/bugzilla/show_bug.cgi?id=12975
Summary: --gc-sections doesn't remove symbols hidden with
version scripts
Product: binutils
Version: 2.21
Status: NEW
Severity: normal
Priority: P2
Com
58 matches
Mail list logo
