Now that binutils 2.46.0 has been introduced and binutils 2.45.1 made
the default version, drop the oldest supported version, binutils 2.43,
keeping only the 3 last versions supported: 2.44, 2.45.1 and 2.46.0.
Signed-off-by: Bernd Kuhls <bernd@kuhls.net>
Now that support for binutils 2.46.0 has been introduced, we follow our
policy of making binutils 2.45.1 the default version.
Signed-off-by: Bernd Kuhls <bernd@kuhls.net>
Signed-off-by: Julien Olivain <ju.o@free.fr>
https://sourceware.org/pipermail/binutils/2026-February/148149.html
" This release contains numerous bug fixes, and also the following new
features:
* Support for new instructions added to AMD, ARM and RISC-V
architectures.
* Support for version 3 of the SFrame standard.
* The readelf program can now display the contents of Global Offset
Tables.
* Improved linker tagging support."
We bring and rebased patches 0001 and 0002 that we carry for binutils
2.45.
Signed-off-by: Bernd Kuhls <bernd@kuhls.net>
Signed-off-by: Julien Olivain <ju.o@free.fr>
Now that binutils 2.45 has been introduced and binutils 2.44 made the
default version, drop the oldest supported version, binutils 2.42,
keeping only the 3 last versions supported: 2.43, 2.44 and 2.45.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Julien Olivain <ju.o@free.fr>
Now that support for binutils 2.45 has been introduced, we follow our
policy of making binutils 2.44 the default version.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Julien Olivain <ju.o@free.fr>
We bring patches 0001 and 0002 that we carry for binutils 2.44. Patch
0002 requires a small update as a nearby configure option has been
removed between 2.44 and 2.45. Patch 0003 that we have for binutils
2.44 is not needed as it is part of the 2.45 release.
Changes in 2.45:
* New versioned release of libsframe: libsframe.so.2. This release introduces
versioned symbols with version node name LIBSFRAME_2.0. Some new symbols
have been added to support the new flag SFRAME_F_FDE_FUNC_START_PCREL and
retrieving flags from SFrame decoder and encoder objects:
- Addition of sframe_decoder_get_flags,
sframe_decoder_get_offsetof_fde_start_addr, sframe_encoder_get_flags,
sframe_encoder_get_offsetof_fde_start_addr.
This release also includes backward-incompatible ABI changes:
- Removal of sframe_get_funcdesc_with_addr.
- Change in the behavior of sframe_decoder_get_funcdesc_v2,
sframe_encoder_add_funcdesc_v2 and sframe_encoder_write.
* On s390 64-bit (s390x), gas, ld, objdump, and readelf now support generating
and processing SFrame V2 stack trace information (.sframe). The assembler
generates SFrame info from CFI directives with option "--gsframe". The
linker generates SFrame info for the linker-generated .plt section and merges
all .sframe sections. Both objdump and readelf dump SFrame info with option
"--sframe[=<section-name>]".
* For SFrame stack trace format, the function start address in each SFrame
FDE has a changed encoding: The 32-bit signed integer now holds the offset
of the start PC of the associated function from the sfde_func_start_address
field itself (instead of the earlier where it was the offset from the start
of the SFrame section itself). All SFrame sections generated by gas and ld
now default to this new encoding, setting the (new)
SFRAME_F_FDE_FUNC_START_PCREL flag.
Relocatable SFrame links are now fixed.
* Readelf now recognizes RISC-V GNU_PROPERTY_RISCV_FEATURE_1_CFI_SS and
GNU_PROPERTY_RISCV_FEATURE_1_CFI_LP_UNLABELED for zicfiss and zicfilp
extensions.
* For RISC-V dis-assembler, the definition of mapping symbol $x is changed,
so the file needs to be rebuilt since 2.45 once used .option arch directives.
* The LoongArch disassembler now properly accepts multiple disassembly
options given by -M, such as "-M no-aliases,numeric". (Previously only the
first option took effect.)
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
[Julien: fix BR2_BINUTILS_VERSION_2_45_X prompt to 2.45.1]
Signed-off-by: Julien Olivain <ju.o@free.fr>
When BR2_BINUTILS_GPROFNG=y, the build currently fails with GCC 15.x,
of course on architectures for which gprofng is actually
available. This issue only exists in binutils 2.42 and 2.43.1, as
binutils 2.44 has the necessary fixes.
Therefore, this commit backports the relevant commits from binutils
2.44 to 2.43.1 and 2.42.
Fixes:
https://autobuild.buildroot.net/results/7cf52019f87625c73ef3c4ca0b6270e20a4ab768/ (binutils 2.42)
https://autobuild.buildroot.net/results/478e24af1a03dea5ec64b77ca611b4f7225c9d39/ (binutils 2.43.1)
The following defconfig, built under a Fedora 42 Docker container
allows to reproduce the issue:
BR2_x86_nehalem=y
BR2_BINUTILS_VERSION_2_42_X=y
BR2_BINUTILS_GPROFNG=y
BR2_INIT_NONE=y
BR2_SYSTEM_BIN_SH_NONE=y
# BR2_PACKAGE_BUSYBOX is not set
# BR2_TARGET_ROOTFS_TAR is not set
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Romain Naour <romain.naour@smile.fr>
The binutils-gdb code base in Git does have a copy of readline that is
used by gdb.
The release tarballs of binutils do NOT contain readline as it's not
needed for binutils itself (only gdb).
However, when we build the ARC binutils, taken from Git, the readline/
folder is present, and therefore readline is built even if it is not
actually needed.
This causes build failures as the copy of readline in ARC binutils
2024.12 (based on binutils 2.43) does not build with GCC 15.x.
To work around this problem, we use a suggestion from Waldemar
Brodkorb: pass --with-system-readline. Indeed, this disables the build
of readline... but it does not actually check that we have a system
readline, since we're not building gdb.
Fixes:
https://autobuild.buildroot.net/results/976243181f7000fe2790b3450bdae51e8c12769e/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Romain Naour <romain.naour@smile.fr>
This commit updates the ARC toolchain components to the 2024.12
release, keeping the same patches.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Julien Olivain <ju.o@free.fr>
Backport the upstream patch that fixes the following build error when
compiling for mips with gcc 15:
In file included from mips-opc.c:29:
mips-opc.c: In function 'decode_mips_operand':
mips-formats.h:86:7: error: expected identifier or '(' before
'static_assert'
86 | static_assert[(1 << (SIZE)) == ARRAY_SIZE (MAP)]; \
| ^~~~~~~~~~~~~
The patch is already part of upstream binutils 2.44, so we only need
it for 2.42 and 2.43.1.
All 3 versions we have of host-binutils were build-tested using the
defconfig from the autobuilder failure (see the link below) and gcc 15
on the host.
Fixes:
- https://autobuild.buildroot.org/results/873/873ec25cf01d5f2b9ae7044e0b1d8d8791b781e6/
Signed-off-by: Raphaël Mélotte <raphael.melotte@mind.be>
Signed-off-by: Julien Olivain <ju.o@free.fr>
This fixes a memory leaks that affects both binutils 2.43 and 2.44,
see https://www.cve.org/CVERecord?id=CVE-2025-3198
Fixes the following CVE:
- CVE-2025-3198: A vulnerability has been found in GNU Binutils 2.43/2.44
and classified as problematic. Affected by this
vulnerability is the function display_info of the file
binutils/bucomm.c of the component objdump.
The manipulation leads to memory leak.
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
Signed-off-by: Julien Olivain <ju.o@free.fr>
Commit 360fd01de2 changed the binutils
package to use 2.43.x as the default version, but got the change wrong
in the BINUTILS_VERSION variable assignment. Indeed we don't support
2.43, but 2.43.1. This issue is causing the following build failure:
ERROR: No hash found for binutils-2.43.tar.xz
Fixes:
https://gitlab.com/vjardin/buildroot/-/jobs/9043815781
Signed-off-by: Vincent Jardin <vjardin@free.fr>
[Thomas: proper fix, and improved commit log]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Now that we have added binutils 2.44, made binutils 2.43 the default,
let's continue to follow our usual policy and drop support for
binutils 2.41.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Now that we've added support for binutils 2.44, follow our usual
process of masking the previous release, 2.43, the default one.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Remaining "nios2" strings come from existing patches that
are not removed with along with BR2_nios2 symbol.
Signed-off-by: Romain Naour <romain.naour@smile.fr>
Signed-off-by: Julien Olivain <ju.o@free.fr>
Now that we have integrated support for binutils 2.43, and made
binutils 2.42 the default, following our tradition, we can drop
support for binutils 2.40.
In addition to the usual things, there is an additional change
in elf2flt.mk, which had a special condition applicable to binutils
2.40, which can be removed now.
Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Now that 2.43 has been introduced, let's make 2.42 the latest version,
following the traditional Buildroot policy.
Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Now that binutils 2.39 is gone, we can drop
BR2_PACKAGE_BINUTILS_HAS_NO_LIBSFRAME, which was only used for
binutils 2.39. This was a blind option, so Config.in.legacy handling
is not needed.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Now that we have integrated support for binutils 2.42, and made
binutils 2.41 the default, following our tradition, we can drop
support for binutils 2.39.
In addition to the usual things, there is an additional minor change
in elf2flt.mk, which had a special condition applicable to binutils
2.39 or 2.40, which can be simplified to only apply to binutils 2.40
now.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Now that 2.42 has been introduced, let's make 2.41 the latest version,
following the traditional Buildroot policy.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Note that the hash of the tarball does not need to be added, as it was
already added as part of commit 11b439ce1b
("package/binutils-bare-metal: new package").
Our existing 2 patches are simply rebased, with minor conflict
resolution required.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
[yann.morin.1998@free.fr: keep version list in strict version order]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
This patch adds a new package for building binutils for a bare-metal toolchain.
The cpu architecture is defined by a toolchain-bare-metal virtual package.
While any cpu architecture could be used, the default configuration will be a
Xilinx microblaze little endian architecture, so that buildroot will be able
to build the microblaze firmware applications for zynqmp and versal.
In order to build the zynqmp pmufw and versal plm applications without error,
binutils version 2.41 or higher is required.
Signed-off-by: Neal Frager <neal.frager@amd.com>
Signed-off-by: Ibai Erkiaga <ibai.erkiaga-elorza@amd.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
While updating ARC GNU toolchain version (see [1]) we unintentionally
got rid of "0001-poison-system-directories.patch" which still makes
sense.
Even though one may think that we do not need these Binutils patches
because we do "poisoning" of system directories in toolchain wrapper,
but these wrappers don't exist for tools originated from Binutils, only
for gcc, g++ etc.
So we need to keep that patch still, sorry.
[1] https://git.buildroot.net/buildroot/commit/?id=045ab73702b8bc09f94ca1f3b8e5bf3acb0b0843
Reported-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Fix the following build failure with oprofile raised since bump of
binutils to version 2.40 in commit
35656482d3:
configure: error: bfd library not found
[...]
configure:17928: checking for bfd_openr in -lbfd
configure:17953: /home/buildroot/autobuild/run/instance-1/output-1/host/bin/arm-linux-gcc -o conftest -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -Os -g0 -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 conftest.c -lbfd -liberty -lpopt -ldl -lintl >&5
/home/buildroot/autobuild/run/instance-1/output-1/host/opt/ext-toolchain/bin/../lib/gcc/arm-buildroot-linux-uclibcgnueabi/12.3.0/../../../../arm-buildroot-linux-uclibcgnueabi/bin/ld: warning: libsframe.so.0, needed by /home/buildroot/autobuild/run/instance-1/output-1/host/arm-buildroot-linux-uclibcgnueabi/sysroot/usr/lib/libbfd.so, not found (try using -rpath or -rpath-link)
Indeed, in this case, libsframe is not installed even after applying
commit 1b4d921e1d because
BR2_BINUTILS_VERSION_2_40_X is not selected by anyone (binutils package
is selected by oprofile and the toolchain is not generated by buildroot)
To fix this issue, invert the logic: install libsframe by default (i.e.
when binutils is selected or with a buildroot toolchain). libsframe will
not be installed only if binutils < 2.40 is detected.
Fixes:
- http://autobuild.buildroot.org/results/af9a2d52823a332b48e6df14d2708b6a4b3833a4
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Now that 2.41.x has been added, that 2.40.x is the default version,
drop support for 2.38.x.
Signed-off-by: Bernd Kuhls <bernd@kuhls.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Now that 2.41 has been released, let's use 2.40.x as the default
binutils version.
Signed-off-by: Bernd Kuhls <bernd@kuhls.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Release notes:
https://sourceware.org/pipermail/binutils/2023-July/128719.html
Copied patches from version 2.40.
Build-tested using these four defconfigs:
----------
BR2_arceb=y
----------
BR2_BINUTILS_VERSION_2_39_X=y
----------
(empty defconfig meaning binutils-2.40)
----------
BR2_BINUTILS_VERSION_2_41_X=y
----------
each using these commands:
$ make host-libiberty
$ make host-binutils
Signed-off-by: Bernd Kuhls <bernd@kuhls.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Now that 2.40.x has been added, that 2.39.x is the default version,
drop support for 2.37.x.
Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
BR2_BINUTILS_GPROFNG was made available for known supported versions
only (binutils 2.39 at this point). As all new versions of binutils
support gprofng, it makes sense to invert the logic, and instead make
BR2_BINUTILS_GPROFNG not available with old binutils versions.
Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Support for the AVX512FP16 instructions was added in binutils
2.38. See the binutils 2.38 releases notes [0] that state:
X86:
[...]
* Add support for Intel AVX512_FP16 instructions.
[0] https://sourceware.org/pipermail/binutils/2022-February/119721.html
It turns out that when building GCC 12.x, some of these AVX512FP16
instructions are now used, and therefore when binutils < 2.38 is used,
the build fails as the assembler does not recognize those
instructions:
/tmp/ccChzL2g.s: Assembler messages:
/tmp/ccChzL2g.s:20: Error: no such instruction: `vmovw 24(%esp),%xmm2'
/tmp/ccChzL2g.s:21: Error: no such instruction: `vmovw 28(%esp),%xmm3'
/tmp/ccChzL2g.s:22: Error: no such instruction: `vmovw 32(%esp),%xmm4'
/tmp/ccChzL2g.s:23: Error: no such instruction: `vmovw 36(%esp),%xmm5'
/tmp/ccChzL2g.s:30: Error: no such instruction: `vcvtsh2ss %xmm2,%xmm6,%xmm6'
/tmp/ccChzL2g.s:36: Error: no such instruction: `vcvtsh2ss %xmm3,%xmm6,%xmm6'
/tmp/ccChzL2g.s:42: Error: no such instruction: `vcvtsh2ss %xmm4,%xmm7,%xmm7'
/tmp/ccChzL2g.s:48: Error: no such instruction: `vcvtsh2ss %xmm5,%xmm1,%xmm1'
/tmp/ccChzL2g.s:80: Error: no such instruction: `vcvtss2sh (%esp),%xmm1,%xmm1'
/tmp/ccChzL2g.s:96: Error: no such instruction: `vcvtss2sh (%esp),%xmm0,%xmm0'
/tmp/ccChzL2g.s💯 Error: no such instruction: `vucomish %xmm1,%xmm1'
/tmp/ccChzL2g.s:103: Error: no such instruction: `vucomish %xmm0,%xmm0'
/tmp/ccChzL2g.s:122: Error: no such instruction: `vucomish %xmm2,%xmm2'
The same issue does not occur with GCC 11.x, but nothing prevents
other packages than GCC to use those instructions, so the problem
really lies on the binutils side missing the support for those
instructions.
Also, in Buildroot, we do not distinguish AVX512 in general from
AVX512FP16 specifically, so our only option is to make binutils 2.37
unavailable for AVX512 systems even if some of them perhaps don't
support AVX512FP16 anyway. This seems like a reasonable trade-off, as
binutils 2.38 is anyway already the default in Buildroot, and
AVX512-capable systems are fairly recent, and therefore using a recent
binutils version should not be a problem on these platforms.
Fixes:
http://autobuild.buildroot.net/results/eb6e28c934654e6d714973415a2fb452f9580279/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Commit 7ff21f8d4b (package/binutils: handle gprofng as an optional
feature) introduced the typo. Although it was noticed, the commit was
not amended before being pushed...
Fix that now...
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
binutils 2.39 added support for gprofng, a new profiler (see
https://www.phoronix.com/news/GNU-Profiler-gprofng).
This new profiler is enabled by default, but it requires bison on the
host.
In order to handle this, this commit:
- Adds a new option BR2_BINUTILS_GPROFNG, which allows to
enable/disable gprofng in host-binutils
- Unconditionnally disables gprofng for the target binutils, based on
the idea that in a Buildroot context the analysis of profiling data
is generally done on the host system. This can of course always be
revisited later by adding a new option to the target binutils
package.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Currently, this option doesn't do anything. It only adds
--enable-plugins --enable-lto to the configure flags, but doesn't
disable them if it is not set. Since both of these default to enabled,
plugins and lto are effectively always enabled.
There really is no need to make this configurable: it adds a bit of size
and build time to host-binutils, but we don't care about that for host
tools. It's still up to individual builds to enable the LTO options.
Therefore, remove the option entirely. For clarity, explicitly pass
--enable-plugins --enable-lto to configure.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Now that recent versions of binutils work with FLAT binaries, we can
drop the old 2.32 version, which was kept only to keep support FLAT
binaries.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Thanks to the bump of elf2flt to version 2021.08, the issue with
recent versions of binutils has been fixed, so we can re-enable using
the recent binutils versions.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Bernd Kuhls