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>
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>
Now that we have 2.35, 2.36 and 2.37, with 2.36 as the default, we can
remove 2.34.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Now that we have added version 2.37, it's time to use the 2.36.x
series as the default version.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
[Arnout: also update default in binutils.mk]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
State of the patches:
- 0001-sh-conf.patch
Refreshed
- 0002-poison-system-directories.patch
Refreshed, but needed some adaptations as the bfd_boolean type no
longer exists, and the standard "bool" type is now used instead.
- 0003-or1k-Fix-issue-with-plt-link-failure-for-local-calls.patch
Drop, present in 2.37, merged upstream as
a76ef689b60405e494cb99e198acf3c82f467f7d
- 0004-or1k-Implement-relocation-R_OR1K_GOT_AHI16-for-gotha.patch
Drop, present in 2.37, merged upstream as
0b3e14c90283c5d234884d0ebe8510bc3c9bc687
- 0005-or1k-Avoid-R_OR1K_GOT16-overflow-failures-in-presenc.patch
Drop, present in 2.37, merged upstream as
3c3de29b048bca6b4aa4235c647b9328e71801b6
- 0006-or1k-Support-large-plt_relocs-when-generating-plt-en.patch
Drop, present in 2.37, merged upstream as
284a1309021a0ef4c29f198470d95652f02b13f0
- 0007-bfd-elf32-or1k-fix-building-with-gcc-version-5.patch
Refreshed
- 0008-or1k-fix-pc-relative-relocation-against-dynamic-on-P.patch
Refreshed
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Release notes:
We are very sorry to have to report that a problem was found with the
GNU Binutils 2.36 release. It turns out that it contained a small
portion of code that was not covered by an FSF copyright assignment.
So we have created a replacement release - 2.36.1 - with that code
removed.
In addition we found that a fix for a theoretical security
vulnerability[1] was itself broken and could result in the archiver
program "ar" misbehaving. So we have chosen to revert the fix from
the 2.36.1 release whilst the problem is properly resolved.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Now that 2.36.x has been added, that 2.35.x is the default version,
drop support for 2.33.x.
Note that we keep binutils 2.32.x as it is the latest version that
works for FLAT binaries (used on noMMU platforms).
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Now that 2.36 has been released, let's use 2.35.x as the default
binutils version.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This commit bumps ARC toolchain to arc-2020.09-release.
ARC GNU tools of version arc-2020.09-release bring some quite significant
changes like:
* Binutils 2.34.50 with additional ARC patches
* GCC 10.0.2 with additional ARC patches
* GDB 10.0.50 with additional ARC patches
Signed-off-by: Evgeniy Didin <Evgeniy.Didin@synopsys.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: arc-buildroot@synopsys.com
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Now that binutils 2.35.1 has been released, it is time to move to
binutils 2.34 as the default binutils version, instead of 2.33.1.
Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Now that binutils 2.34 has been introduced, and we have moved to
2.33.1 as the default version, it is time to drop support for binutils
2.31.1.
Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Now that binutils 2.34 has been released, it is time to move to
binutils 2.33.1 as the default binutils version, instead of 2.32.
Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
This commit bumps ARC toolchain to arc-2020.03-release.
ARC GNU tools of version arc-2020.03-release bring some quite significant
changes like:
* Binutils 2.34 with additional ARC patches
* GCC 9.3 with additional ARC patches
* glibc 2.30 with additional ARC patches
* GDB 10-prerelease with additional ARC patches
Signed-off-by: Evgeniy Didin <Evgeniy.Didin@synopsys.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: arc-buildroot@synopsys.com
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Now that binutils 2.33.1 has been introduced, and we have moved to
2.32 as the default version, it is time to drop support for binutils
2.30.
Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Now that binutils 2.33.1 has been released, it is time to move to
binutils 2.32 as the default binutils version, instead of 2.31.1.
Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This commit bumps ARC toolchain to arc-2019.09-rc1.
We want to test how new toolchain-rc1 builds packages,
so we can make fixes before release of toolcain.
ARC GNU tools of version arc-2019.09-rc1 bring some quite significant changes like:
* Binutils v2_33.20191002 with additional ARC patches
* GCC 9.2.0 with additional ARC patches
* glibc 2.30 with additional ARC patches
Please note that it is a release candidate and it might contain some breakages,
please don't use it for production builds.
Signed-off-by: Evgeniy Didin <Evgeniy.Didin@synopsys.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: arc-buildroot@synopsys.com
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This commit bumps ARC toolchain to arc-2019.09-eng002. We want to
test how new toolchain-eng002 builds packages, so we can make fixes
before release of toolcain.
Please note that it is an engineering build and it might have all
kinds of breakages, please don't use it for production builds
Signed-off-by: Evgeniy Didin <Evgeniy.Didin@synopsys.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: arc-buildroot@synopsys.com
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
This commit bumps ARC toolchain to arc-2019.03-rc1. We want to test
how new toolchain-rc1 builds packages, so we can make fixes before
release of toolcain.
ARC GNU tools of version arc-2019.03-rc1 bring some quite significant
changes like:
* Binutils v2.32.51.20190308 with additional ARC patches
* GCC 8.3.1 with additional ARC patches
* glibc 2.29 with additional ARC patches
Please note that it is a release candidate and it might contain some
breakages, please don't use it for production builds.
Signed-off-by: Evgeniy Didin <Evgeniy.Didin@synopsys.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: arc-buildroot@synopsys.com
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
We are back with 3 binutils version in Buildroot.
Now CFI support is always present for NIOSII.
Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Binutils 2.29 and 2.30 are affected by a bug in ADR and ADRL
pseudo-instruction [1] that was fixed in Binutils 2.31 [2].
* The ADR and ADRL pseudo-instructions supported by the ARM assembler
now only set the bottom bit of the address of thumb function symbols
if the -mthumb-interwork command line option is active.
Due to this issue, we were default to binutils 2.28 for ARM Thumb. But
now that the issue has been fixed in binutils 2.31 and that this
version is the default, the special casing to use 2.28 is no longer
needed.
[1] https://sourceware.org/bugzilla/show_bug.cgi?id=21458
[2] https://sourceware.org/ml/binutils/2018-07/msg00213.html
Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
In commit cd9d58f1fc ("toolchain: bumb
ARC tools to arc-2018.09 release"), the ARC-specific binutils, gcc and
gdb versions were updated to use the 2018.09 release. However, they
are mistakenly pointing to a branch rather than a tag. This commit
fixes that by using the proper release tag.
Signed-off-by: Evgeniy Didin <Evgeniy.Didin@synopsys.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: arc-buildroot@synopsys.com
[Thomas: rework commit log]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
This commit bumps ARC toolchain to arc-2018.09-rc2,
which includes significant changes since arc-2018.09-rc1.
We want to test how new toolchain-rc2 builds packages,
so we can make fixes before release of toolcain.
This makes us closer to toolchain release which will be in a few weeks.
Please note that it is a release candidate and it might
contain some breakages, please don't use it for production builds.
Signed-off-by: Evgeniy Didin <Evgeniy.Didin@synopsys.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: arc-buildroot@synopsys.com
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This commit bumps ARC toolchain to arc-2018.09-rc1.
We want to test how new toolchain-rc1 builds packages,
so we can make fixes before release of toolcain.
ARC GNU tools of version arc-2018.09-rc1 bring some quite significant changes like:
* Binutils v2.31.1 with additional ARC patches
* GCC 8.2.1 with additional ARC patches
* glibc 2.28 with additional ARC patches
Please note that it is a release candidate and it might contain some
breakages, please don't use it for production builds.
Signed-off-by: Evgeniy Didin <didin@synopsys.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: arc-buildroot@synopsys.com
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
As reported by [1], SSP support is missing in the Buildroot toolchain
for microblaze even if it's requested by selecting
BR2_TOOLCHAIN_HAS_SSP config option.
In Buildroot, we are using libssp provided by the C library (glibc,
musl, uClibc-ng) when available. We are not using libssp from gcc.
So for a microblaze glibc based toolchain, the SSP support is enabled
unconditionally by a select BR2_TOOLCHAIN_HAS_SSP.
BR2_microblazeel=y
BR2_TOOLCHAIN_BUILDROOT_GLIBC=y
BR2_KERNEL_HEADERS_4_14=y
BR2_BINUTILS_VERSION_2_30_X=y
BR2_GCC_VERSION_8_X=y
BR2_TOOLCHAIN_BUILDROOT_CXX=y
While building the toolchain, we are building host-binutils which
provide "as" (assembler) and host-gcc-initial wich provide a
minimal cross gcc (C only cross-compiler without any C library).
When SSP support is requested, gcc_cv_libc_provides_ssp=yes is
added to the make command line (see [2] for full details)
With this setting, the SSP support is requested but it's not available
in the end and the toochain build succeed.
When the microblaze toolchain is imported to Biuldroot (2018.05) as
external toolchain with BR2_TOOLCHAIN_EXTERNAL_HAS_SSP set, the build
stop with :
"SSP support not available in this toolchain, please disable BR2_TOOLCHAIN_EXTERNAL_HAS_SSP"
The test is doing the following command line:
echo 'void main(){}' | [...]/host/bin/microblazeel-linux-gcc.br_real -Werror -fstack-protector -x c - -o [...]/build/.br-toolchain-test.tmp
cc1: error: -fstack-protector not supported for this target [-Werror]
When we look at the gcc-final log file (config.log) we can see this
error several time when using the minimal gcc (from host-gcc-initial).
So Why the minimal gcc doesn't support SSP?
When we look at the gcc-initial log file (config.log) we can see an
error with 'as':
configure:23194: checking assembler for cfi directives
configure:23209: [...]microblazeel-buildroot-linux-gnu/bin/as -o conftest.o conftest.s >&5
conftest.s: Assembler messages:
conftest.s:2: Error: CFI is not supported for this target
conftest.s:3: Error: CFI is not supported for this target
conftest.s:4: Error: CFI is not supported for this target
conftest.s:5: Error: CFI is not supported for this target
conftest.s:6: Error: CFI is not supported for this target
conftest.s:7: Error: CFI is not supported for this target
configure:23212: $? = 1
configure: failed program was
.text
.cfi_startproc
.cfi_offset 0, 0
.cfi_same_value 1
.cfi_def_cfa 1, 2
.cfi_escape 1, 2, 3, 4, 5
.cfi_endproc
This is the only relevant difference compared to a nios2 toolchain where
libssp is enabled and available (nios2 is an example).
"CFI" stand for "Control Flow Integrity" and it seems that SSP support
requires CFI target support (see [3] for some explanation).
The SSP support seems to depends on CFI support, but the toolchain
infrastructure is not detailed enough to handle the CFI dependency.
The NiosII toolchains built with binutils < 2.30 are also affected by
this issue.
This patch improve the toolchain infrastructure by adding a new
BR2_PACKAGE_HOST_BINUTILS_SUPPORTS_CFI blind option
Disable SSP support for microblaze entirely.
Disable SSP support for nios2 only with Binutils < 2.30.
Fixes:
https://gitlab.com/free-electrons/toolchains-builder/-/jobs/72006389
[1] https://gitlab.com/free-electrons/toolchains-builder/issues/1
[2] https://git.buildroot.net/buildroot/tree/package/gcc/gcc.mk?h=2018.05#n275
[3] https://grsecurity.net/rap_faq.php
Signed-off-by: Romain Naour <romain.naour@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
[Thomas: adjust how the BR2_PACKAGE_HOST_BINUTILS_SUPPORTS_CFI option
is expressed.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
This enables a riscv64 system to be built with a Buildroot generated
toolchain (gcc >= 7.x, binutils >= 2.30, glibc only).
This configuration has been used to successfully build a qemu-bootable
riscv-linux-4.15 kernel (https://github.com/riscv/riscv-linux.git).
Signed-off-by: Mark Corbin <mark.corbin@embecosm.com>
[Thomas:
- simplify arch.mk.riscv by directly setting GCC_TARGET_ARCH
- simplify glibc.mk changes by using GLIBC_CONF_ENV.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
>From the release note:
This is a point release over the previous 2.31 version, which was
necessary as a source file was accidentally omitted from that
release.
This release also contains a fix for PR gold/23409 where the gold
linker could end up creating duplicate copies of some symbols.
https://sourceware.org/ml/binutils/2018-07/msg00286.html
Signed-off-by: Romain Naour <romain.naour@smile.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
