This reverts commit 0841e2f820 from 2011,
which disabled plugin support in binutils for static toolchains, citing
build system problems. This problem seems to be resolved.
This also reverts part of 45512b003d from
2017, which disabled LTO in gcc for static toolchains, citing problems
on Arch Linux with loading the LTO plugin from a static binary.
Signed-off-by: Chris Copeland <chris@chrisnc.net>
Due to the small flash space on AVR devices the library containing the
standard types in C++ (ˋlibstdc++ˋ) does not get built normally when
enabling the C++ language support.
This option is an easy way to go back to the PC-way where ˋlibstdc++ˋ is
built.
Signed-off-by: Quentin Boswank <qubos@outlook.de>
To build multilib RISC-V toolchain one should use --with-multilib-generator
option instead of --with-multilib-list.
Add corresponding example configuration file.
Signed-off-by: Kirill K. Smirnov <kirill.k.smirnov@gmail.com>
Add zstd to the companion libs witch allows to use lto zstd compression
in a canadian or cross-native enviroment
Signed-off-by: QBos07 <62326551+QBos07@users.noreply.github.com>
Signed-off-by: Quentin Boswank <62326551+QBos07@users.noreply.github.com>
libgccjit is still under development and, despite its name, may also be used for
ahead-of-time compilation.
Documentation can be found on the gcc website:
https://gcc.gnu.org/onlinedocs/jit/internals/index.htmlhttps://gcc.gnu.org/wiki/JIT
With this change it's possible to enable the building of the libgccjit. It's
enabled as a language (with --enable-languages=jit) even if not a language
frontend at all.
The main changes are related to the requirement of having everything host side
built as Position Independent Code (PIC) with --enable-host-shared. GCC has the
needed logic for building its dependencies (mpc, gmp, mpfr, ...) correctly when
built "in-tree", which is not the case with crosstool-ng (see
https://gcc.gnu.org/git/?p=gcc.git;a=commit;h=05048fc29f0)
Signed-off-by: Marc Poulhiès <dkm@kataplop.net>
PIE is not supported by PRU and AVR backends for GCC.
This fixes LD errors when trying to link user programs with a
crosstool-NG toolchain:
/home/dinux/x-tools/avr/lib/gcc/avr/12.2.0/../../../../avr/bin/ld: -pie not supported
collect2: error: ld returned 1 exit status
Signed-off-by: Dimitar Dimitrov <dimitar@dinux.eu>
libsanitizer is only supported on selected architectures. Add
ARCH_SUPPORTS_LIBSANITIZER and have architectures select this option
based on the list of supported configurations from GCC's
libsanitizer/configure.tgt. Support for mips64 was added in GCC12 so
this is an additional condition for the mips architecture.
Fixes#1733
Signed-off-by: Chris Packham <judge.packham@gmail.com>
fix conflict options when compiling multilibs on ARM based architecture.
"--with-arch", "--with-cpu", "--with-fpu", "--with-float", and
"--with-mode" should not be configured with "--with-multilib-list=list"
when configuring multilibs.
Signed-off-by: Chen Tao <t.clydechen@gmail.com>
We don't currently bundle zstd so when performing a canadian build we
need to tell GCC not to enable zstd support for lto otherwise it might
decide to enable it based on the package being installed on the build
machine.
Fixes#1718
Signed-off-by: Chris Packham <judge.packham@gmail.com>
Extend the fix from commit 6b465e15 ("Remove m1 from multilibs for GCC11
on SH arch.") to cover GCC 12 and future releases.
Remove the patch that was added to solve the same problem.
Signed-off-by: Chris Packham <judge.packham@gmail.com>
Rather important option for arm cortex toolchains supporting c++,
avoids pulling in all printf/iostream code by default.
Signed-off-by: Norbert Lange <nolange79@gmail.com>
With libc_headers step before pass-1, there is no need to distinguish
pass-1 and pass-2; they are configured identically (note that with the
current configuration, core pass-2 is only used for win32 - hence, uses
build_libgcc=yes and mode=static).
Signed-off-by: Alexey Neyman <stilor@att.net>
... and the code dependent on them, after the latest wave of obsolete
package removals. This concludes the glorious history of the original
uClibc (non-NG) with lots of kludges removed.
There was a choice here, whether to call the resulting libc "uClibc" or
"uClibc-ng". I opted in favor of giving uClibc-ng the recognition it
deserves, although it had some ripple effect in the ct-ng code.
Signed-off-by: Alexey Neyman <stilor@att.net>
The following versions were marked obsolete in crosstool-ng-1.24.0,
remove them.
- binutils-linaro-2.23.2-2013.10-4
- binutils-linaro-2.24.0-2014.11-2
- binutils-linaro-2.25.0-2015.01-2
- binutils-2.23.2
- binutils-2.24
- binutils-2.25.1
Adjust the milestones now that the old versions have been removed.
Signed-off-by: Chris Packham <judge.packham@gmail.com>
GCC can support using zstd compression for LTO object files. By default
GCC's configure will enable this if libzstd is installed on the machine
building the toolchain. This may be undesirable if the toolchain is to
be used on a different machine. Add an option to control zstd usage and
set the default to the same as the current behaviour (i.e. auto).
Fixes#1579
Signed-off-by: Chris Packham <judge.packham@gmail.com>
Explicitly passing --disable-tm-clone-registry causes gcc to create a
crtbegin.o with a zero-sized .init_array/.fini_array. This in turn
causes ld to complain.
Make CC_GCC_TM_CLONE_REGISTRY a tristate so if it's not explicitly
enabled we can let ./configure decide.
Fixes#1531
Fixes: 1e21a302 ("gcc: Add CT_CC_GCC_TM_CLONE_REGISTRY config")
Signed-off-by: Chris Packham <judge.packham@gmail.com>
This commit adds a new gcc config `CT_CC_GCC_TM_CLONE_REGISTRY` that
enables the GCC transactional memory clone registry feature for libgcc.
Note that the gcc option to control this feature is only available in
gcc 10 and above.
(see gcc commit 5a4602805eb3ebddbc935b102481e63bffc7c5e6)
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
GCC11 somehow has different set of multilibs on SH arch than what GCC10 had.
In particular:
$ gcc10 -print-multi-lib | sed -r -e 's/@/ -/g;'
.;
mb; -mb
m2; -m2
m2e; -m2e
m4; -m4
m4-single; -m4-single
m4-single-only; -m4-single-only
mb/m2; -mb -m2
mb/m2e; -mb -m2e
mb/m4; -mb -m4
mb/m4-single; -mb -m4-single
mb/m4-single-only; -mb -m4-single-only
mb/m2a; -mb -m2a
mb/m2a-single; -mb -m2a-single
$ gcc11 -print-multi-lib | sed -r -e 's/@/ -/g;'
.;
mb; -mb
m2; -m2
m2e; -m2e
m4; -m4
m4-single; -m4-single
m4-single-only; -m4-single-only
mb/m1; -mb -m1
mb/m2; -mb -m2
mb/m2e; -mb -m2e
mb/m4; -mb -m4
mb/m4-single; -mb -m4-single
mb/m4-single-only; -mb -m4-single-only
mb/m2a; -mb -m2a
mb/m2a-single; -mb -m2a-single
mb/m1 fails to build libgcc as libgcc uses opcodes that were not
available in SH-1: libgcc/config/sh/lib1funcs.S uses 'bt/s' and 'dt'
instructions that, according to https://antime.kapsi.fi/sega/files/h12p0.pdf become available in the SH-2 only.
So I removed mb/m1 from the multilibs fog GCC11 and SH arch.
Another option would be to try not to build libgcc for this combination
of the gcc version and archichecture, but I thought this fix would be
more robust.
Signed-off-by: Nik Konyuchenko <spaun2002mobile@gmail.com>
This has gone unnoticed for a while, it looks like all configurations
have probably been using the default floats behaviour. Newer kconfig
tells us the default value is not in the choice so lets fix this in
preparation.
Signed-off-by: Chris Packham <judge.packham@gmail.com>
This adds another mode to do_gcc_core_backend that builds libstdc++
against an alternate libc implementation.
Signed-off-by: Keith Packard <keithp@keithp.com>
While we may use build machine's library when building a simple cross,
we won't have it for the host machine in a canadian build. Until there
are separate selectables for the build/host components, just build our
own local zlib.
Signed-off-by: Alexey Neyman <stilor@att.net>
This allows us to include the component-to-package relation in the
generated kconfig files and make use of that information in the
show-config.sh script.
Signed-off-by: Alexey Neyman <stilor@att.net>
Was previously selected by GCC_5_or_later - but now that one is an
auto-generated option that does not select anything; now HAS_LIBMPX
was made dependent on GCC_5_or_later, but its default was not set to y.
Signed-off-by: Alexey Neyman <stilor@att.net>
Also:
- Move companion_* to comp_* to match the kconfig symbols
- Replace bootstrap with former gen-versions.sh
- Fold *.in.2 into their respective first parts; this moves common
options to the end - if it is undesirable, inclusion of *.in
can be moved where *.in.2 used to be (but that will also move
version selection after common options).
- Retire addToolVersion.sh (may later replace with a more
comprehensive script that tries to download the added tarballs,
copy the patches and try to apply them, and create a version.desc).
Signed-off-by: Alexey Neyman <stilor@att.net>
Make this behavior default in case the core gcc backend is used
for final compiler (i.e., for baremetal configurations). Not
setting this option breaks canadian baremetal configurations,
and not setting it makes little sense at all in any baremetal
configuration (since in baremetal we don't have any libc to begin
with).
Signed-off-by: Alexey Neyman <stilor@att.net>
... when determining if it can be linked statically, and if Python
scripting should default to y.
Prompted by a failure of i686-w64-mingw32,nios2-spico-elf sample
on a system where configure didn't report static linking support.
Signed-off-by: Alexey Neyman <stilor@att.net>
Same as the base release as long as they applied.
MUSL patches didn't, removed.
Also, unobsolete Linaro GCC5 now that they rolled out a new release.
Signed-off-by: Alexey Neyman <stilor@att.net>
Adding new tristate configuration for TLS (Thread Local Storage) to
add "--enable-tls" (y), "--disable-tls" (n) or nothing (m).
Signed-off-by: Jasmin Jessich <jasmin@anw.at>
