This is what GCC manual says [1]:
-------------------------->8----------------------
--enable-__cxa_atexit
Define if you want to use __cxa_atexit, rather than atexit,
to register C++ destructors for local statics and global objects.
This is essential for fully standards-compliant handling of destructors,
but requires __cxa_atexit in libc.
This option is currently only available on systems with GNU libc
...
-------------------------->8----------------------
Important disadvantages of a simple atexit() are that [2]:
-------------------------->8----------------------
1999 C Standard only requires that the implementation support 32
registered functions, although most implementations support many more.
More important it does not deal at all with the ability in most implementations
to remove DSOs from a running program image by calling dlclose
prior to program termination.
-------------------------->8----------------------
Also it seems like all libc's we support in Buildroot (Glibc, uClibc and musl)
support __cxa_at_exit() so enable it unconditionally.
FWIW if we look around we'll see:
1. In OpenEmbedded it is enabled for everything except gcc-cross-initial: [3], [4]
2. In Crosstool-NG it is enabled by default: [5]
3. In OpenWrt it is disabled only for uClibc, otherwise enabled: [6]
So I think we should be good with it as well.
[1] https://gcc.gnu.org/install/configure.html
[2] https://itanium-cxx-abi.github.io/cxx-abi/abi.html#dso-dtor-motivation
[3] https://github.com/openembedded/openembedded-core/blob/master/meta/recipes-devtools/gcc/gcc-configure-common.inc#L59
[4] https://github.com/openembedded/openembedded-core/blob/master/meta/recipes-devtools/gcc/gcc-cross-initial.inc#L23
[5] https://github.com/crosstool-ng/crosstool-ng/blob/master/config/cc/gcc.in#L270
[6] https://github.com/openwrt/openwrt/blob/master/toolchain/gcc/common.mk#L170
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Nicolas Cavallari <Nicolas.Cavallari@green-communications.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Mark Corbin <mark.corbin@embecosm.com>
Cc: Romain Naour <romain.naour@gmail.com>
Cc: Peter Korsgaard <peter@korsgaard.com>
Cc: Bernd Kuhls <bernd.kuhls@t-online.de>
Cc: Claudiu Zissulescu <claziss@synopsys.com>
Cc: Cupertino Miranda <cmiranda@synopsys.com>
Cc: Vineet Gupta <vgupta@synopsys.com>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
We used to exclude GCC's test-suite for quite some time now
mostly for the sake of space reduction.
But:
1. On each GCC version bump we need to revise that functionality
as we need to accommodate changes in GCC sources and this
couldn't be automated
2. The space reduction is significant, but not huge. The two test
suites together take up 290MB, out of 660MB total for GCC (each of
them times two because there is the -initial and -final copy).
However, whenever we build GCC, we also have kernel headers (about
900MB) and a libc (e.g. glibc is 250MB). So at best, it saves less
than 20%.
3. It doesn't really save on build time either.
Below are timings of 2 runs on my laptop:
a) Vanilla master:
--------------------->8---------------------
time make host-gcc-final
real 7m15.114s
user 19m36.611s
sys 2m26.927s
--------------------->8---------------------
b) master + testsuite:
--------------------->8---------------------
time make host-gcc-final
real 7m59.860s
user 20m21.668s
sys 2m36.618s
--------------------->8---------------------
From figures above it's seen that difference is ~45 seconds
or ~10%. On both host-gcc-initial and -final we may save ~1.5
minutes... but these are not the only components we build and
compared to a total toolchain build time IMHO it is not that
much time to care especially traded for maintenance costs
on GCC version bumps.
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Bernd Kuhls <bernd.kuhls@t-online.de>
Cc: Romain Naour <romain.naour@gmail.com>
Cc: Peter Korsgaard <peter@korsgaard.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
[Arnout: add explanation about size impact.]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Since commit 8f8e9162fa ("package/gcc:
do not mourn avr32 for too long..."), in which we dropped AVR32
support, the BR2_GCC_SUPPORTS_FINEGRAINEDMTUNE is always
'y'. Therefore, it is totally useless to keep this option around, and
this commit removes it, making the corresponding code unconditional
along the way.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Reviewed-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
In commit 325bb37942, support for the
Blackfin architecture was removed. This was our only use of
BR2_GCC_TARGET_CPU_REVISION, and since this config option somewhat
complicates the calculation of the --with-cpu/-mcpu option values,
let's drop it.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The BR2_GCC_TARGET_* configuration variables are copied to
corresponding GCC_TARGET_* variables which may then be optionally
modified or overwritten by architecture specific makefiles.
All makefiles must use the new GCC_TARGET_* variables instead
of the BR2_GCC_TARGET_* versions.
Signed-off-by: Mark Corbin <mark.corbin@embecosm.com>
[Thomas: simplify include of arch/arch.mk]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Remove 0860-cilk-fix-build-without-wchar.patch since cilk has
been removed from gcc.
Rebase 0900-remove-selftests.patch on gcc 8.1.0
Remove upstream patches 0001-m68k-coldfire-pr68467.patch and
0892-microblaze-Revert.patch.
Use bz2 release tarball only for gcc 4.9 since all other version
now use xz tarball.
Runtime tested on x86_64 with binutils 2.30, gcc 8.1.0,
kernel headers 4.14 and glibc 2.27.
Buildtime tested with glibc with all supported architectures.
Buildtime tested on x86_64 with musl and uClibc-ng
Signed-off-by: Romain Naour <romain.naour@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
The HOST_GCC_COMMON_WRAPPER_TARGET_FLOAT_ABI variable was incorrectly
used as HOST_GCC_COMMON_WRAPPER_TARGET_FLOATABI_, which would always
be empty.
In practice, this block of code is only used on Blackfin (which has
BR2_GCC_ARCH_HAS_CONFIGURABLE_DEFAULTS unset) and Blackfin doesn't use
float-abi, so this block of code is in fact useless. But even if it's
useless, it's better to have it without typos.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Instead of DL_DIR, the package should now use $(PKG)_DL_DIR to ease the
transition into a new directory structure for DL_DIR.
This commit has been generated with the following scripts:
for i in $(find . -iname "*.mk"); do
if ! grep -q "\$(DL_DIR)" ${i}; then
continue
fi
pkg_name="$(basename $(dirname ${i}))"
[ "${pkg_name}" = "package" ] && continue
raw_pkg_name=$(echo ${pkg_name} | tr [a-z] [A-Z] | tr '-' '_')
pkg_dl_dir="${raw_pkg_name}_DL_DIR"
sed -i "s/\$(DL_DIR)/\$($pkg_dl_dir)/" ${i}
done
Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Upstream gcc now has support for ARC, as of 7.x, with 7.2 being
usable, so let's allow its selection.
GCC from Synopsys GitHub still has a lot of fixes and improvements
most of which will be a part of the next upstream release but let's
see how it goes and if we see vanilla GCC doesn't require way too many
back-ported patches we'll be able to get rid of Synopsys GitHub here
at all.
As of today the only really required back-port is the one for non-uClibc
libc.
Back-ported from upstream master with help of:
-------------------->8--------------------
git format-patch 927c0132251f -1 -- . ':!gcc/ChangeLog' ':!libgcc/ChangeLog'
-------------------->8--------------------
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Currently the possibility to choose the NaN encoding is conditional to
having a sufficiently recent gcc version.
Which means that the architecture selection depends on the gcc version.
But that's opposite to what we've always done in Buildroot: the software
versions are conditional to the architecture options. There is nothing
we can do about the hardware: it is there, we can't change it, while we
can restrict ourselves to using software that is working on said
hardware.
Thus, we inverse the logic, to move the condition onto the software
side: whenever NaN-2008 are selected, we restrict the toolchain
selection to at least a gcc-4.9.
But now, the option with the NaN type is always set, so we must enclose
the code in gcc.mk inside a HAS_NAN_OPTION condition, as is already done
for the external toolchain case.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
float128 is now required on powerpc64le with glibc >= 2.26 toolchains.
Fixes:
running configure fragment for sysdeps/gnu
running configure fragment for sysdeps/powerpc/powerpc64le
checking if [...]powerpc64le-buildroot-linux-gnu-gcc supports binary128 floating point type... no
checking if the target machine is at least POWER8... yes
configure: error: *** binary128 floating point type (GCC >= 6.2) is required on powerpc64le.
In order to prepare the glibc bump to this version, we enable
binary128 floating point in gcc. This require at least gcc >= 6.2.
Disable gcc 4.x and 5.x for internal toolchain.
Signed-off-by: Romain Naour <romain.naour@gmail.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
All patches continue to apply with no changes. 7.2.0 is a bugfix
release of the 7.x branch.
The only change that is not a simple bump is that the 7.2.0 tarball is
now available xz-compressed instead of bz2-compressed.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
GCC 4.9.x requires an old version of ISL to support graphite, and this
old version of ISL is not sufficient to support graphite in modern GCC
versions.
Since GCC 4.9.x is getting older, and to keep things simple, we stop
supporting graphite on old GCC < 5.x. This way, we will be able to
bump ISL to a version that is suitable for modern GCC versions.
In addition, this allows to drop the dependency on cloog, which was
only needed for GCC 4.9.
Signed-off-by: Andrey Yurovsky <yurovsky@gmail.com>
[Thomas: rework commit log.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
MIPS32 support different FP modes (32,xx,64), so give the user the
opportunity to choose between them. That will cause host-gcc to be built
using the --with-fp-32=[32|xx|64] configure option. Also the
-mfp[32|xx|64] gcc option will be added to TARGET_CFLAGS and to the
toolchain wrapper.
FP mode option shouldn't be used for soft-float, so we add logic in the
toolchain wrapper if -msoft-float is among the arguments in order to not
append the -fp[[32|xx|64] option, otherwise the compilation may fail.
Information about FP modes here:
- https://sourceware.org/binutils/docs/as/MIPS-Options.html
- https://dmz-portal.imgtec.com/wiki/MIPS_O32_ABI_-_FR0_and_FR1_Interlinking#5._Generating_modeless_code
Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
MIPS supports two different NaN encodings, legacy and 2008. Information
about MIPS NaN encodings can be found here:
https://sourceware.org/binutils/docs/as/MIPS-NaN-Encodings.html
NaN legacy is the only option available for R2 cores and older.
NaN 2008 is the only option available for R6 cores.
R5 cores can have either NaN legacy or NaN 2008, depending on the
implementation. So, if the user selects a generic R5 target architecture
variant, we show a choice menu with both options available. For well
known R5 cores we directly select the NaN enconding they use.
Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Drop the following patches:
* the Xtensa patches 870 and 871 are upstream now
* patch 942 was backported to GCC 6 branch
Note, that a bz2 release tarball is not provided anymore and is replaced by
a xz tarball file.
Signed-off-by: Jörg Krause <joerg.krause@embedded.rocks>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Since things are no longer installed in $(HOST_DIR)/usr, the callers
should also not refer to it.
This is a mechanical change with
git grep -l '\$(HOST_DIR)/usr' | xargs sed -i 's%\$(HOST_DIR)/usr%$(HOST_DIR)%g'
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Since things are no longer installed in $(HOST_DIR)/usr, the callers
should also not refer to it.
This is a mechanical change with
git grep -l '$(HOST_DIR)/usr/bin' | xargs sed -i 's%$(HOST_DIR)/usr/bin%$(HOST_DIR)/bin%g'
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
... instead of re-computing them over-and-over-again.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Similar to arc architecture this is required.
Reported-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Add support for OpenRISC. See here for more details about
OpenRISC http://openrisc.io.
All buildroot included upstream binutils versions are supported.
Gcc support is not upstream, to be able to enable musl C library
support later, we use the branch with musl support.
At the moment it is possible to build a musl based toolchain,
but bootup in Qemu fails.
Gdb is only working to debug bare-metal code, there is no support
for gdbserver/gdb on Linux, yet.
[Peter: drop ?= for GCC_SOURCE]
Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Tested-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Building this minimal defconfig
BR2_TOOLCHAIN_BUILDROOT_MUSL=y
BR2_GCC_VERSION_6_X=y
fails:
In file included from ../../../../libmpx/mpxrt/mpxrt.c:54:0:
../../../../libmpx/mpxrt/mpxrt.c: In function 'read_mpx_status_sig':
../../../../libmpx/mpxrt/mpxrt.h:52:42: error: invalid application of
'sizeof' to incomplete type 'struct _libc_fpstate'
#define XSAVE_OFFSET_IN_FPMEM sizeof (struct _libc_fpstate)
To fix disable libmpx for musl builds, other projects did the same:
3ec2211548http://git.alpinelinux.org/cgit/aports/commit/main/gcc/APKBUILD?id=1830e485126ea9a95d763317fb0c508c1ff297d2
Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The blind option BR2_GCC_NEEDS_MPC was used to distinguish gcc versions
that rely on the mpc library and the ones that don't. But since a while
already, all the versions we support do need the mpc library. So this
symbol isn't needed anymore.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The current BR2_GCC_ENABLE_TLS can cause users to make incorrect
choices, and is not very useful. This options allows to decide whether
we pass --enable-tls or --disable-tls to gcc, to enable or disable
support for Thread Local Storage.
Its behavior is:
- The option is default to "y" but only exists if we're using
uClibc/NPTL or glibc.
- When we're using uClibc, the option can be disabled.
So, in practice, this means that currently:
- TLS support is always on for glibc
- TLS support is on by default for uClibc/NPTL, but can be disabled in
the configuration. This is in fact bad and causes the build failure
reported in bug #7424 (this bug is still reproducible on master)
- TLS support is always disabled for uClibc/no-thread and
uClibc/linuxthreads.
- TLS support is always disabled for musl. This does not cause any
build failure, but musl can use TLS support, and therefore be more
efficient. According to
http://www.openwall.com/lists/musl/2012/10/04/1, "Note that if you've
been building gcc with --disable-tls, __thread was already working
but gets emulated (very poorly; it's slow and will abort() if it runs
out of memory) through libgcc.".
So, this commit completely removes the BR2_GCC_ENABLE_TLS and instead
makes the right choice inside gcc.mk directly:
- TLS support enabled for glibc, musl and uClibc/NPTL
- TLS support in other cases, i.e uClibc/no-thread and
uClibc/linuxthreads.
We have intentionally *not* added the option to
Config.in.legacy. Indeed, the new behavior is *exactly* the same as the
older behavior, with the exception of:
- People can no longer disable TLS support in uClibc/NPTL, which was
anyway causing a build failure and therefore was not used.
- TLS support is now enabled on musl, but people using musl already had
BR2_GCC_ENABLE_TLS not set, so they wouldn't get the legacy warning.
Fixes bug #7424.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Latest musl release supports ppc64 architecture (both big endian and
little endian), so this commit adds support for this.
Since musl implements the ELFv2 ABI for both big-endian and
little-endian PowerPC64, we have to force using this ABI on PowerPC64
big endian (normally elfv1 is the default).
Also, only gcc 6.x has the necessary changes to support musl on PowerPC
64, so we restrict the gcc version selection accordingly.
Tested with Qemu for big endian and little endian configurations.
Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
[Thomas: add comment about the ABI flag in gcc.mk, rework commit log.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Some architectures, f.e. Blackfin doesn't support to configure GCC with
--with-cpu to set some CPU specific default CFLAGS (-mcpu=foo). Use a
hidden config symbol to give a hint which architecture supports it,
otherwise add defaults to toolchain wrapper for internal toolchains.
Idea from Thomas Petazzoni.
Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
[Thomas:
- simplify the Config.in logic with just one option named
BR2_GCC_ARCH_HAS_CONFIGURABLE_DEFAULTS, defined in package/gcc in one
place.
- improve the organization of the code and name of variables.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
When we build the list of patches to include to compute the hash for
ccache, a typo precented the special gcc-initial/ and gcc-final/
directories from a global patch dir to be included in the calculation.
Fix that by properly expanding the $(PKG) variable.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: He Chunhui <hchunhui@mail.ustc.edu.cn>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
libquadmath requires wchar.
So, turn to positive logic and complete it to only enabling quadmath
support when it is available.
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
When building host gcc, patches stored in global patches directories are
skipped. This patch fixes the unexpected behavior.
Signed-off-by: Chunhui He <hchunhui@mail.ustc.edu.cn>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
[Thomas:
- rename the loop variable from 'D' to 'patchdir'
- add some additional comments
- remove final ; at end of loop when applying the patches, since it's
not needed]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
gfortran supports all options supported by gcc, so it can and should be called
via the toolchain wrapper.
Signed-off-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
BR2_TOOLCHAIN_BUILDROOT_WCHAR is only defined when uclibc is selected,
whereas BR2_USE_WCHAR is always defined. Due to this, we were disabling
quadmath support even with glibc or musl.
So, use BR2_USE_WCHAR to drive the gcc libquadmath option.
In addition, invert the logic of the condition to use positive logic,
and rework the comment to no longer mention gcc 4.6: libquadmath still
exists, and it still requires wchar support.
Signed-off-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
In commit f4682cf933, a hash of the patches applied to gcc was created
to make sure that ccache can properly detect when the toolchain has
changed. The patches applied to gcc consist of the buildroot patches in
package/gcc, but also potentially patches in BR2_GLOBAL_PATCH_DIR.
However, the path to the patches in BR2_GLOBAL_PATCH_DIR was corrected
incorrectly, because it misses a /. So instead of:
$(BR2_GLOBAL_PATCH_DIR)/gcc-initial/*.patch
it would look for
$(BR2_GLOBAL_PATCH_DIR)gcc-initial/*.patch
In other words, if BR2_GLOBAL_PATCH_DIR doesn't end with /, the patches
in BR2_GLOBAL_PATCH_DIR are not taken into account in the ccache hash.
To fix, add the missing /
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Cc: He Chunhui <hchunhui@mail.ustc.edu.cn>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This reverts commit c44cf2cc97.
Now that xtensa gas don't move literals into .init and .fini this fix is
no longer needed.
See upstream commit
https://sourceware.org/git/?p=binutils-gdb.git;a=commitdiff;h=4111950f363221c4641dc2f33bea61cc94f34906,
which was backported to all supported binutils version, under the name
*-xtensa-fix-.init-.fini-literals-moving.patch.
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
[Thomas: add more details in the commit log, as suggested by Yann
E. Morin, and using information provided by Max.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
We have removed gcc 4.5 support since a long time, so this commit
removes dead code that was only used when building a toolchain based
on gcc 4.5.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
See https://gcc.gnu.org/gcc-5/changes.html
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
As reported by Steven Noonan, the variable recently introduced in the
package infrastructure to exclude certain parts of an archive from
being extracted is <pkg>_EXCLUDES, not <pkg>_TAR_EXCLUDES. However,
the gcc code was incorrectly using <pkg>_TAR_EXCLUDES. This commit
fixes that.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reported-by: Steven Noonan <steven@uplinklabs.net>
During the gcc-initial build, we already pass
gcc_cv_libc_provides_ssp=yes explicitly when SSP support will be
available in the C library: at this point in time the C library is not
yet built, so gcc cannot detect if it will support SSP or not.
However, it turns out that there are some situations for which it is
also useful to tell gcc explicitly whether the SSP support is
available or not: the gcc logic to decide whether uClibc has SSP
support or not is broken since uClibc-ng bumped the glibc version it
pretends to be.
So, this commit makes sure that we explicitly pass
gcc_cv_libc_provides_ssp both to gcc-initial and gcc-final, and that
we're always passing either 'yes' or 'no'.
Fixes:
http://autobuild.buildroot.org/results/778/778e6309ba834cc70f8243a4f6c664c0bcaeb7c5/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
If an externally built (non-Buildroot) toolchain also wraps the toolchain
executables, there is a risk that it will also use the '.real' extension.
To minimise this risk, use a more buildroot-specific extension instead:
'.br_real', so we can detect that the external toolchain is built using
Buildroot and get to the raw toolchain binaries.
[Peter: reword description]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Disable libsanitizer for sparc64, too. Same problem as for
sparc, see https://bugs.busybox.net/show_bug.cgi?id=7951
Reported-By: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
$(wildcard ...) in make doesn't sort the files, so the order of the
hashed files is not predictable. Therefore, the ccache hash could
change from one build to another. We don't want that, so sort the
files explicitly.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
One of the gcc configure options that we hash for ccache is
--with-pkgversion which is set to something like Buildroot
2015.11-git-00426-ge7e7e4f - i.e., it will change with every buildroot
commit. That's obviously not wanted, so substitute this away.
Also add a \n to the printf so the output is a bit more readable.
[Peter: update documentation to match]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
When building in a different output directory than the original build,
there will currently be a lot of ccache misses because in many cases
there is some -I/... absolute path in the compilation. Ccache has an
option CCACHE_BASEDIR to substitute absolute paths with relative paths,
so they wil be the same in the hash (and in the output).
Since there are some disadvantages to this path rewriting, it is made
optional as BR2_CCACHE_USE_BASEDIR. It defaults to y because the
usefulness of ccache is severely reduced without this option.
In addition to CCACHE_BASEDIR, we also substitute away the occurences
of $(HOST_DIR) in the calculation of the compiler hash. This is done
regardless of the setting of BR2_CCACHE_USE_BASEDIR because it's
quite harmless.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Our current ccache disables hashing of the compiler executable itself,
because using the default 'mtime' doesn't work in buildroot: we always
rebuild the compiler, so the mtime is always different, so the cache
always misses.
However, in the current situation, if a user changes the compiler
configuration (which would result in the compiler generating different
object files than before) and does 'make clean all', ccache may in fact
reuse object files from the previous run. This rarely gives problems,
because
(1) the cache expires quite quickly (it's only 1GB by default),
(2) radically changing compiler options will cause cache misses because
different header files are used,
(3) many compiler changes (e.g. changing -mtune) have little practical
effect because the resulting code is usually still compatible,
(4) we currently don't use CCACHE_BASEDIR, and almost all object files
will contain an absolute path (e.g. in debug info), so when
building in a different directory, most of it will miss,
(5) we do mostly build test, and many of the potential problems only
appear at runtime.
Still, when ccache _does_ use the wrong cached object files, the
effects are really weird and hard to debug. Also, we want reproducible
builds and obviously the above makes builds non-reproducible. So we
have a FAQ entry that warns against using ccache and tells the user to
clear the cache in case of problems.
Now that ccache is called from the toolchain wrapper, it is in fact
possible to at least use the 'mtime' compiler hash for the external
toolchain and for the host-gcc. Indeed, in this case, the compiler
executable comes from a tarball so the mtime will be a good reference
for its state. Therefore, the patch (sed script) that changes the
default from 'mtime' to 'none' is removed.
For the internal toolchain, we can do better by providing a hash of
the relevant toolchain options. We are only interested in things that
affect the compiler itself, because ccache also processes the header
files and it doesn't look at libraries because it doesn't cache the
link step, just compilation. Everything that affects the compiler
itself can nicely be summarised in $(HOST_GCC_FINAL_CONF_OPTS). Of
course, also the compiler source itself is relevant, so the source
tarball and all the patches are included in the hash. For this purpose,
a new HOST_GCC_XTENSA_OVERLAY_TAR is introduced.
The following procedure tests the ccache behaviour:
Use this defconfig:
BR2_arm=y
BR2_CCACHE=y
make
readelf -A output/build/uclibc-1.0.6/libc/signal/signal.os
-> Tag_CPU_name: "ARM926EJ-S"
Now make menuconfig, change variant into BR2_cortex_a9
make clean; make
readelf -A output/build/uclibc-1.0.6/libc/signal/signal.os
-> Tag_CPU_name: "ARM926EJ-S"
should be "Cortex-A9"
After this commit, it is "Cortex-A9".
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Cc: Danomi Manchego <danomimanchego123@gmail.com>
Cc: Károly Kasza <kaszak@gmail.com>
Cc: Samuel Martin <s.martin49@gmail.com>
Cc: Romain Naour <romain.naour@openwide.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
We have a toolchain wrapper for external toolchain, but it is also
beneficial for internal toolchains, for the following reasons:
1. It can make sure that BR2_TARGET_OPTIMIZATION is passed to the
compiler even if a package's build system doesn't honor CFLAGS.
2. It allows us to do the unsafe path check (i.e. -I/usr/include)
without patching gcc.
3. It makes it simpler to implement building each package with a
separate staging directory (per-package staging).
4. It makes it simpler to implement a compiler hash check for ccache.
The wrapper is reused from the external toolchain. A third CROSS_PATH_
option is added to the wrapper: in this case, the real executable is in
the same directory, with the extension .real.
The creation of the simple symlinks is merged with the creation of the
wrapper symlinks, otherwise part of the -gcc-ar handling logic would
have to be repeated.
The complex case-condition could be refactored with the one for the
external toolchain, but then it becomes even more complex because
they each have special corner cases. For example, the internal
toolchain has to handle *.real to avoid creating an extra indirection
after host-gcc-{final,initial}-rebuild.
Instead of creating the .real files, it would also have been possible
to install the internal toolchain in $(HOST_DIR)/opt, similar to what
we do for the external toolchain. However, then we would also have to
copy things to the sysroot and do more of the magic that the external
toolchain is doing. So keeping it in $(HOST_DIR)/usr/bin is much
simpler.
Note that gcc-initial has to be wrapped as well, because it is used for
building libc and we want to apply the same magic when building libc.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Cc: Fabio Porcedda <fabio.porcedda@gmail.com>
Cc: Jérôme Oufella <jerome.oufella@savoirfairelinux.com>
Reviewed-by: Romain Naour <romain.naour@openwide.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
xtensa libgcc can't be built with -mtext-section-literals flag, now
coming from TARGET_CFLAGS, because it needs to emit literals to
.init/.fini sections, which is not currently supported.
Filter -mtext-section-literals flag out of GCC_COMMON_TARGET_CFLAGS.
Suggested-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
