These instructions are for 64-bit load/store. On ARMv5TE, the CPU
requires addresses to be aligned to 64-bit. When misaligned, behavior is
undefined (effectively either loads the same word twice on LDRD, or
corrupts surrounding memory on STRD).
On ARMv6 and newer, unaligned access is safe.
Removing these instructions for ARMv5TE is necessary, because GCC
ignores alignment information in pointers and does unsafe optimizations
that have shown up as bugs in various places.
This patch was originally added more than 11 years ago in commit b050f87d13,
but got lost 6 years ago, when gcc 9.1 was added in 88c07c6552.
This primarily affects the kirkwood and ixp4xx targets
Signed-off-by: Felix Fietkau <nbd@nbd.name>
GCC 14+ fails to build due to libatomic specific -march handling.
This build error triggers only with glibc and not with musl libc
which is default.
It seems that this patch from GCC14 was forgotten when GCC15 support was
being added [1].
[1] 44ef343500
Fixes: 68cb84183e ("toolchain: add support for GCC 15.1")
Signed-off-by: Robert Marko <robimarko@gmail.com>
All patches automatically refreshed.
Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18600
Signed-off-by: Robert Marko <robimarko@gmail.com>
There is no practical value in keeping GCC11 around, as even OpenWrt 23.05
uses GCC12 as the default one, so drop it.
Signed-off-by: Robert Marko <robimarko@gmail.com>
Remove upstreamed patches:
- 020-MIPS-Include-missing-mips16.S-in-libgcc-lib1funcs.S.patch
- 021-Reuse-scratch-registers-generated-by-LRA.patch
All other patches are automatically refreshed.
Signed-off-by: Shiji Yang <yangshiji66@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/18891
Signed-off-by: Robert Marko <robimarko@gmail.com>
During the build of perl, the following ICE was reported in
https://github.com/openwrt/packages/issues/24565 when targeting PowerPC:
during RTL pass: reload
blocksort.c: In function 'mainSort.isra':
blocksort.c:1011:1: internal compiler error: in patch_jump_insn, at cfgrtl.cc:1303
1011 | }
| ^
0x7d49cee29d8f __libc_start_call_main
../sysdeps/nptl/libc_start_call_main.h:58
0x7d49cee29e3f __libc_start_main_impl
../csu/libc-start.c:392
Please submit a full bug report, with preprocessed source (by using -freport-bug).
Please include the complete backtrace with any bug report.
See <http://bugs.openwrt.org/> for instructions.
The same issue also caused the CI failures in
https://github.com/openwrt/packages/pull/26501.
The issue only occurs with GCC 14.2.0, but not with the head of the
releases/gcc-14 maintenance branch; a bisect found that this patch fixes
it.
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
Link: https://github.com/openwrt/openwrt/pull/18797
Signed-off-by: Robert Marko <robimarko@gmail.com>
Backport patch from upstream GCC 14 branch which fixes linking with
MIPS16 on the pistachio target.
This fixes the following link problem:
```
/builder/shared-workdir/build/staging_dir/toolchain-mipsel_24kc+24kf_gcc-14.2.0_musl/lib/gcc/mipsel-openwrt-linux-musl/14.2.0/../../../../mipsel-openwrt-linux-musl/bin/ld.bfd: ./liblua.so: undefined reference to `__mips16_ledf2'
/builder/shared-workdir/build/staging_dir/toolchain-mipsel_24kc+24kf_gcc-14.2.0_musl/lib/gcc/mipsel-openwrt-linux-musl/14.2.0/../../../../mipsel-openwrt-linux-musl/bin/ld.bfd: ./liblua.so: undefined reference to `__mips16_call_stub_df_2'
/builder/shared-workdir/build/staging_dir/toolchain-mipsel_24kc+24kf_gcc-14.2.0_musl/lib/gcc/mipsel-openwrt-linux-musl/14.2.0/../../../../mipsel-openwrt-linux-musl/bin/ld.bfd: ./liblua.so: undefined reference to `__mips16_muldf3'
```
Link: https://github.com/openwrt/openwrt/pull/18688
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
This reverts commit 57841c83d9.
This is completely breaking the inital GCC configuration and most likely
was not even compile tested, so revert until fixed.
Signed-off-by: Robert Marko <robimarko@gmail.com>
I've observed configuration drift for GCC between musl and glibc
(especially it's final stage):
# musl
lt_cv_prog_compiler_static_works=yes
lt_cv_prog_compiler_static_works_CXX=yes
lt_cv_sys_max_cmd_len=1572864
# glibc
lt_cv_prog_compiler_static_works=no
lt_cv_prog_compiler_static_works_CXX=no
lt_cv_sys_max_cmd_len=512
These changes should prevent this issue in future:
export lt_cv_prog_compiler_static_works=yes
export lt_cv_prog_compiler_static_works_CXX=yes
export lt_cv_sys_max_cmd_len=1572864
Also:
- provide custom autotools/libtool variables via properly named
variable ("GCC_CONFIGURE_VARS"),
- move variables from "GCC_MAKE" to "GCC_CONFIGURE_VARS"
(at this moment only "gcc_cv_libc_provides_ssp=yes" for musl),
- propagate it's usage for both "./configure" and "make".
Signed-off-by: Konstantin Demin <rockdrilla@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18646
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Xcode 16.3 defines TARGET_OS_MAC, it was not defined in prior versions.
zutil.h conditionally defines fdopen as NULL when this macro is defined,
resulting in the following build error:
/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk/usr/include/_stdio.h:318:7: e>
318 | FILE *fdopen(int, const char *) __DARWIN_ALIAS_STARTING(__MAC_10_6, __IPHONE_2_0, __DARWIN_ALIAS(fdopen));
| ^
./zutil.h:147:33: note: expanded from macro 'fdopen'
147 | # define fdopen(fd,mode) NULL /* No fdopen() */
In Xcode 16.2 and earlier, TARGET_OS_MAC was not defined so this entire
block was ignored, gcc and gdb used to compile and work fine.
This may have been used for compatibility with older versions of macOS,
but is no longer needed. By pure luck, the build worked fine for a long
time, because it did not properly detect macOS.
Fixed by removing the check for TARGET_OS_MAC.
Note that since Xcode 16.3, an entire set of TARGET_OS macros
are now defined, most of which are set to 0:
TARGET_OS_LINUX 0
TARGET_OS_MAC 1
TARGET_OS_OSX 1
Signed-off-by: Georgi Valkov <gvalkov@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18467
Signed-off-by: Robert Marko <robimarko@gmail.com>
GCC has changed musl dynamic linker name from
ld-musl-loongarch-lp64d.so.1 to ld-musl-loongarch64.so.1 recently [1].
Meanwhile musl 1.2.5 only supports the new name. So it's better to follow
the new name.
[1] https://gcc.gnu.org/git/?p=gcc.git;a=commit;h=8bccee51f0deac64b79cd9ad75df599422f4c8ff
Signed-off-by: Weijie Gao <hackpascal@gmail.com>
I noticed that CONFIG_GDB was suddenly appearing in the main menuconfig
menu despite the fact that it should be visible only when TOOLCHAINOPTS
is selected and under a dedicated menu.
After some trial and error, it seems that this was caused by the recent
addition of GCC_USE_DEFAULT_VERSION, and after even more trial and error
it gets fixed as soon GCC_USE_DEFAULT_VERSION is placed after GCC_VERSION.
So, lets simply put GCC_USE_DEFAULT_VERSION after GCC_VERSION.
Fixes: 501ef81040 ("config: select KERNEL_WERROR if building with default GCC version")
Signed-off-by: Robert Marko <robimarko@gmail.com>
At the moment we have to manually follow the default GCC version
also in config/Config-kernel.in. This tends to be forgotten at GCC
version bumps (just happened when switching from version 12 to 13).
Instead, introduce a hidden Kconfig symbol which implies KERNEL_WERROR
in toolchain/gcc/Config.in where it is visible for developers changing
the default version.
Also remove the explicit default on BUILDBOT to avoid a circular
dependency and also because buildbots anyway implicitly always select
the default GCC version.
Reference: https://github.com/openwrt/openwrt/pull/15064
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk/usr/include/c++/v1/__locale:550:5: error: '__abi_tag__' attribute only applies to structs, variables, functions, and namespaces
_LIBCPP_INLINE_VISIBILITY
^
/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk/usr/include/c++/v1/__config:891:37: note: expanded from macro '_LIBCPP_INLINE_VISIBILITY'
# define _LIBCPP_INLINE_VISIBILITY _LIBCPP_HIDE_FROM_ABI
^
/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk/usr/include/c++/v1/__config:870:26: note: expanded from macro '_LIBCPP_HIDE_FROM_ABI'
__attribute__((__abi_tag__(_LIBCPP_TOSTRING(_LIBCPP_ODR_SIGNATURE))))
Fixed using backport of upstream commits [1-2] as discussed here
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111632#c21
[1] Include safe-ctype.h after C++ standard headers, to avoid over-poisoning
https://gcc.gnu.org/git/?p=gcc.git;a=commit;h=9970b576b7e4ae337af1268395ff221348c4b34a
[2] libcc1: fix <vector> include
https://gcc.gnu.org/git/?p=gcc.git;a=commit;h=5213047b1d50af63dfabb5e5649821a6cb157e33
Signed-off-by: Georgi Valkov <gvalkov@gmail.com>
Use GCC 13 instead of GCC 12 by default.
All target kernels are building with GCC 13.
Most packages from the feed are building fine.
The root file systems is getting a little bit smaller for MIPS 32 BE
and aarch64.
With GCC 12 I got these sizes for lantiq/xrx200:
7,005,867 openwrt-lantiq-xrx200-tplink_tdw8970-initramfs-kernel.bin
With GCC 13 I got these sizes for lantiq/xrx200:
6,989,754 openwrt-lantiq-xrx200-tplink_tdw8970-initramfs-kernel.bin
With GCC 12 I got these sizes for armsr/armv8:
13,083,836 openwrt-armsr-armv8-generic-ext4-combined.img.gz
4,900,240 openwrt-armsr-armv8-generic-ext4-rootfs.img.gz
20,142,592 openwrt-armsr-armv8-generic-kernel.bin
With GCC 13 I got these sizes for armsr/armv8:
13,068,966 openwrt-armsr-armv8-generic-ext4-combined.img.gz
4,893,078 openwrt-armsr-armv8-generic-ext4-rootfs.img.gz
20,142,592 openwrt-armsr-armv8-generic-kernel.bin
Signed-off-by: Nick Hainke <vincent@systemli.org>
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Some package may needs to enable compatibility option based on the GCC
version.
Currently the GCC version is set based on the default value and doesn't
actually reflect the real value provided by the external toolchain if
used.
Fix this by correctly detecting the GCC version in the external
toolchain and set the correct value in CONFIG_GCC_VERSION.
A new option is added in menuconfig to manually set the GCC version if
needed.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
The PKG_CPE_ID links to NIST CPE version 2.2.
Assign PKG_CPE_ID to all remaining package which have a CPE ID.
Not every package has a CPE id.
Related: https://github.com/openwrt/packages/issues/8534
Signed-off-by: Alexander Couzens <lynxis@fe80.eu>
RISC-V has no support for subword atomic operations; code currently
generates libatomic library calls.
This patch changes the default behavior to fast inline subword atomic
calls that do not require libatomic.
Signed-off-by: Tianling Shen <cnsztl@immortalwrt.org>
While improving access path analysis a typo happened. Now it can happen
that gcc misscompiles. The patch is fixing the issue. However, also
other gcc versions 10.2+ are affected. They also should be bumped or the
fix should be backported.
For more bug information have a look at:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109585
Signed-off-by: Nick Hainke <vincent@systemli.org>
LTO object files will now be compressed using zstd.
Compressing debug sections with zstd will have to wait for GCC 13,
which adds support for -gz=zstd.
NOTE: wiping the ccache is strongly recommended, not doing so might
yield build error later on:
"lto1: internal compiler error: original not compressed with zstd"
Signed-off-by: Andre Heider <a.heider@gmail.com>
This was apparently introduced to recreate the toolchain (wipe
staging_dir/toolchain*, but keep build_dir/toolchain*, followed by a
`make toolchain/compile`).
But it leaves leftovers and causes re-links to happen at src_install phase,
because of the changed paths, possibly adding yet another source of issues.
With the prior commits removing various hacks related to the "initial"
folder we can remove installing it twice altogether.
The recreated toolchain is exactly the same as before.
Signed-off-by: Andre Heider <a.heider@gmail.com>
A modern glibc only links against libgcc_eh for its tests or when building it
static, which doesn't happen here.
Reverts a hunk of: a3edea1b "add support for alternative C libraries..."
Signed-off-by: Andre Heider <a.heider@gmail.com>
Instead of having two different ways to pass flags to the gcc build
process, add them as configure args, which is a reliable way to let
gcc pass them around to its various pieces.
Also add CXXFLAGS, since gcc started to use c++ for itself recently
(~10 years ago now).
Signed-off-by: Andre Heider <a.heider@gmail.com>
Spell out what we want to enable or disable. This prevents host libs to leak in,
so everyone get the same feature set.
Signed-off-by: Andre Heider <a.heider@gmail.com>
Kconfig docs say:
> The default value deliberately defaults to 'n' in order to avoid
> bloating the build.
Apply this rule everywhere, to avoid more cloning of bad examples
Signed-off-by: Tony Butler <spudz76@gmail.com>
Use STAGING_DIR_HOST to reference staging host directory instead of
hardcoding it to default path.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Config evaluation require default with if to be put before the generic
default config with no condition. Putting the default config before any
conditional default results in always selecting the non conditional one.
This results in the version be hardcoded to gcc 12 even if gcc 11 is
selected in the Advanced build options.
Fix this by putting the gcc 12 default option as last after ANY
conditional default config.
Fixes: d9de5252a4 ("toolchain/gcc: switch to version 12 by default")
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
This was build tested with all core packages on all targets
successfully.
Most packages from the feed are also building fine.
This was run tested on the following systems:
* lantiq/xrx200 musl
* pistachio/generic musl
* sunxi/cortexa53 musl
* x86/64 musl
* x86/64 glibc
* armvirt/64 musl
The size of the images stays more or less the same for MIPS BE and
aarch64. I haven't tested other architectures.
With GCC 11 I got these sizes for lantiq/xrx200:
7,219,848 openwrt-lantiq-xrx200-tplink_tdw8970-initramfs-kernel.bin
7,472,208 openwrt-lantiq-xrx200-tplink_tdw8970-squashfs-sysupgrade.bin
With GCC 12 I got these sizes for lantiq/xrx200:
7,217,355 openwrt-lantiq-xrx200-tplink_tdw8970-initramfs-kernel.bin
7,406,674 openwrt-lantiq-xrx200-tplink_tdw8970-squashfs-sysupgrade.bin
The sysupgrade image is probably padded. The initramfs image is 0.03% smaller.
With GCC 11 I got these sizes for armvirt/64:
4,143,943 openwrt-armvirt-64-default-rootfs.tar.gz
10,887,176 openwrt-armvirt-64-Image
24,911,880 openwrt-armvirt-64-Image-initramfs
4,141,572 openwrt-armvirt-64-rootfs.cpio.gz
4,255,854 openwrt-armvirt-64-rootfs-ext4.img.gz
3,391,178 openwrt-armvirt-64-rootfs-squashfs.img.gz
With GCC 12 I got these sizes for armvirt/64:
4,142,778 openwrt-armvirt-64-default-rootfs.tar.gz
10,887,176 openwrt-armvirt-64-Image
24,911,880 openwrt-armvirt-64-Image-initramfs
4,138,105 openwrt-armvirt-64-rootfs.cpio.gz
4,255,463 openwrt-armvirt-64-rootfs-ext4.img.gz
3,390,390 openwrt-armvirt-64-rootfs-squashfs.img.gz
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
The version check which sets GCC_VERSION_FILE to the correct value only
worked when the advanced options menu was active and not when it was not
active.
Thank you Tony Butler for the fix.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
glibc 2.36 changed the definition of enum fsconfig_command, it now
collides with the same definition from sys/mount.h. Remove the include
of linux/fs.h This still compiled with musl too.
This backports a patch which is already in the stable branch of GCC 11
and GCC 12.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Every minor version bump of a major version will result in a huge patch
diff because of the moving of all the patches from version e.g. 11.2.0 to
11.3.0. This commit only use the major version for the patch folders to
differentiate between the different gcc versions. This will significantly
improve the reviewing of the smaller version bump patches and help to see
what really changed in a minor version bump.
Signed-off-by: Nick Hainke <vincent@systemli.org>
libstdcxx-dual-abi needs to be enabled to actually support C++11 ABI.
Enable the config flag to also permit support of .NET 6 development on
OpenWrt.
Signed-off-by: Ivan Maslov <avenger_msoft@mail.ru>
[ reword commit description and title ]
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
GCC 12.1 is out. Add support for it.
Deleted (upstreamed):
011-v12-configure-define-TARGET_LIBC_GNUSTACK-on-musl.patch
931-libffi-fix-MIPS-softfloat-build-issue.patch
Deleted (unneeded?)
970-macos_arm64-building-fix.patch
Other patches manually rebased due to C++ conversion and consequent file name
changing (.c to .cc).
Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
