aboutsummaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
...
* Revert "binutils: Disable glob for better portability"Chris Packham2024-03-011-4/+0
| | | | | | | | | | | | | | This reverts commit 57f59092852dff18fbda68fdbf23f850ad182c40. This was originally added so that a toolchain could be built on a newer system but run on an older one. With the benefit of hindsight that is probably the wrong approach. The best way of achieving that goal would be to use docker/podman container to provide an environment that is the same as the oldest supported system and build inside that. The resulting toolchain should be compatible with the old system and the new one. Closes #2094 Signed-off-by: Chris Packham <judge.packham@gmail.com>
* Ignore a bad id command to allow dockerQuentin Boswank2024-02-181-1/+1
| | | | Signed-off-by: Quentin Boswank <qubos@outlook.de>
* Remove Oracle UEK linux dependencies.Cupertino Miranda2024-02-1811-16/+9
| | | | | | | | | This patch removes any dependency to the Oracle UEK Linux sources since it can be easily replaced by a standard kernel explicitly pointing to the exact kernel version, as the toolchain building only requires the kernel headers. Signed-off-by: Cupertino Miranda <cupertino.miranda@oracle.com>
* glibc: Fix THREAD_SETMEM/THREAD_SETMEM_NC for movq in x86_64Artem Panfilov2024-02-061-0/+60
| | | | | | | | | | | | | | | This patch resolves compilation issues with GCC versions 12 and glibc 2.17. It corrects the constraints used in the THREAD_SETMEM and THREAD_SETMEM_NC macros for the movq instruction in the x86_64 architecture. Backported from: https://github.com/bminor/glibc/commit/b1ec623ed50bb8c7b9b6333fa350c3866dbde87f Closes #1825 Signed-off-by: Artem Panfilov <artem.panfilov@nokia.com>
* Update config.{guess,sub}Bryan Hundven2024-02-022-113/+205
| | | | | | from: git://git.savannah.gnu.org/config.git Signed-off-by: Bryan Hundven <bryanhundven@gmail.com>
* gdb: Explicitly point to target's libgmp for native buildAlexey Brodkin2024-01-301-0/+8
| | | | | | | | | | | | | | | Starting from GDB 11.x, gmp is needed as a dependency to build full gdb. And by default build system of native GDB will try to link with libgmp of the build host. And to make sure that doesn't happen we need to specify location of the target's sysroot so that library search starts from there. Which we do in that change. Fixes [1] & [1]. [1] https://github.com/crosstool-ng/crosstool-ng/issues/2084 [2] https://github.com/crosstool-ng/crosstool-ng/issues/1656 Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
* gcc: Add patch to resolve libsanitizer build issueChris Packham2024-01-302-0/+270
| | | | | | | | | | | | libsanitizer has problems intercepting crypt() and crypt_r() with newer glibcs. Bring in an upstream patch that drops support for these from ASAN. llvm/llvm-project@d7bead8 https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111057 Fixes #2083 Signed-off-by: Chris Packham <judge.packham@gmail.com>
* newlib: Add 4.4.0.20231231Chris Packham2024-01-302-0/+4
| | | | | | Add the latest snapshot of newlib. Signed-off-by: Chris Packham <judge.packham@gmail.com>
* Fix zlib build on win32 near enviromentsQuentin Boswank2023-12-181-2/+0
| | | | | | | We set some variables for linking zlib to '' which we shouldn't do. Let the Makefile sort these things out. Signed-off-by: Quentin Boswank <qubos@outlook.de>
* zstd: Bring in upstream change for cygwin supportChris Packham2023-12-181-0/+68
| | | | | | | Bring in the upstream change that should allow building zstd on cygwin. Fixes #1974 Signed-off-by: Chris Packham <judge.packham@gmail.com>
* gdb: Add 14.1 releaseChris Packham2023-12-183-1/+9
| | | | | | | https://sourceware.org/pipermail/gdb-announce/2023/000137.html Fixes #2077 Signed-off-by: Chris Packham <judge.packham@gmail.com>
* zlib: Add 1.3Chris Packham2023-12-182-0/+8
| | | | | | http://madler.net/pipermail/zlib-announce_madler.net/2023/000014.html Signed-off-by: Chris Packham <judge.packham@gmail.com>
* Remove configure time check for rsyncChris Packham2023-12-182-7/+0
| | | | | | | Now that we're checking for rsync in Kconfig we can remove the check from ./configure. Signed-off-by: Chris Packham <judge.packham@gmail.com>
* config: Detect rsync presence at config timeChris Packham2023-12-183-2/+13
| | | | | | | | | | | | | | | | | | Linux kernel versions newer than 5.3 require rsync in order to export the UAPI headers. Commit f441a6bf ("linux: Add dependency on rsync for Linux >= 5.3") attempted to address this with a check that runs when crosstool-ng is built. That had the downside that if crosstool-ng was built and packaged on a machine that had rsync then run on a machine that didn't then the build would fail due to the missing rsync. Conversely if the first machine didn't have rsync installed when crosstool-ng was built then we would not offer newer kernel versions. We can address this by checking for rsync when the toolchain configuration is updated using some functionality in the newer Kconfig that we've updated to previously. Fixes #1940 Signed-off-by: Chris Packham <judge.packham@gmail.com>
* elf2flt: Update to v2023.09Chris Packham2023-12-083-590/+1
| | | | | | | Update to the v2023.09 release and drop the patches that have been applied upstream. Signed-off-by: Chris Packham <judge.packham@gmail.com>
* binutils: Bring in upstream performance fixChris Packham2023-12-081-0/+156
| | | | | | Bring in an upstream fix for a performance issue in the 2.41 release. Signed-off-by: Chris Packham <judge.packham@gmail.com>
* binutils: Add binutils 2.41Hans-Christian Noren Egtvedt2023-12-0810-0/+622
| | | | | | | | | | https://sourceware.org/pipermail/binutils/2023-July/128719.html Forward ported all patches from binutils 2.40, with only minor adjustment to match new upstream code in patch 0007-poison-system-directories.patch. Signed-off-by: Hans-Christian Noren Egtvedt <hegtvedt@cisco.com>
* scripts/build/binutils: Update path for libbfdChris Packham2023-12-081-1/+1
| | | | | | | | | As of binutils 2.41 libbfd.a is not placed directly in the output directory. Fortunately the libtool .libs location seems to have been in place for some time so we can update the path without worrying about backwards compatibility. Signed-off-by: Chris Packham <judge.packham@gmail.com>
* CI: Add sample for hppa-unknown-linux-gnuChris Packham2023-11-253-0/+8
| | | | | | Add sample for hppa/parisc and include in CI builds. Signed-off-by: Chris Packham <judge.packham@gmail.com>
* config/arch: Mark hppa/parisc as experimentalChris Packham2023-11-251-0/+2
| | | | | | | There are a number of things that don't currently work notably uClibc, C++ and GDB. Mark this architecture as experimental. Signed-off-by: Chris Packham <judge.packham@gmail.com>
* Adjust GUESS tuple for PA-RISC linux.John David Anglin2023-11-251-7/+9
| | | | | | | | | | | | 64-bit kernels can support running both 32 and 64-bit user code. Select hppa-unknown-linux-$LIBC or hppa64-unknown-linux-$LIBC depending on whether compiler defines LP64 or not. Always select hppa-unknown-linux-$LIBC with 32-bit kernel. This will generate 32-bit PA 1.1 code. If PA 2.0 code is desired, user can use a compiler option to select it. Signed-off-by: John David Anglin <dave.anglin@bell.net>
* Map parisc64 to hppa64 in config.sub.John David Anglin2023-11-251-0/+8
| | | | | | | Debian and Gentoo use hppa/hppa64 for the PA-RISC ports. These are the proper CPU designations for configuring most packages. Signed-off-by: John David Anglin <dave.anglin@bell.net>
* Add support for 32 and 64-bit PA-RISC linux build targets.John David Anglin2023-11-253-1/+26
| | | | | | | Note: The 64-bit target lacks a glibc port and doesn't build. Also, there is no uclibc support. Signed-off-by: John David Anglin <dave.anglin@bell.net>
* linux: Add 6.6.1 bump LTSChris Packham2023-11-1523-56/+64
| | | | | | | | | | | | Add 6.6.1 Bump 4.14.326 -> 4.14.329 Bump 5.10.197 -> 5.10.200 Bump 5.15.133 -> 5.15.138 Bump 5.4.257 -> 5.4.260 Bump 6.1.55 -> 6.1.62 Bump 6.5.5 -> 6.5.11 Signed-off-by: Chris Packham <judge.packham@gmail.com>
* strace: Add 6.6Chris Packham2023-11-153-4/+4
| | | | | | | | https://lists.strace.io/pipermail/strace-devel/2023-October/011369.html Add the 6.6 release, delete 6.5. Signed-off-by: Chris Packham <judge.packham@gmail.com>
* mingw-w64: add v11.0.1Mateusz Mikuła2023-11-153-0/+38
| | | | | | This adds access to various new APIs and brings several bug fixes Signed-off-by: Mateusz Mikuła <mati865@gmail.com>
* gdb: Use correct arguments for static linkingChris Packham2023-11-022-1/+10
| | | | | | | | | | | | As of version 13.x GDB uses libtool for linking instead of g++ these take different arguments for static linking. Commit 6146b5a6 ("use -all-static when building a static gdb") attempted to deal with this but had the effect of causing older GDB versions to fail to build statically. Add a new internal flag GDB_CC_LD_LIBTOOL and use this to decide whether to pass `-static` or `-all-static`. Fixes #2053 Signed-off-by: Chris Packham <judge.packham@gmail.com>
* m4/ctng_python.m4: Use AC_PATH_PROGS for full pathChris Packham2023-11-021-1/+1
| | | | | | | | | | | | | | | We use the information from various configure time checks to populate paths.sh. The paths used are all absolute except for the python binary. In the switch to a more comprehensive check for python by commit fa05153e ("Make checking for python more predictable.") we ended up using AC_CHECK_PROGS which checks for the program on the path and sets the variable to the name of the program. This makes python inconsistent with the other programs and seems to cause problems for MSYS2. Use AC_PATH_PROGS instead which does the same check but sets the variable to the absolute name of the program Fixes #2047 Signed-off-by: Chris Packham <judge.packham@gmail.com>
* m4/ctng_python.m4: Fix up URLChris Packham2023-11-021-1/+1
| | | | | | Seems to be the victim of an overzealous search and replace. Signed-off-by: Chris Packham <judge.packham@gmail.com>
* Add gnatls & gnatlinkc-grant2023-11-022-2/+3
| | | | | | | | add gnatls and gnatlink to list of tools since it is needed to support Ada just like gnatmake and gnatbind Update crosstool-NG.sh & TODO Signed-off-by: c-grant <60671494+c-grant@users.noreply.github.com>
* Add more caching to the CIQuentin Boswank2023-10-043-7/+19
| | | | | | | | | | | more specificaly to the tarballs download. The function CT_Fetch now touches the already existing files to be comparable to the not used ones that can araise when a package is updated. This comparsion is needed because if it would not exist the tarball would grow in size due to not used but still cached packages. This would take time but is definitly something to worry about. Signed-off-by: Quentin Boswank <qubos@outlook.de>
* duma: Drop 2_5_15Chris Packham2023-10-046-236/+0
| | | | | | Drop duma 2_5_15 which was marked as obsolete in the last release. Signed-off-by: Chris Packham <judge.packham@gmail.com>
* cloog: Drop 0.18.0Chris Packham2023-10-042-7/+0
| | | | | | Drop cloog 0.18.0 which was marked as obsolete in the last release. Signed-off-by: Chris Packham <judge.packham@gmail.com>
* zlib: Drop 1.2.12Chris Packham2023-10-044-73/+0
| | | | | | Drop zlib 1.2.12 which was marked obsolete in the last release. Signed-off-by: Chris Packham <judge.packham@gmail.com>
* newlib: Drop obsolete versionsChris Packham2023-10-0426-373/+0
| | | | | | Drop versions of newlib that were marked obsolete in the last release. Signed-off-by: Chris Packham <judge.packham@gmail.com>
* expat: Drop 2.4.1Chris Packham2023-10-042-17/+0
| | | | | | Drop expat 2.4.1 which was marked obsolete in the last release. Signed-off-by: Chris Packham <judge.packham@gmail.com>
* isl: Drop 0.11.1Chris Packham2023-10-042-11/+0
| | | | | | Drop isl 0.11.1 which was marked obsolete in the last release. Signed-off-by: Chris Packham <judge.packham@gmail.com>
* mpfr: Drop obsolete versionsChris Packham2023-10-046-51/+0
| | | | | | Drop versions of mpfr that were marked obsolete in the last release. Signed-off-by: Chris Packham <judge.packham@gmail.com>
* picolibc: Drop obsolete versionsChris Packham2023-10-0412-420/+0
| | | | | | Drop versions that were marked as obsolete in the last release. Signed-off-by: Chris Packham <judge.packham@gmail.com>
* strace: Add 6.5, remove old versionChris Packham2023-10-045-5/+5
| | | | | | | Add 6.5, mark 6.4 as obsolete, remove 5.16. https://lists.strace.io/pipermail/strace-devel/2023-September/011351.html Signed-off-by: Chris Packham <judge.packham@gmail.com>
* mpc: Add 1.3.1 drop old versionsChris Packham2023-10-047-10/+5
| | | | | | | | | https://www.multiprecision.org/ Add 1.3.1. Mark 1.2.1 as obsolete. Remove 1.0.3 and 1.1.0. Fixes #2030 Signed-off-by: Chris Packham <judge.packham@gmail.com>
* linux: Add 6.5.5 bump LTSChris Packham2023-10-0423-56/+64
| | | | | | | | | | | | | Add 6.5.5 Bump 6.4 -> 6.4.16 Bump 6.1.35 -> 6.1.55 Bump 5.15.11 -> 5.15.133 Bump 5.10.185 -> 5.10.197 Bump 5.4.248 -> 5.4.257 Bump 4.19.287 -> 4.19.295 Bump 4.14.319 -> 4.14.326 Signed-off-by: Chris Packham <judge.packham@gmail.com>
* m68k: Support MMUChris Packham2023-09-244-0/+8
| | | | | | | | | | Some m68k variants do have a MMU so the architecture can set ARCH_USE_MMU. That means we can have a m68k-unknown-linux-gnu configuration and that m68k-unknown-uclinux-uclibc needs to select LIBC_UCLIBC_NG since it's no longer the default. Fixes #2040 Signed-off-by: Chris Packham <judge.packham@gmail.com>
* picolibc: Bump to 1.8.3Lukas Rusak2023-09-243-4/+4
| | | | Signed-off-by: Lukas Rusak <lorusak@gmail.com>
* use -all-static when building a static gdbChris Copeland2023-09-241-2/+8
| | | | | | | | | | | gdb is linked with libtool, which has a different meaning for -static, and -all-static must be used to get a static executable. The binutils build script already uses this option for static builds. Also remove unnecessary -static from cflags for the gdb build. Signed-off-by: Chris Copeland <chris@chrisnc.net>
* allow plugins and LTO with static toolchainsChris Copeland2023-09-242-3/+0
| | | | | | | | | | | | This reverts commit 0841e2f820de7a4dca1ddc52b04bf834fff2806b 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 45512b003d04b5a89c5c3bb6b674683d82b87f42 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>
* Add a musl variable LIBC_MUSC_EXTRA_CFLAGSNorbert Lange2023-09-242-2/+12
| | | | Signed-off-by: Norbert Lange <nolange79@gmail.com>
* fix toolchain build with linker pluginsNorbert Lange2023-09-241-1/+11
| | | | | | | | | older binutils dont automatically pick up plugins, but need to manually use wrappers like gcc-ar. This fix allows to compile the host toolchain with -ftlo on debian stretch. Signed-off-by: Norbert Lange <nolange79@gmail.com>
* moxiebox: Use newlib 4.2Chris Packham2023-09-241-0/+1
| | | | | | | | | | | | | | moxie-unknown-moxiebox has problems building with newlib 4.3 ld: /lib/libc.a(libc_a-closer.o): in function `_close_r': newlib/libc/reent/closer.c:47: undefined reference to `_close' There are some Makefile changes in newlib 4.3 and it's likely previously this config was picking up `_close` from libsim.a. For now just pin the newlib version back to 4.2 in the moxie-unknown-moxiebox config. Resolves #2036 Signed-off-by: Chris Packham <judge.packham@gmail.com>
* gcc: Bring in upstream fixes for CVE-2023-4039Chris Packham2023-09-17100-0/+15138
| | | | | | | | | Bring in the fixes for GCC 7 through 13. https://rtx.meta.security/mitigation/2023/09/12/CVE-2023-4039.html https://developer.arm.com/Arm%20Security%20Center/GCC%20Stack%20Protector%20Vulnerability%20AArch64 Signed-off-by: Chris Packham <judge.packham@gmail.com>
generated by cgit v1.2.3 (git 2.39.1) at 2025-08-08 03:08:26 +0000