The autobuilder fails when building for sh because of an alignment issue
in assembly that includes an external data file.
Fixes:
http://autobuild.buildroot.net/results/457b88c4ba85edb42b3a36cce48bf90712f590bf/
This fix has been submitted upstream.
Signed-off-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Fix the following build failure with wolfssl raised since bump to
version 3.12.1 in commit 4d85defa71:
/tmp/instance-4/output-1/build/libuhttpd-3.12.1/src/ssl/openssl.c: In function 'ssl_context_new':
/tmp/instance-4/output-1/build/libuhttpd-3.12.1/src/ssl/openssl.c:174:33: error: 'tls13_ciphersuites' undeclared (first use in this function)
174 | SSL_CTX_set_ciphersuites(c, tls13_ciphersuites);
| ^~~~~~~~~~~~~~~~~~
Fixes:
- http://autobuild.buildroot.org/results/25db1e999950bbee069ba7364a228184a6a77808
Note: we have a reference to a pull request on the 'ssl' repository,
because it is a git submodule of libuhttpd.
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
[yann.morin.1998@free.fr: explain the reference to ssl/pull/1]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Highly parallel host-python3 builds sometimes fail with:
Exception in thread Thread-1:
Traceback (most recent call last):
File "/tmp/instance-3/output-1/host/lib/python3.9/threading.py", line 973, in _bootstrap_inner
self.run()
File "/tmp/instance-3/output-1/host/lib/python3.9/concurrent/futures/process.py", line 317, in run
result_item, is_broken, cause = self.wait_result_broken_or_wakeup()
File "/tmp/instance-3/output-1/host/lib/python3.9/concurrent/futures/process.py", line 376, in wait_result_broken_or_wakeup
worker_sentinels = [p.sentinel for p in self.processes.values()]
File "/tmp/instance-3/output-1/host/lib/python3.9/concurrent/futures/process.py", line 376, in <listcomp>
worker_sentinels = [p.sentinel for p in self.processes.values()]
RuntimeError: dictionary changed size during iteration
During the compile_all.py step of host-python3. This issue is reported
upstream at https://bugs.python.org/issue43498, and while not yet
fixed upstream, a PR was proposed with a possible fix for it. Seems
the PR seems reasonable, let's give it a chance and see if it improves
the situation.
Hopefully Fixes:
http://autobuild.buildroot.net/results/ae6c4ab292589a4e4442dfb0a1286349a9bf4d29/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Replace libjson by jsoncpp for C++ dependency which was wrongly added
by commit 74fc60a267
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Use Buildroot's setting for optimization, zstd's build system
overrides CFLAGS (adds -O3), but MOREFLAGS can override again.
Quick tests show that using -O2 (like buildroot)
is actually a little faster than -O3 on x86_64 Atoms.
Disable the legacy format, these are just needed for
decompressing files created with pre-release version.
Signed-off-by: Norbert Lange <nolange79@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Dynamic build without threads is broken since bump to version 1.5.0 in
commit aad8bbf588
Indeed, dynamic library is built with threads support since
91465e23b2
To fix this build failure, add -nomt if needed however libzstd-nomt is a
"special target that builds a library in single-thread mode _and_
without zstdmt_compress.c". For an unknown reason, this target fails to
build and don't create any symlinks so replace libzstd-{mt,nomt} by
lib-{mt,nomt} even if this will have the side effect of building a
shared and a static version of the library.
Move the existing HAVE_THREAD setting in the same condition.
While at it, also replace "libzstd.a libzstd" by "lib"
Fixes:
- http://autobuild.buildroot.org/results/e609601a0fc91c44d88a12c35b29ce937381462f
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
[Arnout: move HAVE_THREAD in the same condition]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Fix the following build failure raised since bump to version 3.4.7 in
commit bb75c4b541:
/tmp/instance-5/output-1/host/opt/ext-toolchain/bin/../lib/gcc/sparc-buildroot-linux-uclibc/9.3.0/../../../../sparc-buildroot-linux-uclibc/bin/ld: ui/qt/CMakeFiles/qtui.dir/sequence_diagram.cpp.o: undefined reference to symbol '__atomic_compare_exchange_4@@LIBATOMIC_1.0'
/tmp/instance-5/output-1/host/opt/ext-toolchain/bin/../lib/gcc/sparc-buildroot-linux-uclibc/9.3.0/../../../../sparc-buildroot-linux-uclibc/bin/ld: /tmp/instance-5/output-1/host/sparc-buildroot-linux-uclibc/sysroot/lib/libatomic.so.1: error adding symbols: DSO missing from command line
Fixes:
- http://autobuild.buildroot.org/results/6617ee0e0046a0452a1515b89e9c704b1c125ec4
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Workaround for gcc bug 68485 doesn't work anymore since bump to version
3.21 in commit 5b3e721399 because CXXFLAGS
is appended after CFLAGS and does not contain -O0:
/tmp/instance-7/output-1/host/bin/microblazeel-linux-gcc -o rtcm2_json.os -c --sysroot=/tmp/instance-7/output-1/host/microblazeel-buildroot-linux-uclibc/sysroot -pthread -Wall -Wcast-align -Wextra -Wimplicit-fallthrough -Wmissing-declarations -Wmissing-prototypes -Wno-missing-field-initializers -Wno-uninitialized -Wpointer-arith -Wreturn-type -Wstrict-prototypes -Wvla -O0 -Os -g0 -pthread -fPIC -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -I/tmp/instance-7/output-1/host/microblazeel-buildroot-linux-uclibc/sysroot/usr/include/dbus-1.0 -I/tmp/instance-7/output-1/host/microblazeel-buildroot-linux-uclibc/sysroot/usr/lib/dbus-1.0/include rtcm2_json.c
during RTL pass: reload
rtcm2_json.c: In function ‘json_rtcm2_read’:
rtcm2_json.c:267:1: internal compiler error: in gen_reg_rtx, at emit-rtl.c:1155
267 | }
| ^
It should be noted that GPSD_CXXFLAGS was added 5 years ago with commit
b4c050e9af but was not actually used in
that commit or any later.
Fixes:
- http://autobuild.buildroot.org/results/fab33f25b08fa218af91640fdbd8c0dcf1d07228
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Reviewed-by: Giulio Benetti <giulio.benetti@benettiengineering.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
The libcamera option requires libdrm, require and depend on it
when available for libcamera plugin.
Fixes:
- http://autobuild.buildroot.net/results/e3bc1d9aa69af1d2cdeee7709df318fccbaa5476
Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Set back -fPIC in CFLAGS which was wrongly removed as a side effect of
commit c5ca521e23.
Another side-effect of that commit was to remove
-DSYSCONFDIR="\"$(PREFIX)/etc/openzwave/\"" - however, we don't need
SYSCONFDIR to be set, since /etc/openzwave is checked first anyway (see
cpp/src/Options.cpp).
Fixes:
- http://autobuild.buildroot.org/results/161f53574697016ab8bdb91aaaf4a2be5410ebb0
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
This restricts the Erlang RISC-V update from
0bab8c1089 to only enabling RISCV_64
rather than enabling builds on all platforms with toolchains that
include atomics.
This fixes build errors on SPARCv8 and other platforms that look like
this:
/tmp/ccD9Q9F7.s: Assembler messages:
/tmp/ccD9Q9F7.s:230: Error: Architecture mismatch on "cas [%g3],%g2,%g1".
/tmp/ccD9Q9F7.s:230: (Requires v9|v9a|v9b|v9c|v9d|v9e|v9v|v9m|m8; requested architecture is v8.)
Fixes:
http://autobuild.buildroot.net/results/a57/a5779e22c41eb63b8400409f26f96b924fdde1bc/
Signed-off-by: Frank Hunleth <fhunleth@troodon-software.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
armadillo can use lapack or openblas as BLAS provider. LAPACK support is
optional.
This patch
- adds an _ARCH_SUPPORTS variable to check if one is available
- adds an option to choose lapack or openblas as BLAS provider
The choice is required since applications may potentially need lapack.
Signed-off-by: Gwenhael Goavec-Merou <gwenhael.goavec-merou@trabucayre.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Force Release build to avoid the following build failure which is raised
since the addition of the package in commit
4634e84978:
/tmp/instance-3/output-1/host/opt/ext-toolchain/bin/../lib/gcc/i586-buildroot-linux-musl/9.3.0/../../../../i586-buildroot-linux-musl/bin/ld: cannot find -lasan
/tmp/instance-3/output-1/host/opt/ext-toolchain/bin/../lib/gcc/i586-buildroot-linux-musl/9.3.0/../../../../i586-buildroot-linux-musl/bin/ld: cannot find -lubsan
Fixes:
- http://autobuild.buildroot.org/results/83274082879f9599bc106e12a253f4b9229ea00b
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Disable -Werror to avoid the following build failures raised since bump
to version 1.2.2 in commit b436b82411:
/tmp/instance-0/output-1/build/open62541-v1.2.2/arch/network_tcp.c: In function 'connection_recv':
/tmp/instance-0/output-1/build/open62541-v1.2.2/arch/network_tcp.c:96:5: error: conversion to 'unsigned int' from 'int' may change the sign of the result [-Werror=sign-conversion]
96 | UA_fd_set(connection->sockfd, &fdset);
| ^~~~~~~~~
/tmp/instance-6/output-1/build/open62541-v1.2.2/plugins/ua_pubsub_udp.c: In function 'UA_PubSubChannelUDPMC_receive':
/tmp/instance-6/output-1/build/open62541-v1.2.2/plugins/ua_pubsub_udp.c:477:21: error: conversion to '__suseconds_t' {aka 'int'} from 'UA_UInt32' {aka 'unsigned int'} may change the sign of the result [-Werror=sign-conversion]
477 | tmptv.tv_usec = (long int)(timeout % 1000000);
| ^
Fixes:
- http://autobuild.buildroot.org/results/911811de81d8abb2a31feb8f27af1592641c6fbc
- http://autobuild.buildroot.org/results/f0187b3f2d62e955fddeef4e90f84ba4fd642bd2
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
To fix this bug we have to use an undocumented test variable to enable
the nightly target-applies-to-host feature so that the target linker
doesn't get used for host binaries like the build-script.
Fixes:
error: failed to run custom build command for `ripgrep v0.8.1 (/home/buildroot/buildroot/output/build/ripgrep-0.8.1)`
Caused by:
process didn't exit successfully: `/home/buildroot/buildroot/output/build/ripgrep-0.8.1/target/release/build/ripgrep-59eeb7069534e1ef/build-script-build` (exit status: 1)
--- stderr
/home/buildroot/buildroot/output/build/ripgrep-0.8.1/target/release/build/ripgrep-59eeb7069534e1ef/build-script-build: /lib/x86_64-linux-gnu/libc.so.6: version `GLIBC_2.33' not found (required by /home/buildroot/buildroot/output/build/ripgrep-0.8.1/target/release/build/ripgrep-59eeb7069534e1ef/build-script-build)
Details:
https://github.com/rust-lang/cargo/issues/3349https://github.com/rust-lang/cargo/pull/9322
Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Support for tpm2-tools is added by the services/tpm2 module in the
SELinux refpolicy.
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Support for tpm2-totp is added by the services/tpm2 module in the
SELinux refpolicy.
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Support for tpm2-abrmd is added by the services/tpm2 module in the
SELinux refpolicy.
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
GCC 11 defaults to C++17. Fix the following build failure with gcc 11:
In file included from _internal/Source/JSONDefs.h:12,
from _internal/Source/JSONDebug.h:4,
from _internal/Source/JSONNode.h:4,
from _internal/Source/JSONNode.cpp:1:
_internal/Source/JSONDefs/GNU_C.h:58:28: error: ISO C++17 does not allow dynamic exception specifications
58 | #define json_throws(x) throw(x)
| ^~~~~
Fixes:
- http://autobuild.buildroot.org/results/1e66dff705bbb38e7e0f0e5864ce794b4345dcc6
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
OpenCV-4's buildsystem will try to detect ccache and use it if
available. This may yield a system-installed ccache.
However, in Buildroot, ccache is entirely hidden away and handled in the
toolchain wrapper.
Forcibly disable detection of ccache.
Commit 505e7f4771 already applied this
change to opencv3.
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Build with libmaxminddb is broken since bump to version 3.0.5 in commit
464d0be380 because of
785958f9b5
So revert this commit until upstream answer to comment to
https://github.com/SpiderLabs/ModSecurity/issues/2131
Reverting this commit requires autoreconfiguring, which itself causes
lots of warnings as configure.ac queries git to know the version of
various parts of libmodsecurity. However, it turns out that those
versions are only used to be displayed in the output of the configure
script, which is quite useless. The only one that is referenced
elsewhere is LIBINJECTION_VERSION, but it's in fact a different thing:
it is defined by others/libinjection/src/libinjection_sqli.c.
The only variable that was AC_SUBST() and therefore visible elsewhere
was MSC_GIT_VERSION, but it is not used anywhere in the code base,
except in the configure script itself.
Note that one patch is 0001 and the other 0003, because there was
already a 0002 patch.
Fixes:
- http://autobuild.buildroot.org/results/4c639fd967faa06f8ae362bacd38f3409c47267c
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Since upstream commit 5da4da99e0bb3d694e93046207ae4ce3b5545043, in
3.13.0, OP-TEE client unconditionally uses the MMC_IOC_MULTI_CMD
ioctl(). But this ioctl() was only introduced in kernel 4.3, so
earlier kernels did not have this definition.
This was clearly known when upstream started using this ioctl. From
the upstream commit 5da4da99e0bb3d694e93046207ae4ce3b5545043 commit
log:
As MMC_IOC_MULTI_CMD is available since linux 4.3, we should be safe to
use it by default on read / write block operations.
So let's propagate this dependency into Buildroot.
Fixes:
http://autobuild.buildroot.net/results/bd270a6793df499d5f52959e3de17f8a92d88305/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Use pkg-config to find numa to avoid the following build failure when
checking for numa_available:
configure:9667: checking for numa_available in -lnuma
configure:9692: /tmp/instance-7/output-1/host/bin/microblazeel-linux-gcc -o conftest -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -Os -static -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -static conftest.c -lnuma >&5
/tmp/instance-7/output-1/host/opt/ext-toolchain/bin/../lib/gcc/microblazeel-buildroot-linux-uclibc/9.3.0/../../../../microblazeel-buildroot-linux-uclibc/bin/ld: /tmp/instance-7/output-1/host/microblazeel-buildroot-linux-uclibc/sysroot/usr/lib/libnuma.a(libnuma.o): in function `numa_node_to_cpus_v1':
(.text+0x2a80): undefined reference to `__atomic_fetch_and_1'
/tmp/instance-7/output-1/host/opt/ext-toolchain/bin/../lib/gcc/microblazeel-buildroot-linux-uclibc/9.3.0/../../../../microblazeel-buildroot-linux-uclibc/bin/ld: /tmp/instance-7/output-1/host/microblazeel-buildroot-linux-uclibc/sysroot/usr/lib/libnuma.a(libnuma.o): in function `numa_node_to_cpus_v2':
(.text+0x2ddc): undefined reference to `__atomic_fetch_and_1'
collect2: error: ld returned 1 exit status
Fixes:
- http://autobuild.buildroot.org/results/577a63432fba2f9ae1ed2c6c2a77c5ce54ac5521
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
GCC 11 defaults to C++17. Fix the following build failure with gcc 11:
In file included from details/shared-ptr/base.cxx:5:
../odb/details/shared-ptr/base.hxx:38:49: error: ISO C++17 does not allow dynamic exception specifications
38 | operator new (std::size_t, odb::details::share) throw (std::bad_alloc);
| ^~~~~
Fixes:
- http://autobuild.buildroot.org/results/cfd5f92f0aa923815edba5fbfcd5b7b312d9d40e
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
When Bluetooth support is enabled ensure to enable the Audio and
HID plugins.
Signed-off-by: Michael Nosthoff <buildroot@heine.tech>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
When BlueZ is enabled ensure to select the BlueZ HID Plugin for wiimote support.
Signed-off-by: Michael Nosthoff <buildroot@heine.tech>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
When BlueZ is enabled ensure to select the BlueZ Audio Plugins.
Signed-off-by: Michael Nosthoff <buildroot@heine.tech>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
When BlueZ is enabled ensure to select the BlueZ Audio Plugins.
Signed-off-by: Michael Nosthoff <buildroot@heine.tech>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
When enabling the BlueZ plugin select the BlueZ Audio Plugins.
Signed-off-by: Michael Nosthoff <buildroot@heine.tech>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
brltty builds with bluetooth support when bluez is enabled.
Ensure the HID Plugin is enabled.
Signed-off-by: Michael Nosthoff <buildroot@heine.tech>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
The A2DP/AVCRP plugins now have to be explicitly selected.
Signed-off-by: Michael Nosthoff <buildroot@heine.tech>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
the HID plugin now needs to be explictly enabled.
Signed-off-by: Michael Nosthoff <buildroot@heine.tech>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
as discussed on the mailinglist this drops the default y
entry for the now optional plugins and tools.
This might break backward compatibility but gives new users
a smaller initial footprint.
Further add a select for the options that required the HID
plugin at runtime.
Signed-off-by: Michael Nosthoff <buildroot@heine.tech>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
