Adding extended attribute support for the mtd tools when the attr
package is selected. This is needed for SELinux support.
Signed-off-by: Clayton Shotwell <clayton.shotwell@rockwellcollins.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
We can delete the patch, as it was integrated upstream.
Signed-off-by: Simon Marchi <simon.marchi@polymtl.ca>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
In OpenCV, OpenGL is only used by highgui module.
OpenGL support is done using extensions from 3rd party framework: either
Qt5OpenGL with Qt5 (with GL support only, not GLES); or gtkglext (which
is not available in Buildroot) with gtk2
So, make OpenGL knob a sub-option of the Qt5 support option.
Note: we enclose both the GUI toolkit choice and the GL option in an
if-block, so that the GL option gets properly indented; having it
depend on WITH_QT5 is not enough, because it does not directly follow
it, so kconfig would not consider it for indenting.
Signed-off-by: Samuel Martin <s.martin49@gmail.com>
[yann.morin.1998@free.fr: tweak commit log about the if-block]
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
OpenCV now also supports gtk3 as a GUI toolkit, in addition to gtk2,
but only one may be enabled at a time.
So, add gtk3 in the choice to select the GUI toolkit.
Signed-off-by: Samuel Martin <s.martin49@gmail.com>
[yann.morin.1998@free.fr: drop the superfluous depends-on for the
kconfig symbol, since they're no longer needed now we depend-on rather
than select]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Starting with the 2.4.6 release, OpenCV supports either Qt4 or Qt5 as GUI
toolkit, so add Qt5 in the GUI toolkit choice.
When Qt4 is enabled (and thus Qt5 is hidden and disabled), no need to
show a comment stating "Qt5 support needs Qt5", because Qt5 is not
selectable.
Conversely, when Qt5 is enabled and Qt4 is not, then no need to show a
comment stating "Qt4 support needs Qt4", because enabling Qt4 would
disable Qt5.
So, we only show the comments when neither toolkit is enabled.
Signed-off-by: Samuel Martin <s.martin49@gmail.com>
Cc: "Yann E. MORIN" <yann.morin.1998@free.fr>
[yann.morin.1998@free.fr: split-out the Qt5 hunk from the
switch-selects-to-depends hunk]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Currently, we only support either Qt4 or gtk2, but OpenCV also
additionally supports Qt5 or gtk3, making for a choice of four
toolkits, one of: Qt4, Qt5, gtk3, gtk2 (and obviously, none).
Since Buildroot does not support coexistence of Qt4 and Qt5, we
can no longer select one and depend on the other, like so:
config BR2_PKG_OCV_WITH_QT4
bool "Qt4"
depends on !BR2_PKG_QT5
select BR2_PKG_QT
config BR2_PKG_OCV_WITH_QT5
bool "Qt5"
depends on !BR2_PKG_QT
select BR2_PKG_QT5
otherwise, we'd get a circular dependency chain in Kconfig, which would
complain with:
package/opencv/Config.in:57:error: recursive dependency detected!
package/opencv/Config.in:57: choice <choice> contains symbol BR2_PKG_OCV_WITH_QT5
package/opencv/Config.in:111: symbol BR2_PKG_OCV_WITH_QT5 depends on BR2_PKG_QT
package/qt/Config.in:5: symbol BR2_PKG_QT is selected by BR2_PKG_OCV_WITH_QT
package/opencv/Config.in:98: symbol BR2_PKG_OCV_WITH_QT is part of choice <choice>
Instead, we need to depend on either Qt version.
So, to have a consistent choice, we make all support for GUI toolkits
actually depend on the toolkit, rather than select it.
Signed-off-by: Samuel Martin <s.martin49@gmail.com>
[yann.morin.1998@free.fr: split-out the switch-selects-to-depends hunk
from the add-qt5 hunk]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
In OpenCV, only one GUI toolkit may be used at any one time, so group
the two existing options into a choice to make this situation explicit.
This will also be useful when we later add support for Qt5 and gtk3.
Suggested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Samuel Martin <s.martin49@gmail.com>
[yann.morin.1998@free.fr: tweak commit log]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Drop the minor version from the Kconfig symbol, so we can seamlessly
update the versions without having to handle legacy stuff.
Note: not adding legacy handling, as we haven't had any release with
those symbols yet.
Reported-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Chris Becker <goabonga@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
refclock_pcf.c contains code using the tm_gmtoff member of struct tm, which
is only available on uClibc if it is built with __UCLIBC_HAS_TM_EXTENSIONS__.
This change date back to:
commit 7129da009c
Author: Eric Andersen <andersen@codepoet.org>
Date: Sat Jan 18 21:27:22 2003 +0000
Merge a bunch of stuff over from the tuxscreen buildroot, with
many updates to make things be more consistant.
-Erik
But nowadays our uClibc configs DO enable __UCLIBC_HAS_TM_EXTENSIONS__, so
it is no longer needed and can be dropped.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Drop sed line which no longer changes anything as upstream has changed to
use strrchr. Worse, it bumps each ntpd/*.c file's modification time, which
sometimes triggers a strange dependency path causing the makefile to attempt
to run the ntpd keyword-gen app, which fails, because it's been
cross-compiled.
Signed-off-by: Danomi Manchego <danomimanchego123@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
If sane is installed system-wide on the build machine, then the
sane-config binary found is the one of the system, which returns
incorrect library paths for cross-compilation.
To fix this, this commit adds a patch to wine to make it support a
SANE_CONFIG environment variable, and then adjusts wine.mk to
explicitly pass the path to $(STAGING_DIR)/usr/bin/sane-config.
Fixes:
http://autobuild.buildroot.org/results/8bd/8bdc1eed55075313403aa8a6c9af6a427bce198e/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
The package comes with usable .service files for smbd, nmbd and
winbind, but does not install them.
[Thomas: use relative paths for the symbolic links.]
Signed-off-by: Alex Suykov <alex.suykov@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
squid comes with a .service file, but does not install it.
[Thomas: use relative path for symlink instead of absolute path.]
Signed-off-by: Alex Suykov <alex.suykov@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Usable .service file is provided in the package.
Signed-off-by: Alex Suykov <alex.suykov@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This change adds 2 patches (backported from upstream) to vlc fixing the
build against opencv2 APIs.
This allows to select the minimal set of opencv modules when opencv
support is enabled.
Signed-off-by: Samuel Martin <s.martin49@gmail.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Starting with the 2.4.10 release, OpenCV supports both Gstreamer-0.10
and Gstreamer-1, but only one can be enabled at the same time (OpenCV
chooses Gstreamer-1 over Gstreamer-0.10 when both are available).
Signed-off-by: Samuel Martin <s.martin49@gmail.com>
Cc: "Yann E. MORIN" <yann.morin.1998@free.fr>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
OpenCV 3.0 support both gstreamer-0.10 and gstreamer-1.x, but only one
is used at the time.
This patch turns the gstreamer support into a choice, in order to prepare
adding the support for gstreamer-1 in a following patch.
Signed-off-by: Samuel Martin <s.martin49@gmail.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
In file included from libavcodec/cabac_functions.h:46:0,
from libavcodec/h264_cabac.c:37:
libavcodec/h264_cabac.c: In function 'ff_h264_decode_mb_cabac':
libavcodec/x86/cabac.h:192:5: error: 'asm' operand has impossible constraints
__asm__ volatile(
To reproduce the bug use this defconfig:
BR2_GCC_VERSION_5_1_X=y
BR2_PACKAGE_FFMPEG=y
BR2_PACKAGE_FFMPEG_GPL=y
BR2_PACKAGE_FFMPEG_NONFREE=y
BR2_PACKAGE_FFMPEG_FFPLAY=y
BR2_PACKAGE_FFMPEG_FFSERVER=y
BR2_PACKAGE_FFMPEG_FFPROBE=y
BR2_PACKAGE_FFMPEG_AVRESAMPLE=y
BR2_PACKAGE_FFMPEG_POSTPROC=y
[Thomas: add a comment in the code.]
Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
In commit f21b2558a0 ("busybox: added
linux-pam support"), we added optional support for PAM in
Busybox. However, this support requires the toolchain to have thread
support, which causes build failures with non-thread capable
toolchains.
This commit therefore enables Busybox PAM support only if the
linux-pam package is available *and the toolchain has thread support.
Fixes:
http://autobuild.buildroot.org/results/1a3/1a380aaca9303b67cc59165d56cf12f35966fe26/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
In case FOO_OVERRIDE_SRCDIR has trailing spaces, like so:
FOO_OVERRIDE_SRCDIR = /path/to/sources\x20
we would end up with a rsync command like so:
rsync -au /path/to/sources / /path/to/build/foo
which would effectively rsync the whole vfs, eventually filling the
whole disk... :-(
So, just qstrip the variable before use.
Reported-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
In the *-install-target phase the manifest file is being updated, if multiply packages try to update it they fail.
To avoid multiple access to the manifest file use flock to sync
multiple luarocks packages.
e.g. installing three luarocks packages:
make lua-cjson-build lua-coat-build lua-coatpersistent-build
make lua-cjson lua-coat lua-coatpersistent -j
Fix error:
Updating manifest for /home/tetsuya/buildroot/br2/output/target/usr/lib/luarocks/rocks
No existing manifest. Attempting to rebuild...
Error: rock_manifest file not found for lua-coat 0.9.1-1 - not a LuaRocks 2 tree?
[Thomas: get rid of LUAROCKS_RUN, and use LUAROCKS_RUN_ENV +
LUAROCKS_RUN_CMD everywhere.]
Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The boost and jack2 packages fail to build when PARALLEL_JOBS is empty
so instead of using an empty PARALLEL_JOBS don't use it in the MAKE
variable when top-level parallel make is being used.
To simplify the use of top-level parallel make, check the MAKEFLAGS
variable to know automatically if the -j option is being used, also use
the "=" operator instead of the ":=" operator because the MAKEFLAGS
variable can be checked only in a "recursively expanded variable".
The "override" keyword must be used in order to change the automatic
variable "MAKE".
When the top-parallel make is being used the sub-make are called without
specifying the "-j" option in order to let GNU make share the job slots
specified in the top make. This is done because GNU make is able
to share the job slots available between each instance of make so if you
want to increase the number of jobs you just need to increase the <jobs>
value in the top make -j<jobs> command.
If we specify the -j<jobs> option in each instance of make, it is less
efficient, e.g. in a processor with 8 cores we specify -j9 in each instance:
the number of processes goes up to 81 because each sub-make can execute
9 processes. The excessive number of processes is not a good thing
because in my tests even -j16 is slower than -j9.
Instead if we don't specify the -j<jobs> option in the sub-make, the top
make share the job slots automatically between each instance, so the
number of process in this examples goes up to 9 that is faster than
using up to 81 processes.
e.g. when the -j3 option is specified only in the top make:
possible state n. 1:
process 1 - <packagea>-build
process 2 - <packagea>-build
process 3 - <packagea>-build
possible state n. 2:
process 1 - <packagea>-extract
process 2 - <packageb>-configure
process 3 - <packagec>-build
possible state n. 3:
process 1 - <packagea>-build make -j1
process 2 - <packageb>-build make -j1
process 3 - <packagec>-build make -j1
Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Fixes one remotely-triggerable issue that was found by the Codenomicon
Defensics tool, one potential remote crash and countermeasures against
the "Lucky 13 strikes back" cache-based attack.
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Other nodejs-related packages will need to call npm with the same set of
arguments as is currently used by the nodejs package itself.
To avoid duplicating this code, set the NPM variable so those packages can
re-use it.
Signed-off-by: Martin Bark <martin@barkynet.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Many packages use node-pre-gyp as a way of deploying precompiled binary
dependencies with fall back to compilation for other targets. Currently
installing node modules that use node-pre-gyp can fail to use the correct
binary for the target. This patch fixes this issue by correctly
configuring node-pre-gyp.
Firstly, node-gyp uses the option --arch to determine its target
architecture (which is already set correctly), however, node-pre-gyp uses
--target-arch. Without this set node.js packages that uses node-pre-gyp
will pick the wrong target architecture.
Secondly, the use of precompiled binary packages is not desirable due to
potential security and licensing issues. To solve this we use the
--build-from-source option to force node-pre-gyp to always build the C++
code.
This patch passes npm_config_target_arch and npm_config_build_from_source
to npm which causes --target-arch and --build-from-source to be passed to
node-pre-gyp.
I have tested this using the node.js package serialport which now
successfully builds and runs.
Signed-off-by: Martin Bark <martin@barkynet.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This patch symlinks all executables in /usr/lib/node_modules/.bin
to /usr/bin so that node.js modules installed using
BR2_PACKAGE_NODEJS_MODULES_ADDITIONAL are accessible from the command line.
Signed-off-by: Martin Bark <martin@barkynet.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The version of the V8 JavaScript engine used by node.js v0.12.5 requires
at least an ARMv6 architecture with VFPv2. For this reason v0.10.39
remains the default for ARMv5 targets, all other targets now default to
v0.12.5.
Signed-off-by: Martin Bark <martin@barkynet.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Lots of fixes have been made to ltrace, including the ones for fixing a
build failure for MIPS architecture. Updating to current master will
allow us to re-enable this package for MIPS and also remove some
upstreamed patches.
At the same time we add a patch made by Jérôme Pouiller to fix a bug
introduced by 5ba9e10 ("Split type definitions from the bundled configs
into their own files"). Two new configuration files are not installed.
Therefore, ltrace fail with messages like :
/usr/share/ltrace/libm.so.conf:333: error: unknown type around 'ldouble
erfl(ldouble);
That patch has been sent upstream.
Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Fixes
checking for /home/fli4l/buildroot/output/host/usr/bin/i586-buildroot-linux-uclibc-gcc option to accept ISO C99... unsupported
configure: error: Building libdrm requires C99 enabled compiler
using this defconfig
BR2_KERNEL_HEADERS_4_0=y
BR2_BINUTILS_VERSION_2_25=y
BR2_TOOLCHAIN_BUILDROOT_CXX=y
BR2_PACKAGE_LIBDRM=y
Patch inspired by
http://git.buildroot.net/buildroot/commit/?id=5cf5b390385fb6325485e37dc9d38e1e3ac1f091
Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
cryptsetup does not actually depend on e2fsprogs, but on libuuid that is a
dependency of e2fsprogs. Remove the e2fsprogs dependency, and add a direct
dependency on util-linux (libuuid provider).
Cc: Martin Hicks <mort@bork.org>
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
host-libtool can fail to build if the host is missing makeinfo, so
disable it.
Signed-off-by: Sam Bobroff <sam.bobroff@au1.ibm.com>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
In this case -DMESA_EGL_NO_X11_HEADERS is needed during compilation.
Fixes this build error
make[1]: Entering directory `/home/fli4l/br8_kodi/output/build/kodi-14.2-Helix/xbmc/cores/dvdplayer'
CPP xbmc/cores/dvdplayer/DVDPlayerVideo.o
In file included from /home/fli4l/br8_kodi/output/host/usr/i586-buildroot-linux-uclibc/sysroot/usr/include/EGL/egl.h:36:0,
from /home/fli4l/br8_kodi/output/build/kodi-14.2-Helix/xbmc/windowing/egl/WinSystemEGL.h:28,
from /home/fli4l/br8_kodi/output/build/kodi-14.2-Helix/xbmc/windowing/WindowingFactory.h:39,
from DVDPlayerVideo.cpp:23:
/home/fli4l/br8_kodi/output/host/usr/i586-buildroot-linux-uclibc/sysroot/usr/include/EGL/eglplatform.h:118:22: fatal error: X11/Xlib.h: No such file or directory
#include <X11/Xlib.h>
^
using this defconfig
BR2_TOOLCHAIN_BUILDROOT_WCHAR=y
BR2_TOOLCHAIN_BUILDROOT_CXX=y
BR2_ROOTFS_DEVICE_CREATION_DYNAMIC_EUDEV=y
BR2_PACKAGE_KODI=y
BR2_PACKAGE_MESA3D=y
BR2_PACKAGE_MESA3D_GALLIUM_DRIVER_SWRAST=y
BR2_PACKAGE_MESA3D_OPENGL_EGL=y
BR2_PACKAGE_MESA3D_OPENGL_ES=y
Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The option is controlled later on inside this if-clause
ifeq ($(BR2_PACKAGE_KODI_OPTICALDRIVE),y)
Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
With raspberrypi_defconfig:
libv4l2rds.c:256:2: error: 'for' loop initial declarations are only allowed in C99 or C11 mode
for (int i = 0; i < tuning->station_cnt; i++) {
^
libv4l2rds.c:256:2: note: use option -std=c99, -std=gnu99, -std=c11 or -std=gnu11 to compile your code
>From build/libv4l-1.6.2/config.log:
configure:4709: checking for .../host/usr/bin/arm-buildroot-linux-uclibcgnueabi-gcc option to accept ISO C99
configure:4858: .../host/usr/bin/arm-buildroot-linux-uclibcgnueabi-gcc -std=gnu99 -c -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -Os -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 conftest.c >&5
conftest.c:54:9: error: unknown type name 'wchar_t'
const wchar_t *name;
The C99 detection problem seems more related to missing wchar_t type support than the compiler option?
Adding LIBV4L_CONF_ENV = ac_cv_prog_cc_c99='-std=c99' gives a lot of compile errors like:
libv4lconvert.c: In function 'dev_ioctl':
processing/../libv4lsyscall-priv.h:85:10: error: 'SYS_ioctl' undeclared (first use in this function)
syscall(SYS_ioctl, (int)(fd), (unsigned long)(cmd), (void *)(arg))
^
libv4lconvert.c:43:9: note: in expansion of macro 'SYS_IOCTL'
return SYS_IOCTL(fd, cmd, arg);
^
processing/../libv4lsyscall-priv.h:85:10: note: each undeclared identifier is reported only once for each function it appears in
syscall(SYS_ioctl, (int)(fd), (unsigned long)(cmd), (void *)(arg))
^
libv4lconvert.c:43:9: note: in expansion of macro 'SYS_IOCTL'
return SYS_IOCTL(fd, cmd, arg);
^
Adding LIBV4L_CONF_ENV = ac_cv_prog_cc_c99='-std=gnu99' solves all configure/compile errors.
Signed-off-by: Peter Seiderer <ps.report@gmx.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Binutils 2.24 produces broken code when compiling the kernel for
ppc64le, so prevent this combination. See:
https://sourceware.org/ml/binutils/2013-12/msg00200.html
The problem manifests early in the boot process with "Kernel access of
bad area, sig: 11" in arch_match_cpu_phys_id().
The fix has been merged upstream as commit
57fa7b8c7e59e35bced580f9bcb9668af43fdbce, which is available since
Binutils 2.25.
Signed-off-by: Sam Bobroff <sam.bobroff@au1.ibm.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This combination causes a compilation failure of the host-gcc-final
recipe like this one:
/br/output/host/usr/mips-buildroot-linux-gnu/bin/ld:
.libs/gload.o: relocation R_MIPS_HI16 against `__gnu_local_gp' can not
be used when making a shared object; recompile with -fPIC
The problem is the file 'libatomic/gload.c' is compiled without -fPIC
when using binutils-2.25. All gcc (with libatomic) versions below 4.9.3
are affected by this issue.
Here is a summary of affected/unaffected versions in Buildroot:
4.7.x: unaffected (doesn't have libatomic)
4.8.x: affected
4.9.x: unaffected (we have 4.9.3 which is fixed)
5.1.x: unaffected
The fix can be found here:
57f5c0954f
However, given the following reasons...
- Upstream gcc 4.8 branch is closed.
- The fix is very hard to backport from 4.9 to 4.8.
- This stuff is insanely sensitive and not working at all could be
better than looking like it works but not quite.
...I think the best choice is to disable that combination in Buildroot.
Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
When working with GCC initial at override source dir mode the
HOST_GCC_INITIAL_POST_PATCH_HOOKS is not called and compilation failes.
The solution is to use HOST_GCC_INITIAL_POST_RSYNC_HOOKS since this hook
is being called at override source dir mode.
Signed-off-by: Tal Zilcer <talz@ezchip.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This is necessary for some packages depending on valgrind, such as
libdrm which will fail with an error like this one:
checking for VALGRIND... no
checking whether to enable Valgrind support... configure: error:
Valgrind support required but not present
package/pkg-generic.mk:146: recipe for target
'/br/output/build/libdrm-2.4.62/.stamp_configured' failed
Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Up to date tor (at least starting from 0.2.6) requires compiler with C99
plus some extensions support.
If default GCC's C standard < c99 (that's the case at least for ARC)
you'll see this on attempt to build tor:
----------------------->8--------------------
src/common/address.c: In function ''tor_addr_parse_PTR_name':
src/common/address.c:502:5: error: 'for' loop initial declarations are only allowed in C99 mode
for (int i = 0; i < 16; ++i) {
^
src/common/address.c:502:5: note: use option -std=c99 or -std=gnu99 to compile your code
----------------------->8--------------------
Once you follow compiler advice and enable c99 support with "-std=c99"
you'll pass that failure but will see tons of other errors, see
https://www.mail-archive.com/tor-dev@lists.torproject.org/msg06273.html
And only g99 resolves all problems at once.
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
ModemManager get started by NetworkManager, in case of systemd init
system. In case of other systems it needs to be started by init script.
Debian [1] solved it by detection in code. For Buildroot it's IMHO
enough to install init script for systemV-like init systems.
[1] https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=770871
[Thomas:
- slightly simplify the script by removing the MODEMMANAGER_BIN
variable which was used at only one place, and use directly $?
instead of an intermediate $ret variable.
- split the too long line added in the .mk file.]
Signed-off-by: Petr Vorel <petr.vorel@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Upstream hash is in sha1 format, like before. It is wrongly announced
as md5 in the release mail, so I added a note to the hash file.
Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
vcdbg is a tool to help debug the communication with the GPU.
It comes as a binary-only, and in two flavours: one for the hard
floating point ABI, one for the software floating point ABI.
Unfortunately, we have no source code for that tool, only a binary that
was dynamically linked with glibc and libraries from rpi-userland.
So, just install that executable, and let's hope there is no symbol
issue at runtime.
Note: vcdbg needs glibc, threads and !static. Since glibc already
implies threads and !static, we only need to depend on glibc.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc; Floris Bos <bos@je-eigen-domein.nl>
Cc: Pascal de Bruijn <pmjdebruijn@pcode.nl>
Cc: Baruch Siach <baruch@tkos.co.il>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Even though ARC gcc understands "-pie" option and attempts to generate
PIE binaries as of today PIE is not really supported for user-space
applications.
So we disable PIE detection if building for ARC.
That first fixes http://autobuild.buildroot.net/results/988/9888a7a30538c9851f4910c16674d8dbb8edeb8f/
and also prevents execution of non-supported PIE binary in runtime.
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Matthew Weber <matthew.weber@rockwellcollins.com>
Reviewed-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
[Thomas: simplify condition on
$(BR2_PACKAGE_GOOGLE_MATERIAL_DESIGN_ICONS_TYPE_PNG) and
$(BR2_PACKAGE_GOOGLE_MATERIAL_DESIGN_ICONS_TYPE_SVG).]
Signed-off-by: James Knight <james.d.knight@live.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
In buildroot we maintain a couple off-the-tree patches.
In particular for binutils and gcc packages.
Because we have many versions of mentioned packages patches for each
particular version reside in a folder which name matches full version
name of the package.
For example we used to have patches for binutils distributed as part of
arc-2014.12 tools in folder ""package/binutils/arc-2014.12". Now with
bump of ARC tools version we need to rename folder with patches to
"arc-2015.06-rc1".
The same applies to gcc.
Should fix http://autobuild.buildroot.net/results/2b2/2b27a4a64c0b225ae479ecfccf7a97f5ea95598c/
As discussed before it's not possible to reproduce reported problem on recent
Fedora 21/22 distros (at least we know about them) but I may confirm that
patches were applied fine and everything was built well. Hopefully reported
build failure goes away now.
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Romain Naour <romain.naour@openwide.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This patch bumps version to 3.1.2 and adds support for the recently
introduced python3 support for py-smbus.
Signed-off-by: Angelo Compagnucci <angelo.compagnucci@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
+ lower kernel headers dependency
According to upstream [1], there is no known minimal kernel-version, nor
minimal required feature-set.
Experimentally tested, that 1.0.2 is works with 3.2 kernel headers, even
some features will be missing [2].
[1] https://mail.gnome.org/archives/networkmanager-list/2015-April/msg00039.html
[2] https://mail.gnome.org/archives/networkmanager-list/2015-April/msg00041.html
[Thomas: add comment in Config.in to indicate that it may work with
earlier kernel versions.]
Signed-off-by: Petr Vorel <petr.vorel@gmail.com>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Only (e)glibc provides libnsl, uclibc provides only a stub, and musl
doesn't implement it at all.
Fixes compilation using this defconfig
BR2_arm=y
BR2_cortex_a7=y
BR2_STATIC_LIBS=y
BR2_TOOLCHAIN_EXTERNAL=y
BR2_PACKAGE_OPENSSL=y
BR2_PACKAGE_EXIM=y
Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The commit 720893b625 "openssl: disable
apps for NOMMU" prevented the openssl binary from being built without
MMU in order to fix a build failure without fork(). However, openssl is
designed to support the lack of fork() with -DHAVE_FORK=0, so allow the
openssl binary to be enabled without MMU thanks to this option.
Signed-off-by: Benoît Thébaudeau <benoit@wsystem.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The original package supported python on target, now we just use
it as part of the host tools. The license was also mis-assigned.
[Thomas: add removed option to Config.in.legacy.]
Signed-off-by: Matthew Weber <matthew.weber@rockwellcollins.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Fix a typo in the monkey package - should be MONKEY_CONF_OPTS
instead of MONKEY_CONF_OPT.
Signed-off-by: Bartosz Golaszewski <bgolaszewski@baylibre.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Include commit ec6cc07f which fixes an issue causing build failures
on several architectures.
Signed-off-by: Bartosz Golaszewski <bgolaszewski@baylibre.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Following commit 36555b4c8d ("ARC:
switch to RC1 of upcoming arc-2015.06 tools"), the binutils package
only contains the hash information for the 2015.06-rc1 ARC version of
binutils (which is in fact no hash). However, when an external
toolchain is used, and binutils is built for the target, it's still
the old 2014.11 binutils version that gets used in binutils.mk,
causing a build failure because there is no hash available for this
version.
This commit fixes that by using 2015.06-rc1 as the default binutils
version on ARC, which is used when no host-binutils has been built.
Fixes:
http://autobuild.buildroot.org/results/8f7/8f772e6fccb4f918120a7bb814da2224432d1c09/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
For some reason, the hash file added in commit
ee6c9f5a1b by Gustavo turns out to be
wrong, so this commit replaces it with the proper hash, also added
after checking the GPG signature.
While at it, we switch to using a .bz2 tarball.
Fixes:
http://autobuild.buildroot.org/results/c44/c441f35119191f47dd5fae96fd76827024e50329/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Monkey is a small, fast and lightweight open source Web Server for
GNU/Linux. It has been designed with focus in embedded devices,
therefore its scalable by nature having a low memory and CPU
consumption and an excellent performance.
[Thomas:
- Add missing dependency on !BR2_STATIC_LIBS (the source code uses
dlopen) and BR2_USE_MMU (the source code uses fork)
- Slightly adjust/reword the description of the
BR2_PACKAGE_MONKEY_SHARED option.
- Remove all the complicated installation logic for the target, and
just use "make install" instead.
- Pass --no-backtrace when uClibc is used, otherwise the build fails
because <execinfo.h> is not available in uClibc.
- Pass $(TARGET_CONFIGURE_OPTS) in the environment of the configure
script., otherwise monkey gets built for the host and not for the
target.
- Add a post install target hook to remove a broken symlink
libmonkey.so installed by Monkey's Makefile when the shared
library is not enabled.
- Use TARGET_MAKE_ENV when calling make, just because we should.
- Pass --malloc-libc so that the libc malloc() is used instead of
the builtin jemalloc allocator, which requires more work to
cross-compile properly.
- Add missing empty line after the .mk header and before the first
variable definition.]
Signed-off-by: Julien Corjon <corjon.j@ecagroup.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Misc fixes and improvements all over the place...
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Floris Bos <bos@je-eigen-domein.nl>
Cc: Pascal de Bruijn <pmjdebruijn@pcode.nl>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Currently, Buildroot does not support building the overlays that are
bundled in the Linux kernel, so all we can do is install the ones
pre-built in rpi-firmware.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Floris Bos <bos@je-eigen-domein.nl>
Cc: Pascal de Bruijn <pmjdebruijn@pcode.nl>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
