[Thomas:
- replicate the ffmpeg "depends on" related to the broken NIOS2
toolchains to the BR2_PACKAGE_SQUEEZELITE_FFMPEG option.
- use the github helper function.
- add hash file.]
Signed-off-by: Hiroshi Kawashima <kei-k@ca2.so-net.ne.jp> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The main .service file is installed by audit.
We are using tmpfiles mechanisms to create the log directory, which
would not exists otherwise since /var/log points to /tmp
[Thomas: add explicit --disable-systemd when systemd is not used.]
Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
[Thomas:
- add patch to fix autoreconf issue, and use LIQUID_DSP_AUTORECONF =
YES instead of an horrible hack calling aclocal/autoconf manually.
- use the github macro instead of hand-coding <pkg>_SITE and
<pkg>_SITE_METHOD. This allows to remove <pkg>_SITE_METHOD
entirely.
- use a full hash as the <pkg>_VERSION
- remove trailing whitespace everywhere.
- use one single assignment of LIQUID_DSP_CONF_OPTS
- fix the comment about the eglibc/musl dependency (it was only
mentioning eglibc, and the condition was inverted)
- add the musl/glibc dependency on the package option itself
- make the package depend on dynamic library support, since the
makefile unconditionally builds a shared library.
- add hash file.]
Signed-off-by: Guillaume William Bres <guillaume.bressaix@gmail.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
wpa_supplicant: Add an explicit option to enable nl80211.
Currently, nl80211 support is conditionnal with libnl being enabled,
using implicit dependencies. This causes problems since it is not
obvious and wpa_supplicant without nl80211 isn't what most user expects.
If nl80211 isn't enabled, then buildroot only enables the wext driver,
which will only work if some deprecated kernel feature isn't left
disabled, or if using a outdated out-of-tree linux driver which doesn't
use the cfg80211 infrastructure.
This makes nl80211 support an explicit option, which
"select BR2_PACKAGE_LIBNL" accordingly. To handle upgrades nicely, it
would have been nice to have "default y if BR2_PACKAGE_LIBNL", but
Kconfig treats this as a circular dependency. So instead, this enables
the option by default, which is less worse than not enabling nl80211
when it was previously implicitely enabled.
[Thomas:
- rewrap Config.in help text
- add comment about thread dependency.]
Signed-off-by: Nicolas Cavallari <nicolas.cavallari@green-communications.fr> Reviewed-by: Gary Bisson <gary.bisson@boundarydevices.com> Tested-by: Gary Bisson <gary.bisson@boundarydevices.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
In order to remove '/home/ftp' and the ftp user from the 'skeleton'
package, we need to add the creation of the ftp users to the package so
it still work out of the box (with an anonymous user).
[Thomas:
- remove the VSFTPD_FIX_HOME_PERMISSIONS variable, and simply put the
additional command in VSFTPD_INSTALL_TARGET_CMDS.
- slightly tweak the comment.]
Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com> Reviewed-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
In order to remove '/home/ftp' and the ftp user from the 'skeleton'
package, we need to add the creation of the ftp users to the package so
it still work out of the box (with an anonymous user).
Signed-off-by Maxime Hadjinlian <maxime.hadjinlian@gmail.com> Reviewed-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Peter Seiderer [Tue, 6 Oct 2015 19:35:30 +0000 (21:35 +0200)]
libdrm: update driver options
- add AMD GPU driver (available since 2.4.63)
- freedreno is no longer experimental (since 2.4.59)
- add NVIDIA Tegra driver (available since 2.4.59)
- update help texts (install vs. installs, change to complete
sentences)
Signed-off-by: Peter Seiderer <ps.report@gmx.net> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Some Buildroot-generated static toolchains for ARM do not have
dlfcn.h header, which breaks builds of qt-4.8.7 because its
QtCore's plugin loader expects this header to be present (and
dynamic libraries to be supported). For certain platforms without
dynamic libraries, there is a QT_NO_DYNAMIC_LIBRARY constant
defined in plugin loader's source. This patch puts that constant
into Qt's build environment if Buildroot is configured without
shared library support.
Peter Seiderer [Wed, 7 Oct 2015 21:56:48 +0000 (23:56 +0200)]
media-ctl: remove package
Package media-ctl is out-dated, source (and developement) have been moved
to v4l-utils since June 2014. Up-to-date version is available in
the libv4l package (BR2_PACKAGE_LIBV4L/BR2_PACKAGE_LIBV4L_UTILS).
Signed-off-by: Peter Seiderer <ps.report@gmx.net> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Peter Seiderer [Thu, 8 Oct 2015 18:35:02 +0000 (20:35 +0200)]
fbterm: fix musl compile
- add missing include, fixes:
fbio.cpp:33:8: error: ‘fd_set’ does not name a type
static fd_set fds;
improxy.cpp:439:3: error: ‘fd_set’ was not declared in this scope
- add missing WAIT_ANY define, fixes:
fbterm.cpp: In member function ‘void FbTerm::processSignal(u32)’:
fbterm.cpp:212:22: error: ‘WAIT_ANY’ was not declared in this scope
s32 pid = waitpid(WAIT_ANY, 0, WNOHANG);
gnuradio: break dependency chain by using 'depends on'
This commit fixes the following recursive dependency:
package/python/Config.in:5:error: recursive dependency detected!
package/python/Config.in:5: symbol BR2_PACKAGE_PYTHON is selected by BR2_PACKAGE_GNURADIO_PYTHON
package/gnuradio/Config.in:46: symbol BR2_PACKAGE_GNURADIO_PYTHON is selected by BR2_PACKAGE_GNURADIO_UTILS
package/gnuradio/Config.in:55: symbol BR2_PACKAGE_GNURADIO_UTILS depends on BR2_PACKAGE_PYTHON_NUMPY_ARCH_SUPPORTS
package/python-numpy/Config.in:1: symbol BR2_PACKAGE_PYTHON_NUMPY_ARCH_SUPPORTS depends on BR2_PACKAGE_PYTHON
by replacing the "select BR2_PACKAGE_PYTHON" by a "depends on
BR2_PACKAGE_PYTHON", and ditto for the BR2_PACKAGE_GNURADIO_PYTHON
selection.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
gnuradio: add missing dependency to python-numpy for python support
Some gnuradio python blocks need python-numpy to be present. Without
python-numpy, gnuradio-python based apps fails, on runtime, with
traceback like:
Traceback (most recent call last):
File "./top_block.py", line 18, in <module>
from gnuradio import analog
File "/usr/lib/python2.7/site-packages/gnuradio/analog/__init__.py", line 35, in <module>
from am_demod import *
File "/usr/lib/python2.7/site-packages/gnuradio/analog/am_demod.py", line 22, in <module>
from gnuradio import gr
File "/usr/lib/python2.7/site-packages/gnuradio/gr/__init__.py", line 44, in <module>
from top_block import *
File "/usr/lib/python2.7/site-packages/gnuradio/gr/top_block.py", line 30, in <module>
from hier_block2 import hier_block2
File "/usr/lib/python2.7/site-packages/gnuradio/gr/hier_block2.py", line 26, in <module>
import pmt
File "/usr/lib/python2.7/site-packages/pmt/__init__.py", line 58, in <module>
from pmt_to_python import pmt_to_python as to_python
File "/usr/lib/python2.7/site-packages/pmt/pmt_to_python.py", line 22, in <module>
import numpy
ImportError: No module named numpy
Signed-off-by: Gwenhael Goavec-Merou <gwenhael.goavec-merou@trabucayre.com> Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be> Reviewed-by: Samuel Martin <s.martin49@gmail.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
opencv3: use BR2_PACKAGE_PYTHON_NUMPY_ARCH_SUPPORTS
Opencv3 select python-numpy. To avoid duplicating architecture
dependencies, an hidden variable has been added to python-numpy
package. This patch propagate this modification to the opencv3
package.
Signed-off-by: Gwenhael Goavec-Merou <gwenhael.goavec-merou@trabucayre.com> Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be> Reviewed-by: Samuel Martin <s.martin49@gmail.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Some packages selects python-numpy. This package has a some CPU specific
code. To simplify and avoid duplicate dependencies this patch introduce an
hidden ARCH_SUPPORTS variable.
[Thomas: fix the definition of the new option to make sure its value
is actually 'y' when a supported architecture is used.]
Signed-off-by: Gwenhael Goavec-Merou <gwenhael.goavec-merou@trabucayre.com> Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be> Reviewed-by: Samuel Martin <s.martin49@gmail.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
- Bump to version 2.0
- Switch the download site to GitHub
- Apply an upstream patch to fix a missing define
- Add a hash file
- Rewrap the help text to 72 characters length
- Change the official website
[Thomas: fix s/MAKE_ENV/TARGET_MAKE_ENV/.]
Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
CVE-2015-5289: json or jsonb input values constructed from arbitrary
user input can crash the PostgreSQL server and cause a denial of
service.
CVE-2015-5288: The crypt() function included with the optional pgCrypto
extension could be exploited to read a few additional bytes of memory.
No working exploit for this issue has been developed.
webkitgtk24: use the correct Config.in option for NEON
In order to enable JIT support on ARM, webkitgtk24 currently looks at
BR2_ARM_ENABLE_NEON, which is not correct: BR2_ARM_CPU_HAS_NEON should
be used instead.
The BR2_ARM_ENABLE_NEON is only visible for cores that select
BR2_ARM_CPU_MAYBE_HAS_NEON, in order to allow to specify if his
particular SoC has chosen to integrate NEON or not. And if so,
BR2_ARM_CPU_HAS_NEON gets selected.
BR2_ARM_CPU_HAS_NEON is in fact selected in two different ways:
- Either directly by the CPU core selection, if NEON is mandatory in
this CPU core.
- Or by BR2_ARM_ENABLE_NEON, for CPU cores where NEON support is
optional.
So really, BR2_ARM_CPU_HAS_NEON is what should be used by packages.
Cc: Gustavo Zacarias <gustavo@zacarias.com.ar> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> Acked-by: Gustavo Zacarias <gustavo@zacarias.com.ar> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Most code is GPLv2+ or compatible except for one file that is
GPLv2-only, hence as a whole is GPLv2 (see Copyright).
Also kill some whitespace and make the only hard dependency
(host-pkgconf) a hard dependency, not an addition, since it's not
conditional.
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
- Bump version to 1.0.25
- Update hash file. The hash value provided by the alioth.debian.org URL
is wrong, so use the one provided by www.sane-project.org. I have
reported this upstream.
- Remove unnecessary patches:
0002-remove-unsafe-headers-path.patch
0003-remove-cups-and-systemd-support.patch
Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Sergio Prado [Sat, 3 Oct 2015 15:40:02 +0000 (12:40 -0300)]
wandboard_defconfig: bump u-boot to 2015.07 and kernel to 3.14.28_1.0.0
Bump U-boot version to 2015.07 from the mainline repository. Since there
is SPL support for the wandboard in the U-boot 2015.07, it is now
possible to boot the same U-boot image for all wandboard variants
(solo/dual/quad).
Bump kernel version to 3.14.28_1.0.0 from the wandboard git repository.
Tested on Wandboard Solo, Wandboard Dual and Wandboard Quad.
Signed-off-by: Sergio Prado <sergio.prado@e-labworks.com> Tested-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
glib-networking: specify GIO_MODULE_DIR for target
Since the upgrade to pkgconf 0.9.12 all directories and variables with
directories are sysroot-prefixed, whereas with the old patch only some
variables were (includedir, mapdir, sdkdir, libdir).
libglib2 uses a giomoduledir=${libdir}/gio/modules variable in
gio-2.0.pc, which when expanded via sysroot by pkgconf at configure time
points to the staging directory (via libdir expansion).
When target install gets called the gio module gets installed to staging
rather than target, with the obvious consequences this has.
Work-around this pkgconfig abuse by overriding GIO_MODULE_DIR at target
install time.
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
python-numpy: uClibc patch to avoid ldexpl and frexpl
It turns out that uClibc only enables the long double math functions
for some architectures (god know why, so what if long double is the
same as double, these functions should still be defined). Since
python-numpy links with those functions, the module will fail to load
on ARM, MIPS and SH.
However, python-numpy actually checks for each function if it is really
available. Only, it overrides that check for ldexpl and frexpl when
long double is the same as double (i.e. in exactly the case the uClibc
doesn't have these functions).
So add another exception for this.
Upstream-status: not applicable (code has changed too much)
- Help the build system to find the zlib headers, otherwise the
configure script will fail like this:
checking if zlib is wanted... yes
checking for inflateEnd in -lz... no
checking zlib.h usability... no
checking zlib.h presence... no
checking for zlib.h... no
configure: error: either specify a valid zlib installation with
--with-zlib=DIR or disable zlib usage with --without-zlib
- Disable instrumentation support by default since it avoids runtime
failures on rtorrent for some architectures like MIPS, PowerPC or ARM:
https://github.com/rakshasa/rtorrent/issues/156
- Tweak the 0001-cross_compile.patch for this new version.
Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com> Signed-off-by: Peter Korsgaard <peter@korsgaard.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>
u-boot: add an option to indicate that DTC is needed
Some U-Boot configurations require the Device Tree compiler to be
available, so we need to depend on host-dtc (example configuration:
zynq_zed).
However, we don't want to build it unconditionally, since the vast
majority of U-Boot configurations don't need it (and host-dtc itself
has a bunch of dependencies).
So, we simply add a Config.in option that allows users to indicate
whether their U-Boot needs DTC or not, and depend on host-dtc if this
option is enabled.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Peter Seiderer [Sat, 3 Oct 2015 19:44:12 +0000 (20:44 +0100)]
dtc: add host build
Having a host variant of dtc is needed for example for U-Boot, which
uses the Device Tree for a number of platforms.
In addition, now that we have a proper host-dtc package, it is no
longer needed for the linux package to install the host dtc compiler:
users interested in having the host dtc compiler can simply enable
this package.
A Config.in.host option is added to build host-dtc, because the
initial reason why the host DTC built by the kernel was installed in
$(HOST_DIR)/usr/bin (commit 707d44d0a28906ebda49584dd5f55985406f0bde
from Thomas DS) was:
Having dtc as a host tool can be useful for users that have a
custom boot scenario where the device tree is not embedded in the
kernel.
[Thomas:
- rework the commit log.]
[Peter: keep the (renamed) linux version] Signed-off-by: Peter Seiderer <ps.report@gmx.net> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
webkitgtk24: fix compilation failure due to unicode quotes
As reported in this upstream bug...
https://bugs.webkit.org/show_bug.cgi?id=128971
...the pretty unicode double quotes break Python stdin. Backporting the
patch that was attached to that bug report and tweak it minimally to
make it apply on Buildroot.
The Barebox, AT91Bootstrap and Linux patches for the Calao QIL A9260
board are all located in the same directory, board/calao/qil-a9260/,
with only a prefix to indicate to which component they apply.
Unfortunately, since commit 0eba4759fab9d093afacac77a0ac31f1f0b3cceb
("packages: apply custom patches using *.patch instead of
<pkg>-*.patch"), the logic to apply custom patches in Barebox and
AT91Bootstrap has been changed from applying <package>-*.patch to
applying *.patch (in order to be consistent with what we do for all
packages in Buildroot).
However, this had the effect of trying to apply both the AT91Bootstrap
and Barebox patches on both packages, which obviously cannot work.
This commit fixes that by moving patches around, and switching to use
BR2_GLOBAL_PATCH_DIR for all of Barebox, AT91Bootstrap and Linux
patches.
In addition, the defconfig was not specifying explicitly which Barebox
version to use, so right now it's trying to use Barebox 2015.09, on
which the board-specific patch does not apply. So we've forced to use
Barebox 2012.08, which was the lastest release available at the time
the defconfig was initially contributed.
Cc: Gregory Hermant <gregory.hermant@calao-systems.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
configs: fix build of calao_usb_a9g20_lpw_defconfig
Both the Barebox and AT91Bootstrap patches for the Calao USB A9G20
board are located in the same directory, board/calao/usb-a9g20-lpw/,
with only a prefix making a difference between whether they apply to
Barebox or AT91Bootstrap.
Unfortunately, since commit 0eba4759fab9d093afacac77a0ac31f1f0b3cceb
("packages: apply custom patches using *.patch instead of
<pkg>-*.patch"), the logic to apply custom patches in Barebox and
AT91Bootstrap has been changed from applying <package>-*.patch to
applying *.patch (in order to be consistent with what we do for all
packages in Buildroot).
However, this had the effect of trying to apply both the AT91Bootstrap
and Barebox patches on both packages, which obviously cannot work.
This commit fixes that by moving patches around, and switching to use
BR2_GLOBAL_PATCH_DIR.
Cc: Gregory Hermant <gregory.hermant@calao-systems.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This defconfig tries to fetch its kernel source code and U-Boot source
code from gitorious.org, which is not longer available. Therefore,
this defconfig has been failing to build since quite some time. Since
there's no obvious other place to grab the kernel and u-boot source
code, and the defconfig was not updated since a long time, let's
simply get rid of it.
[Peter: also drop README] Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
In commit b0f025386f6861fdc21e38e3a33b65669276f2c2 a patch was added
to make sure that python-numpy does work in uClibc. This patch was
needed for the x86 architecture, because uClibc in fact does support
fenv for x86. However, we don't turn on this support in our uClibc
config. Because the Config.in depends still allowed uClibc for x86,
this lead to a built failure, which lead to the fix in commit b0f02538.
But since this fix also fixes it for all other (supported)
architectures, we can just remove the !uClibc dependency completely.
This is also propagated to the reverse dependency in opencv3.
Aurélien Chabot [Sat, 3 Oct 2015 19:29:09 +0000 (21:29 +0200)]
nfs-utils: systemd support
nfs-utils has been patched to install only the needed systemd unit files
supported.
Also tmpfiles was needed because the folder /run/nfs is a tmpfs
and nfs-utils need some files to be there to work correctly.
Signed-off-by: Aurélien Chabot <aurelien@chabot.fr>
[Maxime: Remove patches included originally in the patch as they have
have been merged upstream in v1.3.3]
Signed-off-by Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
[Thomas:
- add comment in nfs-utils_env.sh to explain why an empty script is
needed.
- remove incorrect indentation.] Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>