package/libdrm: move tests programs to the bottom

Tests programs are not that important; at least, they're probably less
important than the drivers selection.

Move it down to after the drivers selection.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/libgpg-error: bump to version 1.23

This patch is based on a patch sent by Vicente Olivert Riera and commented by
Arnout Vandecappelle [1].

- Bump version to 1.23
- Add a hook to fix cross-compilation
- Fix license and license files
- Remove patch applied upstream
- Add a BR2_PACKAGE_LIBGPG_ERROR_ARCH_SUPPORTS variable
- Propagate the dependencies using that variable:
    * package/cppcms
    * package/crda
    * package/gnupg2
        - package/gcr
            - package/midori
    * package/kodi
    * package/libaacs
    * package/libassuan
    * package/libgcrypt
    * package/libgpgme
    * package/libksba
    * package/libmicrohttpd
        - package/janus-gateway
        - package/kodi
        - package/ola
        - package/systemd
    * package/libssh
    * package/libssh2
        - package/php-ssh2
    * package/netatalk
    * package/network-manager
    * package/ntfs-3g
    * package/opkg
    * package/php-gnupg
    * package/rng-tools
    * package/strongswan
    * package/vpnc

[1] http://patchwork.ozlabs.org/patch/416427/

Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Jörg Krause <joerg.krause@embedded.rocks>
[Thomas:
 - rebase on master
 - changing systemd no longer needed, as it no longer selects
   libgcrypt.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
[Maxime:
 - rebase on master
 - bump to new version
 - propagate dependencies to missing packages]
Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Reviewed-by: Romain Naour <romain.naour@gmail.com>
[Thomas:
 - fix hash file.
 - change the way to handle the various arch so that it works properly
   for uClibc.
 - add nios2 arch support.
 - Maxime Hadjinlian learned some basic Emacs-fu to do the final fixups
   of this commit.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

alsa-lib: don't pass --with-softfloat

Passing this option disables the lfloat and ladsp pcm plugins and certain
APIs as explained by:

https://patchwork.ozlabs.org/patch/619527/

But people might want to use these even if they use software floating point,
and the size difference is really minimal, so don't pass this option.

-rwxr-xr-x 1 peko peko 774K Jul  4 18:50 target-orig/usr/lib/libasound.so.2.0.0
-rwxr-xr-x 1 peko peko 709K Jul  4 19:08 target-softfloat/usr/lib/libasound.so.2.0.0

If people don't want the extra pcm plugins then they can simply be disabled
using BR2_PACKAGE_ALSA_LIB_PCM_PLUGINS.

Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

imagemagick: explicitly specify ghostscript font directory

It appears that Imagemagick needs to be configured with
 --with-gs-font-dir=/usr/share/fonts/gs else it will not find
the ghostscript fonts and /etc/ImageMagick-6/type-ghostscript.xml
will contain the wrong font paths. You end up with messages like:
Magick: unable to read font `(null)' @ error/annotate.c/RenderFreetype/1153
They are non-fatal but could get annoying fast if you are using 'convert'
on a lot of files.

Without patch we end up with (depending on what is available on the host):

  Font Configuration:
  Apple fonts       --with-apple-font-dir=default
  Dejavu fonts      --with-dejavu-font-dir=default      none
  Ghostscript fonts --with-gs-font-dir=default  /usr/share/fonts/type1/gsfonts/
  Windows fonts     --with-windows-font-dir=default     none

And with the patch:

  Font Configuration:
  Apple fonts       --with-apple-font-dir=default
  Dejavu fonts      --with-dejavu-font-dir=default      none
  Ghostscript fonts --with-gs-font-dir=/usr/share/fonts/gs      /usr/share/fonts/gs/
  Windows fonts     --with-windows-font-dir=default     none

[Peter: reworked/extended commit message]
Signed-off-by: Robert Sohn <grepper@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

scancpan: improve message when bad host perl version

Signed-off-by: Francois Perrad <francois.perrad@gadz.org>
[Thomas:
 - add comment in scancpan about the version dependency, suggested by
   Yann E. Morin.
 - add comment in perl.mk about the need to sync any version change with
   scancpan, also suggested by Yann E. Morin.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

openblas: pass ONLY_CBLAS=1 only when Fortran is not available

Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/multicat: use double quotes in toolchain comment

To be consistent with the rest of Buildroot.

Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/gpsd: use 'menuconfig' instead of 'config'+'menu'

Since there were multiple sub-menus, their prompts are turned into
comments.

Some option were missing a prompt, so add one; switch integer options to
use the 'int' type instead of 'string'. Add separating lines around
options.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/gd: use 'menuconfig' instead of 'config'+'menu'

[Peter: drop tools comment]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

dvblast: use double quotes around the Config.in comment

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/dvblast: requires NPTL

dvblast uconditionally believes that any Linux system will have support
for clock nanosleep().

However, clock_nanosleep() is only availabe with NPTL.

Fixing dvblast is not really complicated, but not trivial either. So we
just make it depend on NPTL.

Fixes:
    http://autobuild.buildroot.org/?reason=dvblast-3.0
    http://autobuild.buildroot.org/results/dbe/dbe91dd6543b4b88fbf1e5b09d4997d2dc730747/
    http://autobuild.buildroot.org/results/32b/32bee5071f46b86de61db7f35a2cd04e270bc73c/
    [and many more...]

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/pulseaudio: fix typo

commit 7290dc1e87b23fe21c75c50b2238004273a0ec5c added
PULSE_AUDIO_CONF_ENV instead of PULSEAUDIO_CONF_ENV.

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/pulseaudio: explicitly enable/disable WebRTC support

Currently, pulseaudio optionally depends on webrtc if it is enabled.

However, it is never explcitly enabled or disabled, although
pulseaudio's ./configure has --enable/disable-webrtc-aec.

Forcibly enable/disable webrtc support using those options.

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/pulseaudio: allow to build without C++11 support

C++11 support is only required for WebRTC optional support but
pulseaudio can still be built with toolchains based on gcc 4.7 or older
witch doesn't have it.

Add a patch to make C++11 support optional and explicitely check for
gnu++11. Also stop the build only if no C++11 support has been found
and WebRTC support has been resquested.

Fixes:
http://autobuild.buildroot.net/results/d89/d897a94a8b6a52eba03d04c24536f554a7643000

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

toolchain-external: Commonize /lib/<tuple> symlinks for Linaro toolchains

As proposed initialy by Matthew Fornero [1], commonize the creation of
symlinks from {/usr}/lib to {/usr}/lib/<tuple> for Linaro toolchains.

This symlinks are only required for old Linaro toolchains.

[1] http://patchwork.ozlabs.org/patch/624577

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Cc: Matthew Fornero <mfornero@mathworks.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

minnowboard-max: Start X on boot and run a few demos

Now that we've merged nodm, the autologin display manager, we can
get rid of xinit, and startx hack and use nodm to start X on boot.
While here, let's start xterm and glmark2, just as demos.

Signed-off-by: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

configs: chromebook snow: bump to linux 4.6.3

No significant changes but 4.6.3 moved parts of the device tree
into a different file, so dts patch must be updated.

[Peter: explicitly configure kernel headers for 4.6]
Signed-off-by: Alex Suykov <alex.suykov@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

scanpypi: change hash file comments

sha256 checksum will be computed locally either by scanpypi at package
creation or by hand by package updates. Define this checksum as
'computed locally' so that one doesn't need to change this comment by
package updates. Also put comments for both md5 and sha256 in one line.

Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package: move busybox show others dependency to per-package

Having the BR2_PACKAGE_BUSYBOX_SHOW_OTHERS dependencies in
package/Config.in is not very practical: it makes this file not very
readable, and puts the dependency away from the package itself, which
can sometimes be confusing. Therefore, this commit moves the dependency
in each package Config.in file.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

netcat-openbsd: remove stray ~ character at end of Config.in

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/angularjs: drop modules selection and simplify

The angular modules are very small compared to angular:

ls -lahS target/var/www
total 228K
-rw-r--r-- 1 peko peko 142K Jul  4 16:06 angular.js
-rw-r--r-- 1 peko peko  23K Jul  4 16:06 angular-animate.js
-rw-r--r-- 1 peko peko  11K Jul  4 16:06 angular-message-format.js
-rw-r--r-- 1 peko peko 5.9K Jul  4 16:06 angular-sanitize.js
-rw-r--r-- 1 peko peko 4.3K Jul  4 16:06 angular-route.js
drwxr-xr-x 2 peko peko 4.0K Jul  4 16:06 .
drwxr-xr-x 5 peko peko 4.0K Jul  4 15:58 ..
-rw-r--r-- 1 peko peko 3.7K Jul  4 16:06 angular-aria.js
-rw-r--r-- 1 peko peko 3.6K Jul  4 16:06 angular-resource.js
-rw-r--r-- 1 peko peko 3.6K Jul  4 16:06 angular-touch.js
-rw-r--r-- 1 peko peko 2.6K Jul  4 16:06 angular-messages.js
-rw-r--r-- 1 peko peko 1.7K Jul  4 16:06 angular-loader.js
-rw-r--r-- 1 peko peko 1.4K Jul  4 16:06 angular-cookies.js

So drop the suboptions and simplify the package logic.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/open2300: fix static builds

Fixes http://autobuild.buildroot.org/results/036/036c9088cfaae421c15b650f35e7ffe6ab934031/

libmysqlclient needs -lz to build, so we have to pass it explicitly for
static builds.

For almost-two-and-a-half years this error has been happenning, noone
ever tried to fix it. This is now done, at last! ;-]

[Peter: add autobuilder reference]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

docs/manual: document BR2_ROOTFS_POST_FAKEROOT_SCRIPT

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/go-bootstrap: Build host tools with host CC

Use the host compiler when building host tools.

The go build system is not compatable with ccache, so use HOSTCC_NOCCACHE
here.  See https://github.com/golang/go/issues/11685.

Signed-off-by: Geoff Levand <geoff@infradead.org>
Reviewed-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

toolchain-external: wrap gfortran

Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/mplayer: needs BR2_TOOLCHAIN_HAS_SYNC_4

Even though mplayer bundles the ffmpeg code base (and ffmpeg is not
affected by this issue), mplayer uses its own build system and doesn't
test properly for the availability (or not) of atomic operations. In
order to keep things simple, we simply make mplayer depend on the
availability of atomic operations, since it assumes they are available.

The reverse dependency is propagated to libplayer and tovid.

Fixes:

  http://autobuild.buildroot.net/results/6306757da3be9bb7e3cdcbfc8abb4e64a3fb7913/

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

toolchain-external: bump Codescape toolchains to 2016.05

Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/mplayer: Fix static linking with tremor & libogg

Fixes
http://autobuild.buildroot.net/results/bc9/bc98fa585399b53ea181dbaf392b93424145911d/

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/nfs-utils: fix indentation

Use tabs everywhere still we don't have an official policy.

Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/nfs-utils: start/stop daemons consistently

Change init script to start and stop all nfs daemons in a consistent
way. Using the same kill command, start/stop the daemon, printout OK or
FAIL and touch or deleted necessary files.

Signed-off-by: Andreas Ehmanns <universeII@gmx.de>
[Maxime:
   - Make it really consistent to stop the daemons as pointed out by
     Yann E. Morin
   - Remove the -9]
Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

qemu: bump to version 2.6.0

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

host-qemu: add support for vde2

Additional configuration for host-qemu package:
- Enable VDE2 support for qemu

Signed-off-by: Simon Maes <simonn.maes@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

vde2: enable building host package

In preparation of enabling vde2 support in host-qemu package.

Signed-off-by: Simon Maes <simonn.maes@gmail.com>
[Thomas:
 - add HOST_VDE2_CONF_OPTS and HOST_VDE2_MAKE, like we have for the
   target variant.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

qemu: add support for host-qemu-system

This commit adds support for building the system emulation mode in
host-qemu. To do so, it adds the BR2_PACKAGE_HOST_QEMU_SYSTEM_MODE and
BR2_PACKAGE_HOST_QEMU_LINUX_USER_MODE sub-options, making sure that the
latter is selected if the former is not enabled. This ensures that at
least one is enabled *and* that existing configurations continue to
build the user-land emulation (which was the only one we supported until
now).

The list of architectures supported by the system emulation mode is the
same as the one for the user-space emulation mode (as far as the
existing list is concerned), so we simply drop the comment about this
dependency list being related to the user-space emulation only.

Signed-off-by: Simon Maes <simonn.maes@gmail.com>
[Thomas:
 - Make sure either BR2_PACKAGE_HOST_QEMU_LINUX_USER_MODE or
   BR2_PACKAGE_HOST_QEMU_SYSTEM_MODE get selected.
 - Unconditionally enable FDT support in host-qemu when in system
   mode (so the option BR2_PACKAGE_HOST_QEMU_FDT has been removed)
 - Remove the unneeded BR2_PACKAGE_HOST_QEMU_HAS_EMULS option.
 - Remove the SDL related option, we really on the system to provide
   SDL.
 - Remove the BR2_PACKAGE_HOST_QEMU_DEBUG and
   BR2_PACKAGE_HOST_QEMU_STRIP_BINARY options, since they are not really
   useful.
 - Remove HOST_QEMU_SITE and HOST_QEMU_SOURCE definitions, since they
   are automatically derived from QEMU_SITE and QEMU_SOURCE anyway.
 - Group things more logically in the .mk file.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

gcc: uClinux supports POSIX threads

Fixes for example boost compile errors for m68k-nommu or
arm cortex-m4.
Following autobuild failures fixed:
http://autobuild.buildroot.net/results/f95511e839d85ea18c5b150a59e994ef069601a1/
http://autobuild.buildroot.net/results/ff2b8be247e50c170b6317886509cc2922e1c77e/

Tested-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/wireless_tools: Pretty fixes

Remove an empty line and wrap long lines

Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/wireless_tools: Remove install man pages

Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

wpa_supplicant: add option to enable Wi-Fi Display

BR2_PACKAGE_WPA_SUPPLICANT_WIFI_DISPLAY depends on
BR2_PACKAGE_WPA_SUPPLICANT_AP_SUPPORT because CONFIG_WIFI_DISPLAY needs
CONFIG_P2P.

Signed-off-by: Georges Savoundararadj <savoundg@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

toolchain-buildroot: add bfin support

With gcc 6.1.0 and binutils 2.26 internal bfin toolchain can be used. A
gcc patch is required, which was reported upstream.

Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

gcc: add hidden symbol for defaults

Some architectures, f.e. Blackfin doesn't support to configure GCC with
--with-cpu to set some CPU specific default CFLAGS (-mcpu=foo). Use a
hidden config symbol to give a hint which architecture supports it,
otherwise add defaults to toolchain wrapper for internal toolchains.

Idea from Thomas Petazzoni.

Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
[Thomas:
 - simplify the Config.in logic with just one option named
   BR2_GCC_ARCH_HAS_CONFIGURABLE_DEFAULTS, defined in package/gcc in one
   place.
 - improve the organization of the code and name of variables.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/gcc: fix list of files of patch file to hash for ccache

When we build the list of patches to include to compute the hash for
ccache, a typo precented the special gcc-initial/ and gcc-final/
directories from a global patch dir to be included in the calculation.

Fix that by properly expanding the $(PKG) variable.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: He Chunhui <hchunhui@mail.ustc.edu.cn>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/f2fs-tools: update version

Use a tag rather than a sha1, it's more obvious.
Remove the patch, it's been applied upstream.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

support/download: don't over-remove files from git archives

When we now manually create git archives, we removed all .git-related
files. However, we also exclude empty directories.

This means that a directory which only had a .gitignore file is excluded
from the archive.

Fixes:
    http://autobuild.buildroot.org/results/2aa/2aa8954311f009988880d27b6e48af91bc74c346/
    http://autobuild.buildroot.org/results/b45/b45cceea99b9860ccf1c925eeda498a823b30903/
    http://autobuild.buildroot.org/results/5ae/5ae336052fd32057d9631649279e142a81f5651f/
    http://autobuild.buildroot.org/results/5fc/5fc3abf4a1aea677f576e16c49253d00720a8bef/

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libbdplus: select libgcrypt in Config.in as well

Fixes http://autobuild.buildroot.org/results/4d9/4d91521378402f3f5b6402ea453df4c14c391bdc/

libbdplus adds libgcrypt to its _DEPENDENCIES, but forgot to select it in
Config.in.

Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/webrtc-audio-processing: needs gcc >= 4.8

Webrtc-audio-processing require a C++11 compiler

Fixes:
http://autobuild.buildroot.net/results/729/729cd0039ac172cfe31a4325bad902e9a14e146f

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

supertuxkart: new package

Note that this package builds and bundles a number of libraries,
such as GLEW and Irrlicht. We are not interested in doing
non-upstreamable changes, so it is not desirable to modify
this choice in any way.

In addition, Supertuxkart builds a version of the angelscript
interpreter. If a compatible version of angelscript is installed
on the system, it's possible to use it.

[Peter: DOS newlines for 0002 patch, tweak comment header]
Signed-off-by: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

scanpypi: fix help text handling

Make sure a help text is terminated with a full stop.

Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

gmp: dissable assembly for armv7m

It has different assembly constraints which can't be satisfied. Fixes:
http://autobuild.buildroot.net/results/ac6/ac65384a325912788790ceafa7db657cef9382fa/

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

ninja: new package

This new package provides the host variant of Ninja, a small build
system with a focus on speed.

[Peter: use install -D]
Signed-off-by: Eric Le Bihan <eric.le.bihan.dev@free.fr>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

board: add support for Chromebook Snow

Chromebook Snow (Samsung XE303C12) is an Exynos 5 board with
a keyboard, an 11 inch screen and a battery attached.
It is relatively developer-friendly and can run mainline Linux
kernels with little to no effort.

There is barely anything special about this target as far as toolchain
is concerned, but its bootloader only accepts signed kernel images
in a Chromium OS specific format, and is not controllable otherwise.

This config provides a script for building the proper kernel blobs,
and a short manual for booting Buildroot images on this device.

In-tree exynos_defconfig is used for the kernel, with a fragment
to change mwifiex into a module. When built statically, mwifiex
attempts to load its firmware before rootfs is mounted and fails.

[Peter: use BR2_KERNEL_HEADERS_AS_KERNEL=y, lock kernel version,
enable fit support in u-boot mkimage]
Signed-off-by: Alex Suykov <alex.suykov@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

babeld: use TARGET_CONFIGURE_OPTS

As suggested by a previous submission of
babeld (https://patchwork.ozlabs.org/patch/631953/), we can use
TARGET_CONFIGURE_OPTS instead of passing CFLAGS and LDFLAGS in
PLATFORM_DEFINES.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

cdrkit: add explicit dependencies for host variant

Signed-off-by: Julien Floret <julien.floret@6wind.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

libvdpau: new package

Signed-off-by: Damien Lanson <damien@kal-host.com>
[Thomas:
 - Add dependency on BR2_PACKAGE_XORG7, which is necessary when
   selecting BR2_PACKAGE_XLIB_LIBX11 and BR2_PACKAGE_XLIB_LIBXEXT.
 - Fix indentation of Config.in help text.
 - Add host-pkgconf to the dependencies, since the configure script uses
   pkg-config to detect dependencies.
 - Add references for the patches.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

fs/common: add option to execute custom scripts under fakeroot

Some users have the need to be able to tweak the content of the target
rootfs with root-like rights, that is, from inside the fakeroot script.

Add a new system option to allow those users to provide a list of
scripts, like the post-build and post-image scripts, that will be run
from our fakeroot script.

[Peter: pass TARGET_DIR to scripts, tweak help text]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Peter Korsgaard <peter@korsgaard.com>
Cc: Cam Hutchison <camh@xdna.net>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

lvm2: fix m68k build error

Disable PIE for m68k as it is not supported.

Fixes following autobuild failures:
http://autobuild.buildroot.net/results/2f976292ddbb8eddb58cc978f209f4f445b5b756/
http://autobuild.buildroot.net/results/7a2b1eef06004886be28744bf1c53d027f8d0123/

Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/transmission: add optional support for libminiupnpc/libnatpmp

Transmission bundles copies of both libraries and uses them if they are
not provided by the system. This patch allows transmission to use the
system libraries of libminiupnpc/libnatpmp.

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libnatpmp: New package

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/wpa_supplicant: Add optional autoscan support

Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/libbdplus: new package

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/libaacs: new package

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

uclibc: update to 1.0.16

Update uClibc-ng to latest stable release (1.0.16).

Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

efivar: disable for some incompatible Sourcery CodeBench toolchains

linux/nvme.h has been renamed to linux/nvme_ioctl.h since 4.4, and we
have a patch to handle that situation. However, some Sourcery CodeBench
toolchains use 4.4 kernel headers but don't have that header file
renamed, and that is causing build failures. So, prevent this package
to be built using those toolchains.

Fixes:

  http://autobuild.buildroot.net/results/66c/66cf68261fa73c2366610f8d14bd02180ba7cd2d/
  http://autobuild.buildroot.net/results/54c/54cc7d4e3b284b35931e08f9c7f0f902df96a989/

Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
[Romain:
  - add back slash on line breaks
  - propagate efivar dependencies to efibootmgr
  - add comments dependencies]
Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/fftw: add fortran support

Signed-off-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
[Thomas: pass FLIBS explicitly to help the configure script.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

toolchain/toolchain-external: update external toolchains that have Fortran support

Signed-off-by: Samuel Martin <s.martin49@gmail.com>
[Vincent: BR2_TOOLCHAIN_EXTERNAL_BLACKFIN_UCLINUX also has fortran]
Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

toolchain/toolchain-external: add knob for fortran support

Signed-off-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

toolchain/toolchain-external: add lib{gfortran,quadmath} to TOOLCHAIN_EXTERNAL_LIBS

Signed-off-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

toolchain/toolchain-external: enable fortran check when it is selected

Signed-off-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

toolchain/helpers: add fortran check

Signed-off-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
[Thomas: remove extension for the generated temporary file, since it's
really an executable, not an object file.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

pkg-cmake.mk: export the fortran compiler path in the CMake toolchain file

Since the fortran support is conditional, only enable it when needed.

Signed-off-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Cc: Benjamin Kamath <bkamath@spaceflight.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/gcc: select BR2_TOOLCHAIN_HAS_FORTRAN when appropriate

This is only for the Buildroot toolchain backend.

Signed-off-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

docs/manual: document dependency on fortran

Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

toolchain: add hidden symbol for fortran support in the toolchain

This symbol should be used in all packages requiring/testing for fortran
support.

Signed-off-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Reviewed-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/gcc: fix fortran support

Fortran depends on libquadmath when available, make the buildroot
toolchain option depends on this new hidden symbol,

[Vincent: only do "HOST_GCC_FINAL_USR_LIBS += libquadmath" for i386 and
x86_64, otherwise it will fail saying "libquadmath.a: file not found"]

Signed-off-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/gcc: complete the logic on libquadmath

libquadmath requires wchar.

So, turn to positive logic and complete it to only enabling quadmath
support when it is available.

Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

toolchain/toolchain-common.in: add BR2_TOOLCHAIN_HAS_LIBQUADMATH hidden symbol

This hidden symbol allow to know when libquadmath can be built and
installed.

Also, declaring this symbol in toolchain-common.in allows to use it in
both external and buildroot toolchain backend.

This will be needed for adding/improving the fortran support.

Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Cc: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

Package nfs-utils: Added check for OK/FAIL

Removed unconditional 'done' output when starting
and stopping the daemons. Instead check return value
of daemon and print out OK or FAIL.

Signed-off-by: Andreas Ehmanns <universeII@gmx.de>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Reviewed-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/quagga: Add systemd.service file

Use a template service file as all of the daemons use almost
identical arguments and generally appear the same to the init
system.

We "Wants=" zebra as that's the daemon for interfacing to the
kernel, and it's not required for the other daemons to work
but it's probably going to be used in nearly all setups.

/usr/bin/env is needed as systemd doesn't allow the instance
variable (%i) in the executable path.

We don't enable these services by default as this would require
creating configuration and /etc/default files. (And is easily
achieved with an FS overlay)

[Peter: remove killmode/killsignal/restartsec as suggested by Maxime]
Signed-off-by: Nathaniel Roach <nroach44@gmail.com>
Reviewed-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

gcc: apply patches stored in global patches directories

When building host gcc, patches stored in global patches directories are
skipped. This patch fixes the unexpected behavior.

Signed-off-by: Chunhui He <hchunhui@mail.ustc.edu.cn>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
[Thomas:
 - rename the loop variable from 'D' to 'patchdir'
 - add some additional comments
 - remove final ; at end of loop when applying the patches, since it's
   not needed]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

core/pkg-virtual: fix fallout of host-dependecy no auto-derivation

In 4bdb067 (infra: remove auto derivation of host dependencies), the
dependencies of host packages are no longer inherited from the
dependencies of the corresponding target package.

However, for virtual packages, there is no provider declared for the
host variant, so we end up with no dependency for the host variant
of virtual packages.

This causes host-luainterpreter to have no provider, and thus breaks the
build of host-luarocks.

To fix that, we use the host variant of the provider of the target
variant of the virtual package to define the provider of the host variant
of the virtual package (re-read it, it's correct!).

Fixes:
    http://autobuild.buildroot.org/results/d8b/d8baf44b1e2f8e0e32c86558fc3e0694235b512a
    http://autobuild.buildroot.org/results/59a/59ac0fe84e5c76a590c698a1483b2a3ae26e1dcf
    http://autobuild.buildroot.org/results/203/2039f24731d8e14cde24c613b487b3db530dd238

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Julien Floret <julien.floret@6wind.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

iprutils: uses fork(), needs MMU

Fixes:

  http://autobuild.buildroot.net/results/6ee89f9a4e05c89375a96ac7085747afc301b896/

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

pkg-generic: check that SITE has a value when SOURCE

Currently, when an user forgets to define a SITE, he gets some fairly
weird download failure. In order to make things easier to diagnose, this
commit adds a check in the package infrastructure that verifies that if
SOURCE has a value, SITE is not empty.

Cc: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

toolchain-external: remove SOURCE/SITE for locally available toolchain

When the custom external toolchain is locally available, we currently
define SITE/SOURCE to empty variables. Now that the package
infrastructure doesn't define a value for SOURCE when VERSION is empty,
it doesn't attempt to download a file anymore, so we can get rid of
those empty SOURCE/SITE variables in the toolchain-external package.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

mcookie: remove SOURCE variable

Now that the package infrastructure doesn't attempt to download a package
that has an empty version string, there's no need to define the SOURCE
variable in the mcookie package.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

tz: remove SOURCE variable

Now that the package infrastructure doesn't attempt to download a package
that has an empty version string, there's no need to define the SOURCE
variable in the tz package.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

skeleton: remove SOURCE variable

Now that the package infrastructure doesn't attempt to download a package
that has an empty version string, there's no need to define the SOURCE
variable in the skeleton package.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

mkpimage: remove SOURCE

Now that the package infrastructure doesn't attempt to download a package
that has an empty version string, there's no need to define the SOURCE
variable in the mkpimage package.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

mkpasswd: remove VERSION/SOURCE

Now that the package infrastructure doesn't attempt to download a package
that has an empty version string, there's no need to define the VERSION and
SOURCE variables in the mkpasswd package.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

mke2img: remove SOURCE

Now that the package infrastructure doesn't attempt to download a package
that has an empty version string, there's no need to define the SOURCE
variable in the mke2img package.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

initscripts: remove SOURCE

Now that the package infrastructure doesn't attempt to download a package
that has an empty version string, there's no need to define the SOURCE
variable in the initscripts package.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

getent: remove VERSION/SOURCE

Now that the package infrastructure doesn't attempt to download a package
that has an empty version string, there's no need to define the VERSION and
SOURCE variables in the getent package.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

pkg-virtual: remove VERSION/SOURCE

Now that the package infrastructure doesn't attempt to download a package
that has an empty version string, there's no need to define the VERSION and
SOURCE variables in the virtual package infrastructure.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

makedevs: remove VERSION/SOURCE

Now that the package infrastructure doesn't attempt to download a package
that has an empty version string, there's no need to define the VERSION and
SOURCE variables in the makedevs package.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

core: do not attempt downloads with no _VERSION set

Currently, when the package defines no _SOURCE and no _VERSION
variables, we end up trying to download <PACKAGE>-undefined.tar.gz.

This is utterly pointless.

Change the logic to not attempt a download when neither _SOURCE
nor _VERSION is set:
  - do not set an undefined _VERSION to "undefined"
  - do not append an empty (or "undefined") version to the package
    rawname

Consequently, for packages that have no _VERSION, the build
directory will be just the package name (which to some may look
nicer than the current "package-undefined").

Furthermore, the message trace is also a bit leaner (to the same
people that find "package" nicer than "package-undefined").

Signed-off-by: Alex Suykov <alex.suykov@gmail.com>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

cups: bump to version 2.1.4, change upstream location

cups is now available from Github, with tarballs uploaded by the
maintainers, so we use this new upstream location.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

cups: fix static linking with GnuTLS

cups forgets to link against the libs found by pkg-config if it is build with
GnuTLS support.

Otherwise, static build fails with:
../cups/libcups.a(tls.o): In function `http_gnutls_create_credential':
cups-2.1.2/cups/tls-gnutls.c:762: undefined reference to `gnutls_x509_crt_init'
[..]

Fixes:
http://autobuild.buildroot.net/results/205/20509b7d1edc73dc2a7622f2ae7dbf9f6665a918/
http://autobuild.buildroot.net/results/b86/b86d685caf8cf812f9505fca2de36fb8c18baee1/

Patch is sent upstream.

Signed-off-by: Jörg Krause <joerg.krause@embedded.rocks>
[Thomas: refresh and rename patch.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

cups: reformat patch with Git

All patches in the cups package are Git formatted, except
0004-remove-pie.patch, so this commit makes things consistent by also
making this patch Git formatted.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

sunxi-mail: hack to work around autobuilder issue

Fixes http://autobuild.buildroot.org/results/5a5/5a50be9feeb5da7694e9a42f615f0b7276319aeb/

The recent change to use git submodules for sunxi-mali instead of the github
wrapper without changing the version unfortunately doesn't work for people
who have already built sunxi-mali, as the existing github tarball will be
used instead of buildroot doing a clean git clone (with submodules).

The sunxi-mali commit we are using is the latest in the repo, so we cannot
fix it by bumping the version.  Instead simply remove the last character of
the VERSION sha1 so the existing tarballs are no longer used.

Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libminiupnpc: New package

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/sunxi-mali: remove useless package

sunxi-mali-prop only served a feeding prey for sunxi-mali to get the
proprietary libs.

Now that sunxi-mali downloads them (as the git submodule they are), we
no longer need sunxi-mali-prop.

Remove it.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/sunxi-mali: use git submodules

The sunxi-mali git tree is set up with a submodule that contains the
actual libraries. Since there was no support for git submodules so far,
we used a dirty trick to have that submodule cloned by its own package,
with sunxi-mali just vampirising it into its own build dir.

Now that we have support for git submodules, we can simplify stuff a
bit, and have sunxi-mali directly download the proprietary libs itself.

Remove the .hash file since we're not checking hashes for git clones.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>