support/download/git: add support for submodules

Some git repositories may be split into a master repository and
submodules. Up until now, we did not have support for submodules,
because we were using bare clones, in which it is not possible to
update the list of submodules.

Now that we are using plain clones with a working copy, we can retrieve
the submdoules.

Add an option to the git download helper to kick the update of
submodules, so that they are only fetched for those packages that
require them. Also document the existing -q option at the same time.

Submodules have a .git file at their root, which contains the path to
the real .git directory of the master repository. Since we remove it,
there is no point in keeping those .git files either.

Note: this is currently unused, but will be enabled with the follow-up
patch that adds the necessary parts in the pkg-generic and pkg-download
infrastructures.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Tested-by: Matt Weber <matt@thewebers.ws>
Reviewed-by: Matt Weber <matt@thewebers.ws>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

support/download/git: do not use git archive, handle it manually

We currently use git-archive to generate the tarball. This is all handy
and dandy, but git-archive does not support submodules. In the follow-up
patch, we're going to handle submodules, so we would not be able to use
git-archive.

Instead, we manually generate the archive:
  - extract the tree to the requested cset,
  - get the date of the commit to store in the archive,
  - store only numeric owners,
  - store owner and group as 0 (zero, although any arbitrary value would
    have been fine, as long as it's a constant),
  - sort the files to store in the archive.

We also get rid of the .git directory, because there is no reason to
keep it in the context of Buildroot. Some people would love to keep it
so as to speed up later downloads when updating a package, but that is
not really doable. For example:
  - use current Buildroot
  - it would need foo-12345, so do a clone and keep the .git in the
    generated tarball
  - update Buildroot
  - it would need foo-98765
For that second clone, how could we know we would have to first extract
foo-12345 ? So, the .git in the archive is pretty much useless for
Buildroot.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Tested-by: Matt Weber <matt@thewebers.ws>
Reviewed-by: Matt Weber <matt@thewebers.ws>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

support/download/git: do not use bare clones

Currently, we are using bare clones, so as to minimise the disk usage,
most notably for largeish repositories such as the one for the Linux
kernel, which can go beyond the 1GiB barrier.

However, this precludes updating (and thus using) the submodules, if
any, of the repositories, as a working copy is required to use
submodules (becaue we need to know the list of submodules, where to find
them, where to clone them, what cset to checkout, and all those is
dependent upon the checked out cset of the father repository).

Switch to using /plain/ clones with a working copy.

This means that the extra refs used by some forges (like pull-requests
for Github, or changes for gerrit...) are no longer fetched as part of
the clone, because git does not offer to do a mirror clone when there is
a working copy.

Instead, we have to fetch those special refs by hand. Since there is no
easy solution to know whether the cset the user asked for is such a
special ref or not, we just try to always fetch the cset requested by
the user; if this fails, we assume that this is not a special ref (most
probably, it is a sha1) and we defer the check to the archive creation,
which would fail if the requested cset is missing anyway.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Tested-by: Matt Weber <matt@thewebers.ws>
Reviewed-by: Matt Weber <matt@thewebers.ws>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/boost: add boot-type_erasure support

Until now this library wasn't handled by the Buildroot packaging,
so add a new config option to disable it when requested.

Also disable boost-type_erasure for the host variant to keep it as
minimal as possible.

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

Config.in: indentation cleanup

Signed-off-by: Jerzy Grzegorek <jerzy.grzegorek@trzebnica.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/samba4: Change tmpfiles path

Per the documentation:
https://www.freedesktop.org/software/systemd/man/tmpfiles.d.html

The order of path by priorites is:
/etc/tmpfiles.d/*.conf
/run/tmpfiles.d/*.conf
/usr/lib/tmpfiles.d/*.conf

For the user to be able to override our tmpfiles easily, it's better to
place our files in /usr/lib/tmpfiles.d/

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>

sphinxbase: new package

CMUSphinx toolkit is a speech recognition toolkit with various tools
used to build speech applications. CMU Sphinx toolkit has a number of
packages for different tasks and applications. The toolkit is designed
for use on mobile platforms.

Signed-off-by: Jonathan Ben-Avraham <yba@tkos.co.il>
[Thomas:
 - license is BSD-2c, not MIT
 - add --without-lapack]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/systemd: Sort menuconfig alphabetically

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

systemd-bootchart: New package

Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
[Thomas:
 - remove Config.in.legacy option
 - rename the hash file to the correct name
 - add license files.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/systemd: Remove systemd-bootchart

Since systemd v230, bootchart has been removed from the source of
systemd and now lives in its own repository.
A new package will be added in the next commit.

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

package/systemd: Remove SYSTEMD_COMPAT option

Since v230 of systemd, the compat libraries have been fully removed.
https://github.com/systemd/systemd/commit/4de282cf9324ab13d17ac334244d0d7cae2df37d

Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
[Thomas: rewrap Config.in help text in Config.in.legacy and put the
option at the right place.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/systemd: Default yes for networkd

It seems logical to enable a few options along with systemd when it's
the init system.

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

package/systemd: Default yes for timesyncd

It seems logical to enable a few options along with systemd when it's
the init system.

Also change the help as timesyncd is a daemon that implements an SNTP client.

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

package/systemd: Create user only when needed

We don't want to create a bunch of group and users that might not be
used, it's more elegant to create them when the associated features is
selected.

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

package/systemd: Add optional support for resolved

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

package/systemd: Add optional support for myhostname

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

package/ffmpeg: cosmetic change for the mips fpu option

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

package/freerdp: add support for ffmpeg 3.0

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

dmalloc: fix static build

For example coldfire uses FLAT binaries, which are statically
build. Fixes following autobuild failure:
http://autobuild.buildroot.net/results/cc4233a4c3f92275a4071c172330c920c35219ca/

Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
[Thomas: use an intermediate variable rather than duplicating code.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

fftw: Disable combined threads when building with OpenMP support

Otherwise fftw fails to configure:
configure: error: --with-combined-threads incompatible with --enable-openmp

Signed-off-by: André Hentschel <nerv@dawncrow.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/libdvdread: fix static compile by adding libdvdcss as dependency

If libdvdcss support was not found during configure libdvdread tries
to open libdvdcss.so nonetheless:
https://code.videolan.org/videolan/libdvdread/blob/master/src/dvd_input.c#L275

This will obviously fail in a static-only configuration. To fix the
problem add libdvdcss as hard dependency to this package.

Fixes
http://autobuild.buildroot.net/results/415/41534b70a66c6297652f0944f9ecc7e780b26c45/

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

package/libiio: libiio network backend needs threads

When libiio network backend is enabled, NEED_THREADS is set in the
build system and require a toolchain with treads support.

Fixes:
http://autobuild.buildroot.net/results/cfc/cfc70094d8f4008f4e660250b1b90a35a2ab384d

Signed-off-by: Romain Naour <romain.naour@smile.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

babeld: New package

Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
[Thomas: use $(INSTALL) instead of install.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

memcached: bump to version 1.4.28

Removes the 64-bit atomic intrinsics constraint previously introduced.

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

liboping: bump to version 1.9.0

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

configs/qemu: update aarch64 to linux 4.6.3

The console problem goes away with qemu 2.6.0 so it seems it's now a
requirement, update readme.txt accordingly.

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

configs/pc_x86_64_{bios, efi}_defconfig: bump to linux 4.6.3

To match it's friendly qemu counterpart.

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

configs/qemu: bump to linux 4.6.3

In order to keep them in sync with the new configs, and avoid shipping a
vulnerable kernel.

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

package/libuio: fix build with musl

limits.h header is missing in attr.c, base.c and helper.c to provide
PATH_MAX.

Fixes:
http://autobuild.buildroot.net/results/702/7023104e6018ea46c54073ddbe5119d0f66ae5a3

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

package/libuio: disable Werror

Disable -Werror to avoid build issues caused by warnings.

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

package/libuio: fix build with external intl library

When intl library is not provided by the libc, libuio forgets to link
with -lintl.

Fixes:
http://autobuild.buildroot.net/results/8f7/8f715b4e9b826dac13f3d49da25576e6e1d79449

[Peter: simplify using LIBUIO_LIBS as suggested by Yann]
Signed-off-by: Romain Naour <romain.naour@smile.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

a20_ollinuxino: Accommodate U-Boot environment on SD-card

As described by http://linux-sunxi.org/Bootable_SD_card#SD_Card_Layout

The space betweem 544KB..1MB is reserved / used for the u-boot environment,
so mark it as such to make sure genimage doesn't put other partitions here.

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

pcduino: Accommodate U-Boot environment on SD-card

As described by http://linux-sunxi.org/Bootable_SD_card#SD_Card_Layout

The space betweem 544KB..1MB is reserved / used for the u-boot environment,
so mark it as such to make sure genimage doesn't put other partitions here.

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

orangepipc: Accommodate U-Boot environment on SD-card

As described by http://linux-sunxi.org/Bootable_SD_card#SD_Card_Layout

The space betweem 544KB..1MB is reserved / used for the u-boot environment,
so mark it as such to make sure genimage doesn't put other partitions here.

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

cubieboard2: Accommodate U-Boot environment on SD-card

According to current "include/configs/sunxi-common.h"
U-Boot environment is stored on SD-card with offset 544 kB
and its size is limited as 128 kB.

Given memory layout of sunxi boards:
  1. SPL
  2. U-Boot
  3. U-Boot environment
  4. Partitions

(see http://linux-sunxi.org/Bootable_SD_card#SD_Card_Layout for details)

we need to make sure the first real partition starts at least
with 544+128 kB offset. Adding a small safety gap let's put
the first partition with offset 1MB from the beginning of SD-card.

Otherwise current setup gets broken as soon as one saves U-Boot
environment with "saveenv" which basically overwrites FAT partition with
zImage and board no longer boots into Linux kernel.

[Peter: extend description, add padding to u-boot partition instead]
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Erico Nunes <nunes.erico@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Peter Korsgaard <peter@korsgaard.com>
Cc: Carlos Quijano <carlos@crqgestion.es>
Cc: Andi Shyti <andi@etezian.org>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/busybox: support spaces in module aliases in mdev

The new change which enabled automatic module loading on boot does not handle
the cases when module alias includes spaces. It prevents modules to be loaded
since script fails:

  % find /sys/ -name modalias | xargs sort -u
  sort: /sys/devices/platform/Fixed: No such file or directory

First alias in question is "platform:Fixed MDIO bus".

Amend the script to support above like cases.

Fixes: 07f46c2b6dae ("package/busybox: support automatic module loading with mdev")
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/systemd: Force enable blkid

Since systemd select BR2_PACKAGE_UTIL_LINUX_LIBBLKID, we can safely
force this option.

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

package/systemd: Add optional libxkbcommon support

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

package/systemd: add optional support for qrencode

The support of qrencode depends of the journal gateway feature as it is
used to display the key used to seal the journal.

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

package/systemd: add optional support for libidn

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

package/systemd: Disable IMA support

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

package/systemd: Disable the test coverage

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

package/systemd: add optional support for audit

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

package/systemd: remove unnecessary make opts

These flags were originally needed to support uClibc in earlier releases.

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

package/systemd: only create 'systemd-journal' group

systemd-journal is a group used for controlling user access to the
journal when used with tools like journalctl.

While we're at it, sort the users list to improve readability.

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

package/systemd: use 'menuconfig'

systemd has a bunch of options now, so use 'menuconfig' to the group the
options in a submenu.

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

package/systemd: add optional support for hibernation

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

package/systemd: add optional support for polkit

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

package/systemd: add optional support for coredump

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

package/systemd: add optional support for localed

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

package/systemd: add optional support for timedated

Signed-off-by: Gabe Evans <gabe@hashrabbit.co>
[Maxime: Add default y]
Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/systemd: add optional support for hostnamed

Signed-off-by: Gabe Evans <gabe@hashrabbit.co>
[Maxime: add default y]
Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/systemd: add optional support for importd

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

package/systemd: add optional support for machined

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

package/systemd: add optional support for logind

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

package/systemd: add optional support for rfkill

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

package/systemd: add optional support for backlight

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

package/systemd: add optional support for random-seed

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

package/systemd: add optional support for firstboot

Signed-off-by: Gabe Evans <gabe@hashrabbit.co>
[Maxime: Add default y]
Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/systemd: add optional support for sysusers

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

package/systemd: add optional support for tmpfiles

Signed-off-by: Gabe Evans <gabe@hashrabbit.co>
[Maxime: Add default y]
Signed-off-by: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/valijson: indentation cleanup

Signed-off-by: Jerzy Grzegorek <jerzy.grzegorek@trzebnica.net>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/libcoap: indentation cleanup

Signed-off-by: Jerzy Grzegorek <jerzy.grzegorek@trzebnica.net>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

jsmin: add menuconfig entry for host variant

Signed-off-by: Julien Floret <julien.floret@6wind.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

reproducibility: override locale and timezone

And export BR2_REPRODUCIBLE for post-build / post-image scripts.

[Peter: Extend commit message,
move export together with our other exports,
add comment explaining why we override local/timezone]
Signed-off-by: Gilles Chanteperdrix <gilles.chanteperdrix@xenomai.org>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

reproducibility: introduce config knob

Making the builds reproducible involves turning off build timestamps,
which some users may find annoying. So make the reproducible builds
optional.

Signed-off-by: Gilles Chanteperdrix <gilles.chanteperdrix@xenomai.org>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

audit: bump to version 2.6.2

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

package/{libsepol,libselinux}: fix build on ancient distros

Ancient distros (especially enterprise-grade still in use) do not accept
ln --relative.

Revert the upstream commit, since the problem it was trying to fix is
not applicable in the context of Buildroot (even with a merged /usr).

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

audit: add patch to fix EHWPOISON build issue

Fixes:

  http://autobuild.buildroot.net/results/2e147b60481cfb8960749a8e012ff8fa05568a2c/

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

libsepol: add missing dependency on host-flex

Fixes:

  http://autobuild.buildroot.net/results/580037bd376188e18a4a58927c929d21e9b2b03c/

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

python-scapy3k: new package

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

package/cmake: with BR2_ENABLE_DEBUG use RelWithDebInfo

>From the documentation on BR2_ENABLE_DEBUG, the intention is to get
a build with debug symbols and not a "debug build" since that can have
the unintended consequence of being a different code path then a
release build type definition.

Switch the "Debug" to "RelWithDebInfo" in the cmake package support
to accomodate getting the debug symbols and still be a release
build.

Signed-off-by: Charles Hardin <ckhardin@exablox.com>
Reviewed-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/Makefile.in: set variables for legacy f77 checks

For fortran detection, some projects check for fortran availability
using the FC/FCFLAGS variables, and others for the legacy F77/FFLAGS
ones.
So, make sure the legacy fortran F77 and FFLAGS variables are set in
TARGET_CONFIGURE_OPTS.

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

package/Makefile.in: set TARGET_FCFLAGS variables

TARGET_FCFLAGS is already added to TARGET_CONFIGURE_OPTS, but was not
defined so far. This change fixes this.

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

package/gcc: wrap gfortran

gfortran supports all options supported by gcc, so it can and should be called
via the toolchain wrapper.

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/busybox: enable ttyUSB[0-9] when using mdev

Some of our systems are using ttyUSB0 devices as a serial console.
Add them to mdev.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/libuio: new package

Signed-off-by: Joris Lijssens <joris.lijssens@gmail.com>
[Romain:
 - fix autoreconf issue by using GETTEXTIZE and AUTORECONF
 - fix issue with missing makeinfo
 - move version number comment in .mk]
Signed-off-by: Romain Naour <romain.naour@smile.fr>
[Thomas: add GPLv2 license, used for the programs.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/linux-headers: do not default to any version for custom headers

Currently, for a custom headers version, or for the same headers as the
kernel, wedefault to a "very old" version (i.e. 2.6.x in practice).

However, as Vivien explained, when using the same headers as the kernel,
and the kernel is set to use the default version (aka latest version
known to Buildroot) of the kernel, one would expect the headers are
automatically tracking the latest version. Off course, that expectation
is broken because of the above.

However, whatever version we default to, it will probably not be
correct, whether we default to the latest version or to the "very old"
version.

So, simply drop the specific default version, so the default is now the
latest version.

Note: that has the potential to break existing defconfig files that
relied on the "very old" version to be the default. Well, whatever,
they'll get a build failre quite early, and it is easy to fix.

Reported-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

python-xlutils: new package

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

python-xlwt: new package

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

package/gcc: use BR2_USE_WCHAR to set gcc libquadmath option

BR2_TOOLCHAIN_BUILDROOT_WCHAR is only defined when uclibc is selected,
whereas BR2_USE_WCHAR is always defined. Due to this, we were disabling
quadmath support even with glibc or musl.

So, use BR2_USE_WCHAR to drive the gcc libquadmath option.

In addition, invert the logic of the condition to use positive logic,
and rework the comment to no longer mention gcc 4.6: libquadmath still
exists, and it still requires wchar 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>

package/linux-headers: improve help text and prompt

The way the help text and prompt for the kernel header version choice is
written can be misleading to some users.

Rewite them so it is more explicit.

Reported-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

lzop: fix build of host variant with gcc6

Signed-off-by: Jaap Crezee <jaap@jcz.nl>
[Thomas: minor tweaks in commit log and patch.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

efivar: depends on target gcc >= 4.7

The efivar internal build system uses flags such as -std=gnu11 ,
-Wmaybe-uninitialized , -flto which are available in gcc >= 4.7 only.
These flags are passed both to target build and to build a host tool.
For the host part, this has been worked around by overriding
'gcc_flags', but doing that for the target would remove all flags that
upstream intended for the target build.

Buildroot doesn't support building gcc 4.6 in its own toolchain anymore
but it's possible to use an external toolchain with gcc 4.6 which would
be unable to build this package.

This patch adds a limitation on toolchains with target gcc >= 4.7 to
make sure that the flags are available in the chosen toolchain.

Signed-off-by: Erico Nunes <nunes.erico@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

openblas: new package

OpenBLAS is optimised for specific CPU models, which don't fully match
with the GCC code generation options. Therefore, we can't automatically
select BR2_PACKAGE_OPENBLAS_TARGET based on the CPU choice. Instead, let
the user select the TARGET name, but offer a sensible default. Other
possible solutions were deemed too complicated: adding choice options in
the ambiguous cases, or only making the option user-visible when there
is ambiguity.

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

efivar: force shared libgcc for Nios II

efivar for Nios II also has to be linked with shared libgcc to avoid an
error due to FDE encoding in static libgcc when linking libefiboot.so.

Fixes:
http://autobuild.buildroot.net/results/0c9/0c90e3e7ad41d21dd832f6f266af35fc19185170/

Signed-off-by: Erico Nunes <nunes.erico@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/upx: fix build with host gcc 6.x

Disable the -Werror gcc option because is causes the following error
when using gcc 6.x on the host.

 p_vmlinx.cpp:100:5: error: this ‘if’ clause does not guard...
  [-Werror=misleading-indentation] if (a->p_paddr > b->p_paddr) return  1;

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

package/ucl: fix build with host gcc 6.x

ucl fails with the following error when using gcc 6.x on the host:

 checking whether your compiler passes the ACC conformance test... FAILED
 configure:
 configure: Your compiler failed the ACC conformance test - for details see
 configure: `config.log'. Please check that log file and consider sending
 configure: a patch or bug-report to <markus@oberhumer.com>.
 configure: Thanks for your support.
 configure:
 configure: error: ACC conformance test failed. Stop.

Fix the issue by using the ISO C90 standard.

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

python-tomako: new package

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

python-tornado: fix runtime dependencies

According to setup.py tornado requires backports.ssl_match_hostname
for Python < 3.2 and certifi for Python < 3.4.

Actually both requirements are optional and tornado alone can work
without them, but Python packages like python-circus check requirements
at runtime and though they don't use this funtionality, they insist on
having these packages installed.

An upstream patch fixes backports.ssl_match_hostname dependency and
selecting python-certifi for Python 2 fixes certifi dependency till
this is fixed upstream.

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

python-socketio: new package

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

python-engineio: new package

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

configs/qemu_mips64r6el_malta_defconfig: new configuration

Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Tested-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

configs/qemu_mips64r6_malta_defconfig: new configuration

Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Tested-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

configs/qemu_mips32r6el_malta_defconfig: new configuration

Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Tested-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

configs/qemu_mips32r6_malta_defconfig: new configuration

Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Tested-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

configs/qemu_mipsel_malta_defconfig: rename to qemu_mips32r2el_malta_defconfig

Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Tested-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

configs/qemu_mips_malta_defconfig: rename to qemu_mips32r2_malta_defconfig

Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
Tested-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

netsnmp: add OK/FAIL output in init script

This commit reworks the output of the start(), stop() and reload()
functions. The return values of start-stop-daemon are now checked and a
OK or FAIL message is printed out.

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

netsnmp: add missing reload of snmptrapd in init script

In case of a reload command the snmpdtrapd wasn't restarted.

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