sdl_mixer: fix static linking issue

LDFLAGS must be passed when creating binaries.

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

uclibc: add various upstream patches

Fix some static linking issues for Xtensa/Blackfin.
Optimize the size for non-threaded static binaries and
fix static C/C++ applications when running on the target.
Fix noMMU madvise declarations.

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

DEVELOPERS: add Clayton Shotwell for policycoreutils

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

DEVELOPERS: regroup the two entries for Clayton Shotwell

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

policycoreutils: new package

This package contains the core policy utilities that are required
for basic operation of an SELinux system.

Signed-off-by: Clayton Shotwell <clayton.shotwell@rockwellcollins.com>
Signed-off-by: Matt Weber <matthew.weber@rockwellcollins.com>
Signed-off-by: Niranjan Reddy <niranjan.reddy@rockwellcollins.com>
Tested-by: Bryce Ferguson <bryce.ferguson@rockwellcollins.com>
Signed-off-by: Bryce Ferguson <bryce.ferguson@rockwellcollins.com>
[Thomas:
 - Move the Config.in comment at the top of the Config.in file rather
   than between the main option and its sub-options, as this breaks
   menuconfig indentation.
 - Fix the propagation of the libsemanage dependencies. libsemanage
   depends on BR2_PACKAGE_AUDIT_ARCH_SUPPORTS and
   BR2_TOOLCHAIN_USES_GLIBC which were not accounted for. Since it
   depends on BR2_TOOLCHAIN_USES_GLIBC, then all the gettext related
   handling becomes useless and has been removed.
 - Rename the prompt of the restorecond sub-option to just
   "restorecond".
 - Use TARGET_CONFIGURE_OPTS and HOST_CONFIGURE_OPTS instead of
   passing LDFLAGS, CC, etc. manually.
 - Use make "foreach" function for loops instead of shell "for" loops.
 - Rework the explanation of why we're passing DESTDIR at build time.
 - Minor formatting tweaks here and there.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

DEVELOPERS: add Ash as developer for PRU packages

Signed-off-by: Ash Charles <ash.charles@savoirfairelinux.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

uclibc-ng-test: fix compile errors with glibc toolchains

We need to link with -lrt for some tests.
Update to latest upstream version.

Fixes:
 http://autobuild.buildroot.net/results/7388ffcc51a6686b8b708d7373427368d3755870
 http://autobuild.buildroot.net/results/926716ed127dea879c5b26c4d04dc7f0615c8f34
 http://autobuild.buildroot.net/results/737f1651a8bba7d3d815f46ec796955d09ba96bd
 http://autobuild.buildroot.net/results/56a8d67a45469d2a478da8d7d1ad7c5fb5e6c0c5

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

package/qemu: separate user and system arch on powerpc64le

QEMU provides a single system emulator that supports both powerpc64
and powerpc64le with a target called 'ppc64-softmmu', but it provides
a different usermode emulator for each one (with targets
'ppc64le-linux-user' and 'ppc64-linux-user').

Due to this asymmetry it is not possible to support both cases with
the single arch value used in the package file. This patch introduces
an additional value into the package configuration,
HOST_QEMU_SYS_ARCH, so that both cases can be supported.

Fixes commit d2ff457e88ff4106f97e29516d467978827d30c2
and autobuilder failture
http://autobuild.buildroot.net/results/a2d63e21c3e82c36f4a975e90ed56faba18e97a5

Signed-off-by: Sam Bobroff <sam.bobroff@au1.ibm.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

DEVELOPERS: add Ezequiel Garcia as developer for mimic package

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

mimic: new package

Signed-off-by: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>
[Thomas:
 - Rewrap Config.in help text
 - Use "depends on" rather than "select" for the audio backend options
 - Slightly simplify some of the prompts for the audio backend selection
 - Remove MIMIC_INSTALL_STAGING = NO, that's the default
 - Use += when assigning MIMIC_DEPENDENCIES
 - Remove double quotes when setting --with-audio=.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

cjson: bump to version 1.1.0

Signed-off-by: Fabrice Fontaine <fabrice.fontaine@orange.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

swupdate: replace BR2_PREFER_STATIC_LIB with BR2_STATIC_LIBS

The symbol to control static compilation was renamed in 2015.02, but missed
when swupdate was added.

Signed-off-by: Danomi Manchego <danomimanchego123@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/efl: add Wayland support

The EFL Wayland support was removed with commit [1] since the dependecy
on libdrm was missing. Also it requires OpenGL ES with EGL, Evas DRM
and Evas GLES DRM support [2].

As stated in configure, Evas GLES DRM engine support (gl_drm) depends
on wayland-client to build (wayland-client >= 1.8.0).

So, enable gl_drm only when wayland support is selected.

[1] 4f04be1659f186765f506c68f5bfbf6845fc40dc
[2] https://www.enlightenment.org/about-wayland

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Cc: Pierre Floury <devpfl@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/efl: enable libdrm support

drm can be used without compositor, just like it was with
the framebuffer for standalone applications

As stated in configure.ac, libdrm support needs libdrm, elput,
libxkbcommon and libgbm.

libgbm is only provided by mesa3d package when OpenGL EGL support is
enabled, so add a direct dependency on mesa3d.

Rework the libxkbcommon dependency since it's now required for
elput and libdrm support.

[1] https://www.enlightenment.org/about-wayland

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Pierre Floury <devpfl@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/xen: Disable Werror checking

The strict unused-const-variable checking was causing autobuilder errors
when trying to build Xen tools/libxl as the migrate_*[] arrays are not
always accessed.

To avoid the error edit the Makefile to stop all general warnings being treated
as errors, by removing the -Werror flag.

Fixes:
  http://autobuild.buildroot.net/results/0e0/0e0d4aa4a05da5804821951289c0a4049b009c61/

Signed-off-by: Alistair Francis <alistair.francis@xilinx.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

ezxml: disable mmap code for nommu

Fixes:
http://autobuild.buildroot.net/results/8e1/8e10052a0d0ad4f47622925349ec5fddb7b672f0/
http://autobuild.buildroot.net/results/2f1/2f15b81beadd2d902c73d6d8bd7fb06b03f62ef5/

The mmap support uses madvise which isn't available on nommu uClibc, so disable it.

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

sysstat: bump to version 11.4.2

[Peter: add local sha256 hash as suggested by Vicente]
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

dovecot-pigeonhole: bump version to 0.4.16

Release notes:
  http://www.dovecot.org/list/dovecot-news/2016-October/000330.html

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

dovecot: bump version to 2.2.27 (security)

Fixes CVE-2016-8652 : http://www.securityfocus.com/bid/94639/

Release notes:
  http://www.dovecot.org/list/dovecot-news/2016-December/000333.html

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

links: security bump to version 2.14

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

php: security bump to version 7.0.14

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

gvfs: bump to version 1.30.2

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

grep: bump to version 2.27

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libglib2: bump to version 2.50.2

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

vala: bump to version 0.34.4

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

gnutls: bump to version 3.4.17

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

harfbuzz: bump to version 1.3.4

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

poppler: bump to version 0.49.0

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

mpv: bump to version 0.22.0

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

cairo: bump to version 1.14.8

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

coreutils: bump to version 8.26

Pass on INSTALL_PROGRAM to configure since starting with version 8.26
when building natively it tries to use the self-built install to
install, and when cross-compiling it expects INSTALL_PROGRAM to point to
a real install, or otherwise fails when recursively trying to expand it.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libunistring: bump to version 0.9.7

Add new GPLv2 license option.
Unfortunately COPYING is GPLv3, so avoid it.

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

gtksourceview: bump to version 3.22.2

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

libinput: bump to version 1.5.3

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

linux: bump default to version 4.8.13

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

linux-headers: bump 4.{8, 4, 1}.x and 3.{18, 12}.x series

Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

rpm: add optional dependencies on elfutils and binutils

Since these are detected by the configure script, we should explicit
their dependencies in rpm.mk.

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

rpm: add patches to fix bfd.h related build issues

This commit adds two patches to the rpm package to fix two separate but
related build issues:

 - The first patch fixes a build that occurs when rpm is built after
   elfutils, but without binutils. In this case, dwarf.h is present, so
   rpm enables the build of a number of additional tools. But one of
   them uses bfd.h, provided by binutils, without checking for its
   existence. So the first patch fixes that by checking for bfd.h
   existence before enabling the tool.

   Fixes:

     http://autobuild.buildroot.net/results/810/810e24cab65f67d143da29c78c0f89d37a851cd7/build-end.log

 - The second patch fixes a build issue that occurs when rpm is built
   after both dwarf and binutils. In this case, bfd.h complains because
   config.h is not included. That's a weird and silly issue in bfd.h
   that the binutils developers don't want to fix, and you have to
   define PACKAGE or PACKAGE_VERSION before including bfd.h to use it
   outside of binutils.

   Fixes:

     http://autobuild.buildroot.net/results/362/362502f89631c9ba1d71906527674657ccff01ed/build-end.log

Thanks a lot to James Knight <james.knight@rockwellcollins.com> for the
initial investigation about the first issue.

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

package/efl: bump to version 1.18.4

Bug fix release:
https://www.enlightenment.org/news/efl-1.18.4

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

systemd: don't build systemd-firstboot by default

systemd-firstboot is never invoked since systemd's first boot detection
logic checks whether /etc/machine-id exists. Since the file is created
automatically by systemd.mk, systemd will never detect first boot and
therefore the systemd-firstboot.service unit file will never get run.

Additionally, if /etc/machine-id is removed to allow systemd-firstboot
to run, it interactively prompts for the system locale. This makes it
seem unlikely that an embedded system would want to use it.

Signed-off-by: Frank Hunleth <fhunleth@troodon-software.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/xen: Bump Xen to 4.7.1

Signed-off-by: Alistair Francis <alistair.francis@xilinx.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

arch/Config.in.arm: support thumb2 instructions for ARMv8 in 32bit mode

The ARMv8 cores all support thumb2 instructions when running in aarch32 mode.

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

arch/Config.in.arm: only enable BR2_ARM_CPU_HAS_NEON for ARMv8 in 32bit mode

A number of packages use BR2_ARM_CPU_HAS_NEON to know if the target handles
aarch32 neon instructions, which is only true for ARMv8 cores when they are
running in 32bit mode.

Notice: These cores do support neon-like instructions using a different
encoding in 64bit mode (it is a required part of ARMv8, similar to the FPU).

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

arch/Config.in.arm: only enable BR2_ARM_CPU_HAS_ARM for ARMv8 in 32bit mode

Fixes:
http://autobuild.buildroot.net/results/5e6/5e67cc067a06f7364cde1a8393ea72608fe7fef1/

A number of packages use BR2_ARM_CPU_HAS_ARM to know if the target handles
classic A32 instructions, which is only true for ARMv8 cores when they are
running in 32bit mode.

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

uclibc-ng-test: disable some parts of test-suite

The math tests are experimental at the moment and require more
adaptions before they can be enabled again.
The locale tests are not compatible with musl toolchains,
so disable them.

Use latest git version from upstream for other bugfixes.

Fixes:
 http://autobuild.buildroot.net/results/74e/74e7add310797772bc51f9ea9847408a8b05d643/

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

qt5xmlpatterns: install examples

Signed-off-by: Matt Kraai <kraai@ftbfs.org>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

DEVELOPERS: update my entry list

Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/openvpn: bump version to 2.3.14

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

efibootmgr: depends on wchar

After commit 3ae07b4746 recently, efibootmgr now selects
BR2_PACKAGE_GETTEXT if the toolchain requires it.
gettext depends on wchar, so this dependency should be propagated as
well.
menuconfig currently complains loudly if you select efibootmgr, with an
error such as:

  warning: (... && BR2_PACKAGE_EFIBOOTMGR ... && ) selects
  BR2_PACKAGE_GETTEXT which has unmet direct dependencies
  (BR2_USE_WCHAR)

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

efivar: introduce BR2_PACKAGE_EFIVAR_ARCH_SUPPORTS

efivar only makes sense on platforms that support UEFI.
UEFI is only supported by some architectures at the moment, being mostly
employed on platforms such as x86, x86_64 and aarch64. Some other
platforms such as MIPS and PowerPC may have some unofficial UEFI
support. UEFI is also limited to little endian architectures.

efivar was being supported in Buildroot without architecture
restrictions so far, however this has led to the creation of a number of
hacks in the recipes, mostly for architectures that are not supported by
UEFI.
In order to avoid spending more time to debug these failures and
maintaining more hacks for unsupported architectures, efivar can be
restricted to that platforms where it makes sense and where it is more
likely to receive some testing and actual usage.

The existing hacks for the now unsupported architectures are removed,
and the dependency is propagated to efibootmgr as it depends on efivar.

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

efivar: fix comment after uClibc compatibility patch

uClibc support was recently added to efivar through a small
compatibility patch.
This commit updates a comment in the efivar recipe to reflect this, as
we no longer have glibc as the only supported C library.

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

package/freeswitch: bump version to 1.6.13

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

linux: generate KBUILD_BUILD_TIMESTAMP date whith LC_ALL=C

Fix kernel reproducible build if a non-C locale is used on the host
system.

When building the Linux kernel, scripts/gen_initramfs_list.sh does 'date
-d"$KBUILD_BUILD_TIMESTAMP" +%s'. In linux.mk, Buildroot sets
KBUILD_BUILD_TIMESTAMP to "$(shell date -d @$(SOURCE_DATE_EPOCH))".

For example, if LANG=fr_FR.UTF-8 is defined in the host system, it does
not work:

- LC_ALL=C date -d"$(LC_ALL=C date)" : ok
- LC_ALL=C date -d"$(LC_ALL=fr_FR.UTF-8 date)" : error

LANG/LC_ALL variables exported in the main Makefiles are not passed in
the $(shell ...) sub-shells.

Signed-off-by: Jean-Baptiste Trédez <jean-baptiste.tredez@basystemes.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

qt5connectivity: install examples

Signed-off-by: Matt Kraai <kraai@ftbfs.org>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

madplay: fix build on noMMU platforms

A recent change in uClibc-ng removed the MADV_* definitions for
madvise(), but kept the madvise() function itself. This defeats the
logic used in madplay: it checks if madvise() is available, and if it
is, uses it and assumes the MADV_* definitions exist.

This inconsistency has been reported to upstream uClibc-ng, but in the
mean time, we can simply tell madplay to not use madvise(), which is
anyway useless on noMMU platforms.

Fixes:

  http://autobuild.buildroot.net/results/3291554ea013e5f4a8f3447e10e664dffa8b131b/

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

uhttpd: add patch to remove -Werror

Fixes:

  http://autobuild.buildroot.net/results/febbd094090e14fe7a79e6f691c290075da4fd59/

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

lttng-tools: fix build on musl

Add patch to add a missing <pthread.h> include to fix the build on musl.

Fixes:

  http://autobuild.buildroot.net/results/ce0b438ca4dcf7ccef95c7050305d8cd58fbb24a/

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

pru-software-support: fix legal info again

Commit 00810e0ea3976c5b26a7fd264870f15e17a2eaa2 still didn't fix the
license information properly, so this is a new commit (properly tested
this time) fixing the license information.

Fixes:

  http://autobuild.buildroot.net/results/ece1407b9f32472fb55823bb554dc026546cd26d/

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

php: disable FPM interface on Sparc

Building fails with "Sparc v8 and predecessors are not and will not be
supported (see bug report 53310)", so we disable FPM on Sparc.

Fixes:

  http://autobuild.buildroot.net/results/6a86f6cf07f6b6dccd7bfaab6d7682f9faf3527b/

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

pkg-autotools: search in the entire source directory for ltmain.sh

Fixes static linking of berkeleydb package, where ltmain.sh
is not in the sub-directory that contains configure.
Always search the complete source directory.

Fixes:
 http://autobuild.buildroot.net/results/f0a96f671644d9f9efcf245b354affdc84f7d7da

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

package/easydbus: new package

Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Reviewed-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

cmake: adjust dependencies added for libuv

Commit 1ffcf364b6e9894a876dc581a090f87685945412 updated cmake to 3.7.0,
which requires selecting the libuv package. At the time, the libuv
package only depended on BR2_TOOLCHAIN_HAS_THREADS. However, later on,
it was changed in master to depend on BR2_TOOLCHAIN_HAS_THREADS_NPTL, a
change which was not taken into account in the cmake 3.7.0 bump that was
merged in the next branch.

Due to this, builds of cmake is attempted on architectures that don't
provide NPTL thread support, causing a build failure. This commit fixes
that by adjusting the dependency.

Fixes:

  http://autobuild.buildroot.net/results/16a5e1cbb57c0124537c4f3dc0807ba1eaa975ec/

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

package/ffmpeg: bump version to 3.2.2

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

gcc: fix powerpc64le arch config name

Buildroot references powerpc64 little endian as BR2_powerpc64le and not
BR2_powerpc64el.

Signed-off-by: Erico Nunes <nunes.erico@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

toolchain-external: disallow sourcery codebench ARM toolchain for ARMv8 cores

This toolchain uses GCC 4.8.x, which doesn't support the ARMv8 cores.

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

gcc: disallow 4.9.x for ARM cortex-a72

a72 support was only added in the 5.1 cycle.

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

gcc: disallow 4.8.x for ARMv8 cores

Notice: A53/A57 were in fact supported in aarch64 mode in 4.8 (in aarch32
mode only from 4.9), but it doesn't handle --with-abi, and as there is
unlikely to be any aarch64 based legacy projects unwilling to use a newer
GCC version it is simpler to disallow it for all modes.

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

Makefile: move SED definition into the main Makefile

Since commit f71a621d91ec27f175fc84012962f88b1107305f, we are using the
SED variable in the main Makefile. However, this variable is only
defined in package/Makefile.in, which gets included only when a
configuration is defined.

This means that, if you do:

 $ make menuconfig savedefconfig

without a configuration defined, it fails with:

/bin/bash: /BR2_DEFCONFIG=/d: No such file or directory
Makefile:898: recipe for target 'savedefconfig' failed
make[1]: *** [savedefconfig] Error 127

This issue affects users of the "buildroot-submodule" project, which
does menuconfig+savedefconfig automatically. They worked around this
issue in commit
https://github.com/Openwide-Ingenierie/buildroot-submodule/commit/d12676b608a58529c6b551aa176464166a200428,
but really "make menuconfig savedefconfig" should work out of the box.

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

mbedtls: bump to version 2.4.0

This change also fixes
http://autobuild.buildroot.net/results/27b/27b35d03b576392955e21498ac1e62590ef5f5db/build-end.log

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

ffmpeg: fix build for ffplay

Upstream has dropped SDL support for ffplay in favor of SDL2.
This results in silently not building ffplay even if it is selected
in Buildroot config.

[Peter: propagate !static dependency from sdl2]
Signed-off-by: Phil Eichinger <phil@zankapfel.net>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/shairport-sync: bump version to 2.8.6

Signed-off-by: Jörg Krause <joerg.krause@embedded.rocks>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

sysklogd: fix build on musl

This commit add a stack of small patches that make sysklogd build fine
with the musl C library. Build with uClibc and glibc has been tested
with those patches applied as well.

The first patch is slightly rework (better description and capital
letter to the title) in preparation for upstream submission.

Fixes:

  http://autobuild.buildroot.net/results/8fa2bf73f983330884bce2e5ac31e01dee112ba9/

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

raspberrypi-usbboot: new package

This new package currently installs the "rpiboot" utility, which is
needed to access via USB mass storage the built-in eMMC of Raspberry Pi
compute modules.

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

DEVELOPERS: remove Tom Sparks

Tom privately notified that he is not able to support the packages he
originally added in Buildroot, so this commit removes him from the
DEVELOPERS file.

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

toolchain-external-linaro-{arm,armeb}: allow on ARMv8

The Linaro toolchains are currently only available on ARMv7-A, but can
in fact also be used to generate 32 bits code for ARMv8 platforms. This
commit therefore adjusts their architecture dependency.

Example, a 32 bits ARM build produces a 32 bits busybox binary:

$ file output/target/bin/busybox
output/target/bin/busybox: setuid ELF 32-bit LSB executable, ARM, EABI5 version 1 (SYSV), dynamically linked, interpreter /lib/ld-linux-armhf.so.3, for GNU/Linux 2.6.32, BuildID[sha1]=16a7a70eb9cac08759e52a260478b9c287f59238, stripped

Which was built for Cortex-A72:

$ ./output/host/usr/bin/arm-linux-gnueabihf-readelf -A output/target/bin/busybox
Attribute Section: aeabi
File Attributes
  Tag_CPU_name: "Cortex-A72"
  Tag_CPU_arch: v8
  Tag_CPU_arch_profile: Application
  Tag_ARM_ISA_use: Yes
  Tag_THUMB_ISA_use: Thumb-2
  Tag_FP_arch: FP for ARMv8
  Tag_ABI_PCS_wchar_t: 4
  Tag_ABI_FP_rounding: Needed
  Tag_ABI_FP_denormal: Needed
  Tag_ABI_FP_exceptions: Needed
  Tag_ABI_FP_number_model: IEEE 754
  Tag_ABI_align_needed: 8-byte
  Tag_ABI_align_preserved: 8-byte, except leaf SP
  Tag_ABI_enum_size: int
  Tag_ABI_VFP_args: VFP registers
  Tag_CPU_unaligned_access: v6
  Tag_MPextension_use: Allowed
  Tag_Virtualization_use: TrustZone and Virtualization Extensions

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

arch/Config.in.arm: add Cortex-A57 and Cortex-A72

Add two popular ARM64 cores to the list of supported cores: Cortex-A57
and Cortex-A72.

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

arch/Config.in.arm: Add Cortex-A53 CPU

Adds the Cortex-A53 CPU to the target architecture variant choice. This
sets the toolchain to use Cortex-A53 as the target. The effect is that
various Cortex-A53 tunings are enabled for the compilation of packages.

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

arch/Config.in.arm: specify ABI for ARM64

There's currently only one widely supported ABI for ARM64, called lp64,
so we define BR2_GCC_TARGET_ABI to the appropriate value.

Note that there is another ABI for ARM64 being worked on, ilp32, but its
support is not fully upstream in the kernel, so we're not adding support
for it for the moment.

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

arch/Config.in.arm: add FPU related options for ARMv8 cores

The ARMv8 cores have a mandatory FPU unit called FP-ARMv8, so we:

 - add a new hidden Config.in option for the availability of this
   unit (BR2_ARM_CPU_HAS_FP_ARMV8)

 - allow the selection of a possible choice in the "Floating point
   strategy", and add two new choices: BR2_ARM_FPU_FP_ARMV8 and
   BR2_ARM_FPU_NEON_FP_ARMV8.

 - specify the -mfpu values for BR2_ARM_FPU_FP_ARMV8 and
   BR2_ARM_FPU_NEON_FP_ARMV8 cases, when used on ARM 32 bits (-mfpu
   doesn't exist on ARM64, instead -mcpu modifiers are used, so they
   will be added on a per-core basis).

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
[yann.morin.1998@free.fr: drop the FP strategy dependency]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

arch/arm: drop legacy dependency for floating point strategy

The floating point strategy currently depends on EABI || EABIHF. The
reason was that, wayback when we also supported OABI, we only exposed FP
for EABI or EABIHF, and hide it for OABI, which did not support FP.

It's been a while now that we do not support OABI, but the dependency
stuck all along.

Remove it as it is no longer needed, and is always true.

However, the choice is empty for AArch64, as we still have no entry for
their floating point strategy yet.

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>

arch/Config.in.arm: add BR2_ARM_CPU_ARMV8

In order to prepare the addition of ARM64 cores, add the blind
BR2_ARM_CPU_ARMV8 option.

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

arch/Config.in.arm: prepare addition of 64 bits cores

Until now the "Target Architecture Variant" choice was not visible on
AArch64. In order to prepare the addition of the 64 bits core to this
choice, this commit adds a "depends on !BR2_ARCH_IS_64" dependency to
all currently supported cores (that are 32 bits only).

Following this commit, the "Target Architecture Variant" choice appears
on AArch64, but is for now empty.

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

arch: merge Config.in.aarch64 into Config.in.arm

The 64 bits ARM processors are capable of running 32 bits ARM code, and
some platforms are indeed using this capability. Due to this, if we were
to keep the separation between Config.in.aarch64 and Config.in.arm, we
would have to duplicate the definition of all 64-bits capable ARM cores
into both files.

Instead of going down this route, let's take the same route as the x86
one: a single Config.in.x86 file, used for both x86 32 bits and x86 64
bits, with the appropriate logic to only show the relevant cores
depending on which architecture is selected.

In order to do this, we:

 - Make the "ARM instruction set" choice only visible on ARM 32 bits,
   since we currently don't support ARM vs. Thumb on AArch64.

 - Add the relevant values for the BR2_ARCH option.

 - Add the relevant values for the BR2_ENDIAN option.

 - Make the "aapcs-linux" BR2_GCC_TARGET_ABI value only used on ARM 32
   bits, since this ABI doesn't mean anything on AArch64.

 - Make the BR2_GCC_TARGET_FPU option depends on ARM 32 bits, since
   there is no -mfpu option on AArch64.

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

vim: be more careful when removing the documentation

The current VIM_REMOVE_DOCS hook removes all .txt files from
/usr/share/vim. Unfortunately, this also removes the rgb.txt file,
which is needed at runtime for vim, as reported in bug #9466.

This commit changes VIM_REMOVE_DOCS to remove only
/usr/share/vim/vim*/doc/. Size-wise, it's equivalent because:

 - We are no longer removing a few README.txt in other directories,
   taking more space.

 - We are now removing the /usr/share/vim/vim*/doc/ folder entirely,
   which contained a few files not named *.txt

So overall, the size of /usr/share/vim/ before and after this patch is
still 11MB.

Fixes bug #9466.

Reported-by: Mateusz Furdyna <sir.ferdek@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

pkg-autotools: generic configure fix for powerpc64

Many (100+) packages supported by buildroot contain old configure
scripts (or build them from old versions of autotools) that are unable
to determine how to link shared libraries on powerpc64 and
powerpc64le. This causes that test to erroneously fail on toolchains
that are not "bi-endian" (which is the case for toolchains built by
buildroot), which causes configure to build static libraries instead
of dynamic ones. Although these builds succeed, they tend to cause
linker failures in binaries later linked against them.

Because affected configure files can be discovered automatically, this
patch introduces a hook (enabled only when building for powerpc64 and
powerpc64le) that uses a script to scan and fix each package.

Signed-off-by: Sam Bobroff <sam.bobroff@au1.ibm.com>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

pru-software-support: fix licensing information

Fix the path to the license file, and adjust the licensing information
according to the contents of this file.

Fixes:

  http://autobuild.buildroot.net/results/213cdda78b0a4762ee4fecc0f178df1225062c6f/

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

package/liquid-dsp: fix build failure on powerpc64le

Commit 7d435d8510db69ed2f9108abaa52479bce127b72 fixed building on
powerpc and powerpc64. This extends that fix to cover powerpc64le,
which suffers from the same problem.

Fixes:
http://autobuild.buildroot.net/results/45c4651a833e4667d4ef68e619cf4b5749a288da

Signed-off-by: Sam Bobroff <sam.bobroff@au1.ibm.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

nginx-dav-ext: depends on BR2_PACKAGE_NGINX_HTTP_DAV_MODULE

This package adds missing WebDAB commands to NGINX. So, it makes sense
iff Nginx is compiled with the WebDAV support.

Fixes:

  http://autobuild.buildroot.net/results/6e997faf553dc70d0aa1ee811fd669650349cba2/

Signed-off-by: Johan Oudinet <johan.oudinet@gmail.com>
[Thomas: fix indentation, add autobuilder reference.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

nginx-naxsi: depends on BR2_PACKAGE_NGINX_HTTP

This package adds XSS and SQLi protection to Nginx HTTP server.
So, it makes sense iff the HTTP part of nginx is enabled.

Fixes:

  http://autobuild.buildroot.net/results/1df67325f11a6f1c88d0c9cc5f4feab0b57bc2a6/

Signed-off-by: Johan Oudinet <johan.oudinet@gmail.com>
[Thomas: fix indentation, add autobuilder reference.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

efivar: not available for static builds

efivar uses dlfcn.h which is not available in static builds
configuration. Also propagate dependency to efibootmgr. This commit
also does s/requires/needs/ in comment while at it.

Fixes:
  http://autobuild.buildroot.net/results/ab3/ab32ccacfac2e65e1fcb307058d06c3d0122979d/

Signed-off-by: Rahul Bedarkar <rahul.bedarkar@imgtec.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

gstreamer1: fix compile for microblaze and xtensa architectures

Fix unaligned access support detection introduced to gstreamer
by commit 'gstconfig.h: Detect unaligned access support at compile-time' ([1])
by assuming no unaligned support for unknown architectures.

Fixes [2], [3]:

  ../gst/gstconfig.h:112:4: error: #error "Could not detect architecture; don't know whether it supports unaligned access! Please file a bug."
  #  error "Could not detect architecture; don't know whether it supports unaligned access! Please file a bug."

[1] https://cgit.freedesktop.org/gstreamer/gstreamer/commit/?h=1.10&id=6ef601367e0f5adb7a8b02cec94dec04fccf701a
[2] http://autobuild.buildroot.net/results/c1cd60d3f44f2938e5f42453d034ea070898f837
[3] http://autobuild.buildroot.net/results/c3a6662f0acfda179a58cdad9165f36826229826

Signed-off-by: Peter Seiderer <ps.report@gmx.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

chrony: bump version to 2.4.1

Fixes a crash issue with smoothtime directive.

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

ti-cgt-pru: fix typo in comment

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

package/openpowerlink: fix build with Linux kernel 4.7+

dev->trans_start was replaced by netif_trans_update helper in kernel
4.7 by commit 860e9538a9482bb84589f7d0718a7e6d0a944d58.

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

package/openpowerlink: bump to v2.5.0

 - Rebase existing patches
 - Use a "menu" instead of "choice" to select Ethernet Powerlink
   Drivers since we can now build several kernel modules at once.
 - Select the Intel 82573 driver by default.
 - Disable library with simulation interface.
 - Disable zynq/FPGA (PCIe) interface otherwise the kernelpcp library
   is always build.

Signed-off-by: Romain Naour <romain.naour@smile.fr>
[Thomas: minor Config.in tweaks.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

qextserialport: remove redundant patch 0001-gui

The same thing was done in upstream commit 220d7a2f "QtGui module only
needed by Windows", though it was done differently upstream.

Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

DEVELOPERS: add developer for nginx-dav-ext

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

nginx: group the external modules enabling logic

This commit moves the logic that allows to enable the naxsi external
module below the "external modules" comment, which was already used for
the upload and dav-ext modules.

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

nginx-dav-ext: new package

Nginx built-in support for webdav is missing support for two commands:
PROPFIND and OPTIONS. This commit adds a new package that provides an
external nginx module with improved webdav support.

Signed-off-by: Johan Oudinet <johan.oudinet@gmail.com>
[Thomas:
 - Remove the BR2_PACKAGE_NGINX_HTTP_DAV_EXT_MODULE sub-option of the
   nginx package. The BR2_PACKAGE_NGINX_DAV_EXT option is sufficient.
 - Move the nginx.mk code together with another external module being
   enabled, nginx-upload.
 - Add LICENSE and LICENSE_FILES variables.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

erlang: odbc support requires thread support

Fixes:

  http://autobuild.buildroot.net/results/cba60679fa916128796a7b4131ed50403bc74798/

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