support/scripts/check-package: new script

Create the infra to check the style of new packages before submitting.
The overall function of the script is described inside a txt file.
It is designed to process the actual files and NOT the patch files
generated by git format-patch.

Also add the first check function, to warn if a file (Config.*, *.mk,
*.hash, *.patch) has no newline at the last line of the file, see [1].

Basic usage for simple packages:
support/scripts/check-package -vvv package/newpackage/*

Basic usage for packages with subdirs:
support/scripts/check-package -vvv $(find package/newpackage/ -type f)

See "checkpackage" in [2].

[1] http://patchwork.ozlabs.org/patch/631129/
[2] http://elinux.org/Buildroot#Todo_list

Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

syslinux: build with the target toolchain

Until now, the host toolchain was used to build syslinux, as it was
not possible to build a 32-bit syslinux with a x86-64 toolchain.

However, syslinux requires gnu-efi, and gnu-efi is built using the
target toolchain. Mixing different toolchains doesn't work well, so
this commit changes the syslinux package to use the target toolchain
for syslinux as well. This is made possible by patches
0003-Fix-ldlinux.elf-Not-enough-room-for-program-headers-.patch and
0004-memdisk-Force-ld-output-format-to-32-bits.patch.

Since syslinux also contains some utilities that have to run on the
host, those have to continue being built with the host toolchain,
which requires patch 0005-utils-Use-the-host-toolchain-to-build.patch.

Patch 0006-lzo-Use-the-host-toolchain-for-prepcore.patch is about
building prepcore, another utility with the host toolchain as it is
required at build-time.

This was tested using a Buildroot's built x86_64 toolchain, and
checked that the output binaries are 32-bits. It was tested as well if
they actually boot on hardware.

Signed-off-by: Benoît Allard <benoit.allard@greenbone.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

configs/beaglebone_qt5: increase size of ext4 filesystem image

In commit c6bca8cef0310bc649240b451989457ce94a8358, the
auto-calculation logic of the ext4 filesystem size was removed. The
default size of 60MB is too small for the beaglebone_qt5_defconfig,
should it should be increased.

Signed-off-by: Michael Trimarchi <michael@amarulasolutions.com>
Suggested-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

autofs: fix Git formatting of patch

The patch 0001-include-linux-nfs.h-directly-in-rpc_sub.patch was Git
formatted, except that a Upstream status statement was added above the
patch, which makes it unapplicable by "git am". So fix this by putting
the Upstream status statement where it should be.

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

pkg-python: ignore user site-packages directory

According to PEP 370 Python will also search for the packages in the
user site-packages directory. This can affect build reproducibility.

The solution is to use PYTHONNOUSERSITE=1 for all Python packages,
i.e. both host and target variants.

Fixes bug #9791.

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

gnupg: enable support for SHA-256 and SHA-512 hashes

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

uboot: qstrip BR2_TARGET_UBOOT_CUSTOM_DTS_PATH

We need to avoid failing of copy of custom dts using cp -f <>
command. So, just qstrip the variable as is done for other
configuration options.

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

qt5declarative: fix examples compile without qt widgets module

Fixes:

  .../qt5declarative-5.8.0/examples/quick/demos/photoviewer/photoviewer.pro
  Project ERROR: Unknown module(s) in QT: widgets

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

batman-adv: bump version to 2017.0.1

Signed-off-by: Matthias Fritzsche <txt.file@txtfile.eu>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/linux-headers: bump 3.2.x and 4.{4, 9, 10}.x series

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

linux: bump default version to 4.10.9

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

package/libpciaccess: bump version to 0.13.5

Add additional md5, sha1 & sha256 hashes according to buildroot docs:
http://buildroot.uclibc.org/downloads/manual/manual.html#adding-packages-hash

> If upstream provides more than one type of hash (e.g. sha1 and sha512),
> then it is best to add all those hashes in the .hash file.

Removed patch 0001, applied upstream:
https://cgit.freedesktop.org/xorg/lib/libpciaccess/commit/?id=6bd2f7f92eae713663f4e13f6e2cb23526607b8c

Renumbered remaining patch.

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

package/libpthread-stubs: bump version to 0.4

libdrm 2.4.77 needs it since
https://cgit.freedesktop.org/mesa/drm/commit/?id=63120172668610cebfd01e89f67e73037037dfd9

Fixes:
http://autobuild.buildroot.net/results/732/732652f34f4865462007a3ebeb40d6642fde028b/
http://autobuild.buildroot.net/results/b5d/b5d7667737c92de5d19313ec6bb0ec9ce81707ea/

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

mosquitto: enable WITH_ADNS for glibc builds

WITH_ADNS option has been added in version 1.4.11

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

mosquitto: bump to version 1.4.11

- This version requires a patch (sent upstream) to remove -lanl from
all Linux builds as this library is only needed for adns support
 - sha512 must be computed locally as eclipse.org does not give it for
this version

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

package/libdrm: bump version to 2.4.78

Add additional md5, sha1 & sha256 hashes according to buildroot docs:
http://buildroot.uclibc.org/downloads/manual/manual.html#adding-packages-hash

> If upstream provides more than one type of hash (e.g. sha1 and sha512),
> then it is best to add all those hashes in the .hash file.

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

DEVELOPERS: update email address of Rahul Bedarkar

I'm no longer with Imagination Technologies. Update email address to an
active one.

Signed-off-by: Rahul Bedarkar <rahulbedarkar89@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

ethtool: bump to version 4.10

Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

ethtool: update homepage link

Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

python-pillow: bump version to 4.1.0

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

lft: bump version to 3.79

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

icu: legal-info: refer to the real license file

License.html now just contains the string:

The ICU license is now in plain text format, see <a href="./LICENSE">LICENSE</a>.
Update links and software appropriately.

So refer directly to that file instead.

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

freetype: legal-info: include docs/LICENSE.TXT

LICENSE.TXT gives an overview and explains in detail that freetype is dual
licensed under the FTL and GPLv2+, so also include it in the license files.

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

jemalloc: bump version to 4.5.0

Signed-off-by: Eric Le Bihan <eric.le.bihan.dev@free.fr>
[Thomas: use release tarball, which avoids the need for autoreconf.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

gst1-plugins-base: remove legacy freetype handling

In 2010 commit 32d319e6f "gst-plugins-base: ensure <stdint.h> is used"
introduced a typo (missing backslash) that made the code ineffective.

In 2013 commit f8e7fdcd3 "gst1-plugins-base: add gstreamer1 base
plugins" copied the code.

It can be confirmed by looking at the output of:
$ make printvars | grep '^GST1_PLUGINS_BASE_CONF_ENV\|^FT2_CONFIG'
FT2_CONFIG=/bin/false ac_cv_header_stdint_t="stdint.h"
GST1_PLUGINS_BASE_CONF_ENV=

Direct use of freetype was dropped in version 1.7.2 by upstream [1], so
remove the code instead of fixing it.

Found using [2]:
check-package --include-only Indent $(find * -type f)
and manually removed.

[1] https://cgit.freedesktop.org/gstreamer/gst-plugins-base/commit/?id=183610c035dd6955c9b3540b940aec50474af031
[2] http://patchwork.ozlabs.org/patch/729669/

Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

tesseract-ocr: fix typo in variable name

TESSERACT_OCF_CONF_OPTS
            ^

Found using [1]:
check-package --include-only TypoInPackageVariable $(find * -type f)
and manually fixed.

[1] http://patchwork.ozlabs.org/patch/729672/

Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

python-pylibftdi: use package name for .mk file

Found using [1]:
check-package --include-only TypoInPackageVariable $(find * -type f)
and manually fixed.

[1] http://patchwork.ozlabs.org/patch/729672/

Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

python-pylibftdi: check hash

Currently the hash is not checked, returning this message:
WARNING: no hash file for pylibftdi-0.15.0.tar.gz

Rename the .hash file to use the package name.

Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package: remove trailing backslash

Occurrences were searched using [1]:
check-package --include-only TrailingBackslash $(find * -type f)
and manually removed.

[1] http://patchwork.ozlabs.org/patch/729669/

Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package: add missing empty line after header

Occurrences were searched using [1]:
check-package --include-only PackageHeader $(find * -type f)
and manually fixed to comply to [2].

[1] http://patchwork.ozlabs.org/patch/729669/
[2] http://nightly.buildroot.org/#writing-rules-mk

Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package: normalize separator size to 80

Occurrences were searched using [1]:
check-package --include-only PackageHeader $(find * -type f)
and manually fixed to comply to [2].

[1] http://patchwork.ozlabs.org/patch/729669/
[2] http://nightly.buildroot.org/#writing-rules-mk

Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package: remove consecutive empty lines

Occurrences were searched using [1]:
check-package --include-only ConsecutiveEmptyLines $(find * -type f)
and manually removed.

[1] http://patchwork.ozlabs.org/patch/729666/

Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package: remove empty line at end of file

Occurrences were searched using [1]:
check-package --include-only EmptyLastLine $(find * -type f)
and manually removed.

[1] http://patchwork.ozlabs.org/patch/729666/

Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package: remove trailing whitespace

Occurrences were searched using [1]:
check-package --include-only TrailingSpace $(find * -type f)
and manually removed.

[1] http://patchwork.ozlabs.org/patch/729666/

Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

toolchain-external: fix handling of ld.so

Commit ba6bac138331cea449592f877e558f84176a70bf made a change in copying of
the dynamic loader, with the goal of reducing toolchain-specific fixups.
Any ld*.so file found in the toolchain's lib directory would be copied to
the staging/lib directory.

For the toolchains that previously needed fixup, this new behavior is fine.
The reason they needed fixup was that the normal copy action did not include
any dynamic loader.

However, for certain other toolchains this new behavior actually breaks
things: regardless of ARCH_LIB_DIR, which may be lib64 instead of lib, the
dynamic loader from lib is copied _over_ any previously correct dynamic
loader.

This has been witnessed with the CodeSourcery x86_64 and
CodeSourcery MIPS64 toolchains. In both cases, a 32-bit dynamic loader was
copied to staging/lib, while a 64-bit version was expected.

Fix the problem by only performing this explicit dynamic loader copy if no
dynamic loader is found in staging/lib.

Fixes http://autobuild.buildroot.net/results/8bf/8bffe54032aad9cc710a22411ef3bff4a2c93e55/

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

package/libdrm: bump version to 2.4.77

Replace sha256 hash by sha512.

Removed patch 0003, applied upstream:
https://cgit.freedesktop.org/mesa/drm/commit/?id=6796859fcce7b9d7157f066a8b0dac3ede6438dd

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

package/libvpx: bump version to 1.6.1

Rebased patch 0001.

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

linux: Add CIP SLTS selection option

This patch adds an easy way to select the CIP project SLTS
kernel within the kernel menu.

Signed-off-by: Angelo Compagnucci <angelo.compagnucci@gmail.com>
[Thomas: remove BR2_CIP_KERNEL_REPO_URL option.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

python-pudb: bump to version 2017.1.2

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

fmt: new package

Signed-off-by: Semyon Kolganov <semenak94@mail.ru>
[Thomas:
 - add entry to DEVELOPERS file
 - add missing wchar dependency
 - minor tweaks.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

configs: bump linux kernel to 4.10.8 in snps_archs38_vdk_defconfig

With this commit we update snps_archs38_vdk_defconfig with the
following:
  - bump linux kernel version to 4.10.8
  - set up host linux headers to 4.10

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

support/test-pkg: print number of toolchains and progress

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

docs/manual: document the test-pkg script

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

docker-engine: bump to version 17.04.0-ce

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

docker-containerd: bump to version 422e31ce907fd9c3833a38d7b8fdd023e5a76e73

This is a runtime dependency of docker-engine in version 17.04.0-ce

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

runc: bump to version 9c2d8d184e5da67c95d601382adf14862e4f2228

This is a runtime dependency of docker-engine in version 17.04.0-ce

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

libnspr: use __nios2__ instead of nios2

Our patch adding nios2 support to libnspr uses the built-in compiler
define "nios2". However, this doesn't work with C++11, where only the
__nios2__ define is available. Since __nios2__ is always available,
use that instead:

$ ./output/host/usr/bin/nios2-linux-gcc -dM -E - < /dev/null | grep -E "( nios2 | __nios2__ )"

$ ./output/host/usr/bin/nios2-linux-gcc -std=c++11 -x c++ -dM -E - < /dev/null | grep -E "( nios2 | __nios2__ )"

Patch 0001-nios2.patch is therefore changed to use __nios2__ (the rest
of the change noise is due to using quilt to format the patch). Patch
0002-microblaze.patch is simply updated to apply correctly on top of
the modified 0001-nios2.patch.

This fixes the build of the poppler library on nios2. It is built with
-std=c++11, and includes nspr headers (through nss), causing a build
issue.

Fixes:

  http://autobuild.buildroot.net/results/9fee58076157d814616fa0da51afde8da21a8973/

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

libnspr: fix Microblaze patch

Due to a mistake, the Microblaze patch was adding another #elif
defined(nios2), which doesn't make any sense. This commit gets rid of
it.

The rest of the noise in the change is due to the use of quilt to
generate the patches.

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

docs/website: update for 2017.02.1

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

Update for 2017.02.1

[Peter: drop Makefile changes]
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 476067a78d4fd1ddb141a8cccc63cf3e75cbe7a3)

arp-scan: new package

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

Add out-of-source U-Boot device tree support

Similaly to Linux, this patch adds the ability to copy in and build
out-of-source device tree sources during a U-Boot build.

Signed-off-by: James Balean <james@balean.com.au>
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

beaglebone: add DT for BeagleBone Green

Signed-off-by: Matthias Fritzsche <txt.file@txtfile.eu>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

syslinux: use Git formatted patches

In preparation for the addition of more patches to the syslinux
package, reformat the two existing patches as proper Git formatted
patches.

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

toolchain: copy_toolchain_lib_root: copy symlinks instead of recreating them

copy_toolchain_lib_root handles symlinks by recreating them, disregarding
the original destination and assuming the destination is in the same
directory as the link itself.
When a library link points to the real library file in another directory,
for example:
    usr/lib/octeon2/libcrypt.so -> ../../../lib32/octeon2/libcrypt.so.1
then the link created by copy_toolchain_lib_root is broken.

It is more robust to copy the symlink to keep the destination intact. The
destination path should be present, possibly through other symbolic links.

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

toolchain: copy_toolchain_lib_root: clarify input parameter

The input to copy_toolchain_lib_root is not one library, not a list of
libraries, but a library name pattern with glob wildcards.
This pattern is then passed to 'find' to get the actual list of libraries
matching the pattern. Reflect this using an appropriate variable name.

Note: if the root of the buildroot tree contains a file matching one of
these library patterns, the copying of libraries from staging to target will
not be correct. It is not impossible to fix that, e.g. using 'set -f', but
maybe it's not worth it.

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

toolchain: copy_toolchain_lib_root: clarify logic

copy_toolchain_lib_root has slightly different logic depending on the type
of library object: file or link. All actions related to links are not
relevant in case you are working with a file. Hence, try to increase clarity
by not executing unnecessary lines in the 'file' case.

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

toolchain: copy_toolchain_lib_root: remove unused variable LIBDIR

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

toolchain-external: simplify previously-broken symbolic links

A previous commit rewrote broken symbolic links in staging, caused by a
non-singular ARCH_LIB_DIR. In this case, the symbolic links are typically
using one or more intermediate directory symlinks, which can be simplified
using the newly introduced simplify_symlink helper.

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

toolchain helpers: introduce simplify_symlink

The external toolchain logic flattens the directory layout in the staging
directory.  Regardless of the number of levels present in the extracted
toolchain, libraries are always copied to lib/ and usr/lib/, and directory
symbolic links are provided to make the original paths available as well.

Due to this, the same library may be reachable through a number of paths:
one path without any symbolic link, and one or more paths using directory
symlinks.

Using a direct path in a symlink destination is generally preferred because
it is clearer, but it is also more robust against accidental removal of an
intermediate directory symlink.

Introduce a helper function to simplify a symlink's destination to such a
direct path. This function will be used in a subsequent patch.

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

toolchain-external: cover multilib toolchains with lib/<variant> layout

The toolchain from the Cavium Octeon SDK has a sysroot layout as follows:

./lib32
./lib32/octeon2
./lib32-fp
./lib64
./lib64/octeon2
./lib64-fp
./usr
./usr/lib
./usr/lib32
./usr/lib32/octeon2
./usr/lib32-fp
./usr/lib64
./usr/lib64/octeon2
./usr/lib64-fp
./usr/bin
./usr/bin32
./usr/bin32-fp
./usr/bin64-fp
./usr/libexec
./usr/libexec32
./usr/libexec32-fp
./usr/libexec64-fp
./usr/sbin
./usr/sbin32
./usr/sbin32-fp
./usr/sbin64-fp
./usr/include
./usr/share
./sbin
./sbin32
./sbin32-fp
./sbin64-fp
./etc
./var

with the following selections:
- lib64          : default
- lib64/octeon2  : -march=octeon2
- lib64-fp       : -march=octeon3
- lib32          : -mabi=n32
- lib32/octeon2  : -mabi=n32 -march=octeon2
- lib32-fp       : -mabi=n32 -march=octeon3

In case of '-mabi=n32 -march=octeon2' (but same is true for n64+octeon2)the
original Buildroot toolchain logic would copy both the libraries in
lib32 as the subdirectory lib32/octeon2, which means that every library is
installed twice (but only one of each is really needed).

While ARCH_LIB_DIR is determined by the location of libc.a, which in this
case is effectively:
    <sysroot>/usr/lib32/octeon2/libc.a
the variable only retains 'lib32' and not 'lib32/octeon2' as expected.

To make Buildroot cope with this style of toolchain layout, we need to adapt
the calculation of ARCH_LIB_DIR to also include the second part.
This, in turn, means that ARCH_LIB_DIR is no longer guaranteed to be a
singular path component, resulting in some additional changes.

Certain older Linaro toolchains actually had the same layout. Libraries were
located in lib/<tuple> rather than lib directly. Previously, this was
handled by adding a toolchain-specific fixup that creates a symlink
lib/<tuple> -> lib, but with this patch this would no longer be needed.
Note that one difference with the Octeon case is that these Linaro
toolchains are not actually multilib, i.e. there is just one location with
the libraries and thus there is no problem with duplicated libraries.

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

toolchain helpers: introduce function relpath_prefix

The helper function copy_toolchain_sysroot has some logic to transform a
path into a number of '../' components based on the depth of that path.

As this same logic will be needed in another place in a subsequent patch,
extract it into a separate helper relpath_prefix.

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

toolchain-external: handle ld.so fixups centrally

Normally, the Buildroot toolchain logic copies all required libraries from
the external toolchain to the staging directory, including the dynamic
loader ld-*.so.
There are cases, however, where the dynamic loader is _not_ automatically
copied to staging. This happens when the dynamic loader is not inside
ARCH_LIB_DIR itself (e.g. lib64), but instead resides in 'lib' (assume, of
course, that ARCH_LIB_DIR != 'lib').

Currently, this is fixed in a toolchain-specific fixup, e.g. by recreating a
missing symlink or copying over a missing file. Such toolchain specific
fixups are not very nice.

Moreover, in a subsequent patch, the value of ARCH_LIB_DIR changes for some
toolchains, causing them to have the same problem of a missing dynamic
loader. This used to be the case for older Linaro toolchains with libraries
in 'lib/<tuple>': Buildroot used to set ARCH_LIB_DIR=lib but the mentioned
patch changes it to 'lib/<tuple>' instead. As a result, the files directly
under 'lib/' will no longer be copied. There should be none, but the dynamic
loader is a notable exception.
[Note: support for these older Linaro toolchain has been removed in 2016.11]

Instead, copy over the ld.so file(s)/link(s) from the extracted toolchain
into staging, in the central copy_toolchain_sysroot function. The existing
toolchain logic will then handle the copy of these files from staging to
target.
This means the toolchain-specific fixups can be removed.

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

toolchain-external: clarify rsync excludes in copy_toolchain_sysroot

The copy_toolchain_sysroot helper features a complex rsync loop that copies
various directories from the extracted toolchain to the staging directory.
The complexity mainly stems from the fact that we support multilib toolchain
tarballs but only copy one of the multilib variants into staging.

Increase understandability of this logic by explicitly restricting the
rsync excludes to the iteration of the for loop they are relevant for.
Additionally, update the function comment.

Note: all attempts to reduce duplication between both rsync while keeping
things nice and readable failed. One has to be extremely careful regarding
line continuation, indentation, and single vs double quoting. In the end, a
split up rsync seemed most clean.

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

DEVELOPERS: add Benoît Allard for boot/syslinux

Signed-off-by: Benoît Allard <benoit.allard@greenbone.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

DEVELOPERS: update the e-mail address of Bartosz Golaszewski

I'm no longer with BayLibre. Change my address to an active one.

Signed-off-by: Bartosz Golaszewski <brgl@bgdev.pl>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

Revert "odroid-mali: mark package as BROKEN"

This reverts commit e2ac546c1951ba8a3e88965d3023eb38a9493f28.

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

odroid-mali: fix broken headers

Upstream has updated the OpenGL
headers (https://github.com/mdrjr/c2_mali/issues/1) it should fix
failures such as:

  http://autobuild.buildroot.net/results/ed8d562ae5fdb472a83f9a07b2f755c80c972c34/

in addition, when the framebuffer variant (i.e not X11) is used, we
update the cflags in egl.pc to pass -DMESA_EGL_NO_X11_HEADERS so that
X11 headers are not included.

Reported-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Dagg Stompler <daggs@gmx.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

uclibc: add patch fixing build failures related to time structures

The latest bump of uClibc-ng 1.0.23 contains a commit touching <fcntl.h>
that causes a large number of build failures related to time
structures. For the moment, let's revert this patch to avoid the build
failures.

Fixes:

  http://autobuild.buildroot.net/results/fcd024f6658890861fd7165e15edc65f77a9d8e7/
  http://autobuild.buildroot.net/results/431135d9f985f3b3f935cc29fe16728b2412dd3a/

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

libselinux: query for python site-packages dir directly

With the bump to version 2.6, the following commit needs
to be taken into consideration for overloading paths.
https://github.com/SELinuxProject/selinux/commit/8162f10e670da963eb65ccf1e7de69ea85aba30d

The PYLIBVER is no longer used and the PYTHONLIBDIR is
renamed to PYSITEDIR with slightly different pathing.

More details can be found in the issue ticket which was
marked as a non-issue after analysis that a Buildroot fix
was the resolution.
https://github.com/SELinuxProject/selinux/issues/51

Signed-off-by: Matthew Weber <matthew.weber@rockwellcollins.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

less: bump version to 487

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

nano: bump version to 2.8.0

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

gnupg2: bump version to 2.1.20

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

squid: bump version to 3.5.25

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

freeswitch: bump version to 1.6.16

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

libiio: Fix tools not linked with pthreads

Add a patch cherry-picked from the upstream libiio project.

Fixes:
http://autobuild.buildroot.net/results/595/595dfd7a17e2bba71844536c9071f7608879e319
http://autobuild.buildroot.net/results/693/69390ba1f7b1460c3217ef88125faeeeda445f20
http://autobuild.buildroot.net/results/71e/71ecc80e891a36305b3dacbd12bf8130fe4fd31f
http://autobuild.buildroot.net/results/4d1/4d1d798ac740403e814b843083b14c9a890dfd54
http://autobuild.buildroot.net/results/00f/00ffdfce0e7780a0f5a5dfadf6bfd63e51c7bbb8
http://autobuild.buildroot.net/results/863/863fa72f443c45793df4837992c5c8dc65590c48

Signed-off-by: Paul Cercueil <paul.cercueil@analog.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

gst-ffmpeg: work-around bogus configure logic on SPARC

The libav version built into the gst-ffmpeg code produces a bogus
binary on SPARC, which causes the following error of the
check-bin-arch script:

  ERROR: architecture for ./usr/lib/gstreamer-0.10/libgstffmpeg.so is Sparc v8+, should be Sparc
  ERROR: architecture for ./usr/lib/gstreamer-0.10/libgstpostproc.so is Sparc v8+, should be Sparc
  ERROR: architecture for ./usr/lib/gstreamer-0.10/libgstffmpegscale.so is Sparc v8+, should be Sparc

The problem is the following bit of code in
gst-lib/ext/libav/configure:

elif enabled sparc; then

    enabled vis && check_asm vis '"pdist %f0, %f0, %f0"' -mcpu=ultrasparc &&
        add_cflags -mcpu=ultrasparc -mtune=ultrasparc

I.e, it checks if the architecture supports the pdist
instruction... but forces -mcpu to ultrasparc while doing so. So it's
like "let's see if this Ultrasparc instruction exists when I force the
compiler to think I'm using Ultrasparc", which is non-sensical. This
has been fixed later on in libav upstream:

  https://git.libav.org/?p=libav.git;a=commit;h=6aa93689abe8c095cec9fa828c2dee3131008995

However, this commit cannot be backported as-is since the shell
function check_inline_asm did not exist in the old libav version
bundled in gst-ffmpeg.

Therefore, we take the simpler route of disabling the VIS
optimizations on SPARCv8 and Leon3.

Fixes:

  http://autobuild.buildroot.net/results/e82d179c3d4f92ad7423693a4b1d42379a3f5411/

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

Revert "qemu: allow to build host variant statically"

This reverts commit a6afa968fc9080080b943885b8599475df9e36c4, which
causes too many different build issues:

  http://autobuild.buildroot.net/results/d5c/d5ced6c9738b285e6fda8c3d41c3c66ee0cf1edc/build-end.log

  http://autobuild.buildroot.net/results/fca/fca01cc5ce13384df0a7f8fad75e2acd05598f99/build-end.log

  http://autobuild.buildroot.net/results/f17/f1755f26a3eacd84de9bd901b214a5bbb964988c/build-end.log

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

package/systemd: better patch to avoid ln --relative

We currently have two patches that address the ln --relative issue
differently. The first one changes the behaviour to generate absolute
symlinks, which is incorrect; the second provides an ad-hoc solution
for a single case.

Replace both of them with a single patch that mimics ln --relative when
the host ln does not support it.

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

portmap: remove package as it is obsolete

The upstream tarball isn't available, no releases since ten years. The
latest change to upstream git is from 2014.  Better use rpcbind for any
RPC portmapper service.

Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
[Thomas: make the legacy option select rpcbind, as suggested by Arnout.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

binutils: remove support for binutils 2.25

Now that we have switched to binutils 2.27 as the default binutils
version, it's time to get rid of binutils 2.25. So this commit remove
the 2.25 version choice, the hash file entry, the patches, and adds a
Config.in.legacy option.

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

binutils: switch to 2.27 as the default version

Now that binutils 2.28 is available, switch to binutils 2.27 as the
default version, for both the host variant and the target variant. Note
that the target variant, when no host variant is built, was still
2.25.1: we forgot to update it to 2.26 when the host version was updated
to 2.26.

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

binutils: remove 2.24 hash

Commit d8878112ad3df2a3d9468cf40a2401b581e7e432 ("binutils: remove
deprecated 2.24.X") removed support for binutils 2.24, but forgot to
remove the hash from binutils.hash. This commit fixes that.

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

binutils: bump to 2.28

 - Patch 0100-elf32-arm-no-data-fix.patch is upstream as of commit
   6342be709e8749d0a44c02e1876ddca360bfd52f, so it is removed.

 - Patch 0130-tc-xtensa.c-fixup-xg_reverse_shift_count-typo.patch is
   upstream as of commit 78fb7e37eb8bb08ae537d6c487996ff17c810332, so
   it is removed.

 - Patch
   0900-Revert-part-Set-dynamic-tag-VMA-and-size-from-dynami.patch is
   upstream as of commit c646b02fdcae5f37bd88f33a0c4683ef13ad5c82, so
   it is removed.

 - All other patches are kept, just refreshed to apply cleanly on the
   new binutils version.

Signed-off-by: Adam Duskett <Adamduskett@outlook.com>
[Thomas: improve commit log.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

uclibc: update to 1.0.23

Bugfix release, internal RPC implementation removed.
All patches applied upstream.

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

uhttpd: fix TCP_FASTOPEN related compile error

Fixes [1]:

  .../uhttpd-a8bf9c00842224edb394e79909053f7628ee6a82/listen.c: In function 'uh_setup_listeners':
  .../uhttpd-a8bf9c00842224edb394e79909053f7628ee6a82/listen.c:120:30: error: 'TCP_FASTOPEN' undeclared (first use in this function)

[1] http://autobuild.buildroot.net/results/56e/56e0727ccd1255b05e03d1b79dc238bd88701230

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

dc3dd: gettextize to match newer gettext version

Because dc3dd is being autoreconfigured and comes with an old gettext
infra, gettextize needs to be called so that the infra is updated to
match the newer version used in Buildroot.

Commit b36d57fab included a patch to add the definition of MKDIR_P to
po/Makefile.in.in in order to fix autobuild failures that ocurred when
host-gettext was built before dc3dd. This patch is no longer necessary
as gettextize adds a new Makefile template which contains the needed
definition, so drop it.

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

python-serial: fix legal info

python-serial is provided under BSD-3-Clause license.

Signed-off-by: Rahul Bedarkar <rahulbedarkar89@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

libplatform: use SPDX short identifier for PHP license

Signed-off-by: Rahul Bedarkar <rahulbedarkar89@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

php-ssh2: add version to license string

LICENSE file refers to PHP license version 3.01

Signed-off-by: Rahul Bedarkar <rahulbedarkar89@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

php-memcached: fix legal info

LICENSE file refer to PHP license version 3.01. License header in all
source files refer to PHP license version 3.01 except source files for
fastlz library which is provided under MIT license and g_fmt.{c,h}
which has ISC-like license header.

Signed-off-by: Rahul Bedarkar <rahulbedarkar89@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

php-imagick: add version to license string

LICENSE file refers to PHP license version 3.01.

Signed-off-by: Rahul Bedarkar <rahulbedarkar89@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

php-geoip: add version to license string

License header in geoip.c refers to PHP license version 3.01.

Signed-off-by: Rahul Bedarkar <rahulbedarkar89@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

php-amqp: add version to license string

LICENSE file refers to PHP license version 3.01.

Signed-off-by: Rahul Bedarkar <rahulbedarkar89@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

package/tyrian: fixes compilation with static libs

Has been tested with: "./support/scripts/test-pkg -c tyrian.cfg -p opentyrian"
Fixes: http://autobuild.buildroot.net/results/0e2345db82b33f591958fc0f72ad914adafe0522
and some similar previous build failure.

Thanks Thomas for the tip ;-).

Signed-off-by: Julien BOIBESSOT <julien.boibessot@armadeus.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

btrfs-progs: bump version to 4.10.2

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

sqlite: bump version to 3180000 (3.18.0)

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

libsoc: add host-pkgconf dependency

Without pkg-config PKG_CHECK_MODULES won't be expanded
and ./configure script produces following error message:

./configure: line 12237: syntax error near unexpected token `PYTHON,'
./configure: line 12237: ` PKG_CHECK_MODULES(PYTHON, python-"$PYTHON_VERSION")'

Fixes:

http://autobuild.buildroot.net/results/86e/86e04bd2b10527130306451e56a7693ed4b4befd

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

libmaxminddb: improve --disable-tests patch

Following review of libmaxminddb patch by Thomas Petazzoni, here is an
enhanced version of the --disable-tests patch which has been sent
upstream.

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

linux-headers: bump 4.{4, 9, 10}.x series

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