engicam: icorem6: Add OF C.TOUCH 12 Kit

This patch add support for Open Frame C.TOUCH 12 Kit.

Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

board: icorem6: Add Qt5 configs for 1280x800 LVDS display

i.CoreM6 OF C.TOUCH 10.1 Kit has 1280x800 LVDS display, this patch adds
relevant qt5 configs with supports opensource Etnaviv graphical stack.

Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

engicam: icorem6: Add OF C.TOUCH 10.1 Kit

This patch add support for Open Frame C.TOUCH 10.1 Kit.

Since the defconfig is no longer only for the starter kit, remove that
from BR2_TARGET_GENERIC_ISSUE.

Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

board: icorem6: Add Qt5 configs for 800x480 LVDS display

i.CoreM6 Starter Kit has 800x480 LVDS display, this patch adds
relevant qt5 configs with supports opensource Etnaviv graphical stack.

Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
[Arnout: mention the qt5 defconfig in readme.txt, otherwise the demo
 applications don't exist]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/freeswitch: install to staging

Needed for upcoming package freeswitch-mod-bcg729.

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
[Arnout: made comment more explicit]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

lbase64: compile with -fPIC

Some architectures (e.g. x86_64, MIPS, microblaze, or1k) require to
build with -fPIC to build a shared library, but the lbase64 build
system doesn't add it. Therefore, we add it in CFLAGS.

Fixes:
http://autobuild.buildroot.net/results/b56/b567096cbfdb89b1895924051265c56204ffb512/

Signed-off-by: Francois Perrad <francois.perrad@gadz.org>
[Arnout: write commit log]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

libressl: fix musl build with older kernel headers

musl provides its own SYS_getrandom definition, but not GRND_NONBLOCK.
This breaks the build with kernel headers older than v3.17. Add a patch
adding a local definition of GRND_NONBLOCK to fix the build.

The following defconfig reproduces the build failure:

BR2_x86_pentium_mmx=y
BR2_TOOLCHAIN_BUILDROOT_MUSL=y
BR2_KERNEL_HEADERS_3_12=y
BR2_PACKAGE_LIBRESSL=y

The getentropy_linux.c file is in upstream tarball, but not in its git
repository. It originates from OpenBSD. For this reason the patch is
against the tarball, but not git formatted.

Cc: Adam Duskett <aduskett@gmail.com>
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
[Arnout: change filename to correspond to how git creates it]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

trinity: undefined UIO_MAXIOV need sys/io.h

This upstream patch fixes the build error produced by building for
powerpc 32bit after the glibc bump to 2.26 with the following internal
toolchain config.

gcc5.4.0
glibc2.26
bin2.28.1
linux4.1.43

Failure log:
  CC net/proto-ipv4.o
net/proto-ipv4.c: In function ‘ip_setsockopt’:
net/proto-ipv4.c:231:49: error: ‘UIO_MAXIOV’ undeclared (first use in this function); did you mean ‘_IO_MAGIC’?
   so->optlen = rnd() % sizeof(unsigned long)*(2*UIO_MAXIOV+512);
                                                 ^~~~~~~~~~
                                                 _IO_MAGIC

Signed-off-by: Matthew Weber <matthew.weber@rockwellcollins.com>
[Arnout: add Matthew's Sob to the patch]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

support/testing: fix remaining code style

Fix the remaining code style warnings from flake8:
 - properly indent continuation lines;
 - use proper code to test a parameter is not None.

Cc: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

.flake8: add config file for Python code style

We recommend wrapping at 80 columns but we accept 132 columns when it
makes more readable.

Follow up of discussion at this thread:
http://lists.busybox.net/pipermail/buildroot/2017-October/203624.html

Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Peter Korsgaard <peter@korsgaard.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

testing/tests/init: use lowercase method names

Use method naming convention from PEP8 as other test cases already do.

sed \
  -e 's,startEmulator,start_emulator,g' \
  -e 's,checkInit,check_init,g' \
  -e 's,checkNetwork,check_network,g' \
  -i support/testing/tests/init/*.py

Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

support/testing: fix code style

Fix the trivial warnings from flake8:
 - remove modules imported but unused;
 - use 2 lines before class or module level method;
 - remove blank line at end of file.

Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

support/testing: standardize defconfig fragments style

Change all defconfig fragments to take advantage of
"cf3cd4388a support/tests: allow properly indented config fragment".

Make each defconfig fragment:
 - start after a backslash;
 - be declared as a multi-line string literal;
 - be indented one level more than the variable that contains it.

Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

support/testing: allow to indent ccache defconfig fragment

Move the strip out of leading spaces in defconfig fragments from the
BRTest class to the Builder class. It actually postpones the strip out,
consequentially allowing test cases to post-process the defconfig in
their own __init__ before calling the __init__ method from BRTest.

Moving this code to the Builder class also allows any new test class
that inherits from BRTest to reuse the same code even if the new class
overrides the setUp method.

At same time, prepend a newline to the jlevel handling otherwise it
would stop working for defconfig fragments that don't end in a newline.

Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

u-boot: add option to specify config fragments

U-Boot nowadays also uses kconfig, so we can handle config fragments like we
do for barebox/busybox/linux.

Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

mosh: bump to version 1.3.2

And add hashes for the license files while we're at it.

Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/apache: bump to version 2.4.28

Fix for CVE-2017-9798 is included in this release, so this patch is
removed.

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
[Update commit log: not a security bump]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

wireguard: bump to version 0.0.20171005

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

configs/atmel: addition of the sama5d27 som1 ek board

Add the new SAMA5D27 SOM1 Evaluation Kit with linux4sam_5.7 components.

Signed-off-by: Nicolas Ferre <nicolas.ferre@microchip.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

configs/atmel: bump to linux4sam_5.7

Bump at91sam9x5ek, atmel_sama5d2_xplained, atmel_sama5d3_xplained and
atmel_sama5d4_xplained all variants to linux4sam_5.7. The 3 foundation
components have their tags changed (AT91Bootstrap, U-Boot, Linux kernel).

Signed-off-by: Nicolas Ferre <nicolas.ferre@microchip.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/x11r7/xserver_xorg-server: security bump to version 1.19.4

Fixes CVE-2017-13721 & CVE-2017-13723:
https://lists.x.org/archives/xorg-announce/2017-October/002809.html

Added all hashes provided by upstream.

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

package/x11r7/xkeyboard-config: bump version to 2.22

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

wireguard: needs 3.10+ kernel

The dependency is actually only for the kernel module (and thus on the
runtime kernel version rather than kernel headers), but as we don't know the
runtime version in kconfig and the user space part doesn't make much sense
without the kernel module this will have to do.

[Peter: extend commit message, add comment to Config.in as suggested by
Arnout, fix off-by-one in comment text]
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

cryptsetup: add host variant

E.G. for generating dm-verity hashes or dm-crypt data at build time in a
post-image script.

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

lvm2: add host variant

This is only for the device-mapper library, so a user selectable
Config.in.host option isn't needed.

[Peter: add host-pkgconf dependency as pointed out by Arnout]
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

shairport-sync: bump to version 3.1.3

Fixes a runtime startup issue on SysV init systems:
```
# shairport-sync -d
Daemon returned 2 as return value.
```

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

libcurl: security bump to version 7.56.0

Drop upstreamed patch.

Fixes CVE-2017-1000254 - FTP PWD response parser out of bounds read:

https://curl.haxx.se/docs/adv_20171004.html

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

package/iprutils: bump version to 2.4.15.1

Signed-off-by: Samuel Mendoza-Jonas <sam@mendozajonas.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

qemu: change to .tar.xz format

And use the official download location.

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

qemu: security bump to version 2.8.1.1

Fixes the following security issues and adds a number of other bigfixes:

2.8.1: Changelog:
https://lists.gnu.org/archive/html/qemu-devel/2017-03/msg06332.html

CVE-2017-2615 - display: cirrus: oob access while doing bitblt copy backward
mode

CVE-2017-2620 - display: cirrus: out-of-bounds access issue while in
cirrus_bitblt_cputovideo

CVE-2017-2630 - nbd: oob stack write in client routine drop_sync

2.8.1.1 Changelog:
https://lists.gnu.org/archive/html/qemu-devel/2017-04/msg03460.html

CVE-2017-7471 - 9p: virtfs allows guest to change filesystem attributes on
host

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

openntpd: fix musl build with older kernel headers

musl provides its own SYS_getrandom definition, but not GRND_NONBLOCK.
This breaks the build with kernel headers older than v3.17. Add a patch
adding a local definition of GRND_NONBLOCK to fix the build.

The getentropy_linux.c file is in upstream tarball, but not in its git
repository. It originates from OpenBSD. For this reason the patch is
against the tarball, but not git formatted.

Fixes:
http://autobuild.buildroot.net/results/623/6235ddc095bc1984439e7339d590b11581a3e257/
http://autobuild.buildroot.net/results/648/648e7c91edcf893fa780de9bb44c05f5f842a2e8/
http://autobuild.buildroot.net/results/7da/7da71263a6e9b780031f1f6185c760d11dfc994e/

Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

vboot-utils: disable static futility

Possibly fixes
http://autobuild.buildroot.net/results/fbe/fbe7892539cc1bf18f89d18c6609a5aa747647d0/
http://autobuild.buildroot.net/results/4bc/4bc388618ff02dde26ae09db021434a936c8cc1b/
and similar errors.

Signed-off-by: Alex Suykov <alex.suykov@gmail.com>
Tested-by: Matt Weber <matthew.weber@rockwellcollins.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/tor: fix openssl detection for static builds

Fixes
http://autobuild.buildroot.net/results/8e6/8e639ab8912e7d884fd8e6dbb1ca8b49451dd766/

/home/test/autobuild/run/instance-1/output/host/arm-buildroot-linux-uclibcgnueabi/sysroot/usr/lib/libcrypto.a(c_zlib.o):
In function `zlib_stateful_expand_block':
c_zlib.c:(.text+0x54): undefined reference to `inflate'
/home/test/autobuild/run/instance-1/output/host/arm-buildroot-linux-uclibcgnueabi/sysroot/usr/lib/libcrypto.a(c_zlib.o):
In function `zlib_stateful_compress_block':
c_zlib.c:(.text+0xd4): undefined reference to `deflate'

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

HOST_DIR/lib: symlink respectively to lib32/64

Discovered the issue on a RHEL7.4 machine where
the cmake build dynamically selected HOST_DIR/lib64
as the installation path for the lzo2 library.

Fixes failures like the following:
host-mtd
http://autobuild.buildroot.net/results/d31/d31581d2e60f35cf70312683df99c768e2ea8516/

host-squashfs
http://autobuild.buildroot.net/results/d9c/d9c95231ac774ed71580754a15ebb3b121764310/

Signed-off-by: Matthew Weber <matthew.weber@rockwellcollins.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

linux: bump default to version 4.13.5

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

linux-headers: bump 4.{4, 9, 13}.x series

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

qemu: drop obsolete "--disable-uuid" configuration parameter

./configure: --disable-uuid is obsolete, UUID support is always built

Change-Id: I9e278418d19e15bbbd3ea233658cd62f75e3385c
Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/turbolua: Fix build with gcc7

gcc7 has -Wimplicit-fallthrough enabled by default. Together with
-Werror in turbolua makefile it gives -Werror=implicit-fallthrough=
build error.

Add upstreamable patch [1], which fixes typo in FALLTHROUGH comment,
so it makes the fallthrough explicit for gcc.

[1] https://github.com/kernelsauce/turbo/pull/323

Fixes:
http://autobuild.buildroot.net/results/6f107b62f105508044185576b071323e5d73dd83
http://autobuild.buildroot.net/results/d303929960c667a21181d0bbdab64ba8a8d56a4f

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

lua-basexx: new package

Signed-off-by: Francois Perrad <francois.perrad@gadz.org>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

lua-compat53: new package

Signed-off-by: Francois Perrad <francois.perrad@gadz.org>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

lua-cqueues: new package

Signed-off-by: Francois Perrad <francois.perrad@gadz.org>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

lua-fifo: new package

Signed-off-by: Francois Perrad <francois.perrad@gadz.org>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

qt: Allow enabling of QtWebKit with GCC 6+

Building Qt with QtWebKit on configuration step there is
a check which disables QtWebKit build with GCC 6+.
Back in the day nobody thought about building Qt with GCC
version greater than 5.x. And now with modern GCCs like
6.x and 7.x this assumption gets in the way.

Given in Buildroot today we don't have GCC older than 4.9
it should be safe to remove now meaningless check completely
by adding patch to qt.

Signed-off-by: Evgeniy Didin <didin@synopsys.com>
Cc: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

wireguard: bump version to 0.0.20171001

And add a hash for the license file while we're at it.

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

package/bridge-utils: fix headers path

Currently, bridge-utils wants to grap the headers from the linux-headers
package, so we point it directly there, as has been the case since we
first added bridge-utils in 2003 (c8eea31d3f), and then further refined
in 2005 (178a317d26) which is the first moment we pointed to the linux-
headers directory.

However, ther are two things wrong with that.

First, the headers are not directly in $(LINUX_HEADERS_DIR). Instead,
they are in a sub-directory thereof. So, we could not have found them
the way we are doing now.

Second, this definitely does not work when using an external toolchain,
because there is not linux-headers package enabled then.

Yet, against all odds, bridge-utils has valiantly deflected all rocks
thrown its way, day-in day-out building without any issue in every
autobuilders it's been confronted with. Good boy, good boy. :-)

And indeed, it turns out that the required headers are easily found from
within the sysroot of the toolchain. Wonders! :-)

But there's still a gotcha: the default search path is still a hard
coded path pointing to the installed kernel source tree on the host.

So, we still have to pass this option, but we can simply point to the
sysroot.

[Peter: point to sysroot instead of non-existing directory]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/netsniff-ng: add missing header

Like for nfs-utils and strongswan, add stdint.h header for
UINT16_MAX definition

Reported-by: Ross <grunpferd@netscape.net>
Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

nfs-utils: add missing include of stdint.h

The Glibc bump to 2.26 exposed this missing header when building with the
following combination using an i386 internal toolchain:

gcc5.4.0
bin2.28.1
linux4.1.43

Logfile snippet:

2017-10-02T00:08:11 ^
2017-10-02T00:08:12 rpc.c: In function ‘nsm_recv_getport’:
2017-10-02T00:08:12 rpc.c:469:13: error: ‘UINT16_MAX’ undeclared (first use in this function)

Upstream: https://bugzilla.linux-nfs.org/show_bug.cgi?id=312

Signed-off-by: Matthew Weber <matthew.weber@rockwellcollins.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

strongswan: add missing include of stdint.h

Recent releases of glibc don't include the full stdint.h
header in some network headers included by utils.h.
Upstream is targetting a 5.6.1 release of the fix.

Ustream: https://wiki.strongswan.org/issues/2425

Signed-off-by: Matthew Weber <matthew.weber@rockwellcollins.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

dump1090: fix librtlsdr package name typo

Fixes:
http://autobuild.buildroot.net/results/720/720a4b23a47aa6677bde4418f21e6b073a72ea51/
http://autobuild.buildroot.net/results/14e/14edf051d113a6ec54b61738748764adeac7ba9d/

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

package/icu: remove xlocale.h header include

Remove xlocale.h header include since it's removed from glibc 2.26.

Reported-by: Ross <grunpferd@netscape.net>
Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/php: bump version to 7.1.10

Changelog: http://www.php.net/ChangeLog-7.php#7.1.10

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

package/{mesa3d, mesa3d-headers}: bump version to 17.2.2

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

x11r7: xdriver_xf86-input-tslib: add the hash for the COPYING license file

The COPYING file is part of the xf86-input-tslib project's source tree.
Let's include the hash for it's current version.

Signed-off-by: Martin Kepplinger <martink@posteo.de>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

wireguard: add new package

[Peter: Fix help text, drop suboptions, drop patches, fix dependencies,
 handle optional bash dependency, use kernel-module infrastructure,
 enable needed kernel options]
Signed-off-by: Aurélien Chabot <aurelien@chabot.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

jsoncpp: bump to version 1.8.3

Signed-off-by: Jörg Krause <joerg.krause@embedded.rocks>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

arc/bfin: remove 60x cores

Those cores are not supported in upstream gcc, not even in master.

The only toolchain that supported those core was the 2014R1 ADI
rebuilt toolchain, but we removed it in 311bc13 (toolchain: kill
ADI Blackfin toolchain) because there was too many issues with it.
ADI has not released any newer toolchain since then.

There is little hope for those cores now, so remove them.

Support for those cores has been useless and unusable for a while
without nobody noticing, therefore we intentionally skip adding
Config.in.legacy. This would require keeping code in
arch/Config.in.bfin since the options being removed are inside a
choice...endchoice block.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
[Thomas: explain why we don't add the options to Config.in.legacy.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

arch/bfin: internal backend not suitable for some cores

Some cores are not supported by upstream gcc.

Use the newly-introduced symbol to state so, rather than have the
exclusion in the toolchain choice.

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

arch/mips: internal backend not suitable for some cores

Some cores are not supported by upstream gcc.

Use the newly-introduced symbol to state so, rather than have the
exclusion in the toolchain choice.

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

arch/csky: internal backend not suitable

Upstream gcc does not have support for C-Sky, and we do not have a
vendor tree for it either (yet?).

Use the newly-introduced symbol to state so, rather than have the
exclusion in the toolchain choice.

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

arch: add option to disable internal toolchain backend

Some architectures or specific cores do not have support in upstream
gcc. Currently, they are individually listed as exclusions in the
toolchain choice.

This poses a maintainance burden, as the knowledge about what gcc
version supports what architecture is split across many places: the
toolchain choice, the gcc version choice, the external toolchains.

As a first step, add a blind option that architectures or individual
cores may select to indicate they lack support in our internal backend.

Actual use of the option will come in followup patches.

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

tslib: add the hash for the COPYING license file

Since tslib's COPYING file in part of the source tree, we can easily
include it's hash for the given version here.

Signed-off-by: Martin Kepplinger <martink@posteo.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

monit: bump to version 5.24.0

Signed-off-by: Jörg Krause <joerg.krause@embedded.rocks>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

dnsmasq: simplify build configuration

Drop direct sed'ing of config.h for HAVE_CONNTRACK, HAVE_LUASCRIPT, and
HAVE_DBUS. Use MAKE_OPTS COPTS parameters instead, like we do already
for all other options.

Rename DNSMASQ_ENABLE_LUA to DNSMASQ_TWEAK_LIBLUA since it now does only
that.

Merge two conntrack and three dbus conditional sections.

Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

dnsmasq: security bump to version 2.78

Supported Lua version is now 5.2.

Add licenses hash.

Fixes a number of security issues:

CVE-2017-13704 - Crash when DNS query exceeded 512 bytes (a regression
in 2.77, so technically not fixed by this bump)

CVE-2017-14491 - Heap overflow in DNS code

CVE-2017-14492 - Heap overflow in IPv6 router advertisement code

CVE-2017-14493 - Stack overflow in DHCPv6 code

CVE-2017-14494 - Information leak in DHCPv6

CVE-2017-14496 - Invalid boundary checks allows a malicious DNS queries
to trigger DoS

CVE-2017-14495 - Out-of-memory Dos vulnerability

Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

Config.in: rename xbmc -> kodi

This is a left-over of the xbmc->kodi rename patch
https://git.buildroot.net/buildroot/commit/package/kodi?id=3578459748d80cd469729319543b199380a4370f

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

sdl2: Bump version to 2.0.6

Bump version and remove patches that were merged upstream

Signed-off-by: Olivier Schonken <olivier.schonken@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

configs/raspberrypi*: bump kernel version

Now based on 4.9.52 (from 4.9.36).

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

package/urg: fix extraction commands

Currently, the extraction commands entirely remove the urg directory,
which means the downloaded stamp will get removed, and thus a subsequent
build would try to re-download it.

It turns out that the directory extracted by urg is already correctly
named, so we just need to extract out of the build directory. This
highly simplifies the command.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

bananapro: bump Linux to version 4.13.4

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

bananapro: bump U-Boot to version 2017.09

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

package/weston: fix build on musl

As usual, musl is more conservative than the alternatives, wrt to
headers that are internally included.

Fixes:
    http://autobuild.buildroot.org/results/a7a/a7a6b17dff09a45a35185a0e02704523b815dd57/

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

dump1090: new package

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

package/gdb: bump to version 8.0.1

https://sourceware.org/ml/gdb-announce/2017/msg00004.html

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

package/binutils: bump to 2.29.1

There is a problem between the git tag and the archive release content,
they are not in sync [1].

The version used to generate the archive allow us to remove the xtensa
patches.

Otherwise, see the announce [2].

[1] https://sourceware.org/ml/binutils/2017-09/msg00337.html
[2] https://sourceware.org/ml/binutils/2017-09/msg00311.html

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

package/nodejs: bump version to 8.6.0

See https://nodejs.org/en/blog/release/v8.6.0/

Signed-off-by: Martin Bark <martin@barkynet.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

configs: nexbox_a95x_defconfig: bump to kernel 4.13.4

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

package/supertuxkart: needs freetype

https://github.com/supertuxkart/stk-code/blob/master/CMakeLists.txt#L231

Build would fail without this patch using this defconfig:

BR2_x86_64=y
BR2_TOOLCHAIN_BUILDROOT_LOCALE=y
BR2_TOOLCHAIN_BUILDROOT_CXX=y
BR2_PACKAGE_SUPERTUXKART=y
BR2_PACKAGE_MESA3D=y
BR2_PACKAGE_MESA3D_DRI_DRIVER_I965=y
BR2_PACKAGE_MESA3D_OPENGL_EGL=y
BR2_PACKAGE_XORG7=y

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Acked-by: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>
Acked-by: Romain Naour <romain.naour@gmail.com>
Tested-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

zsh: add hash for license file

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

zsh: bump to version 5.4.2

Upstream does not provide a .tar.xz archive for this release. Revert to
.tar.gz.

Cc: Phil Eichinger <phil.eichinger@gmail.com>
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

zsh: disable NIS support

The configure test for NIS and NIS+ looks for the ypcat and nisls host
utilities, respectively. This is not compatible with cross compilation.
Disable both unconditionally.

Fixes:
http://autobuild.buildroot.net/results/54c/54c9c5dff3df0a43573b03896cb0629fcae08478/
http://autobuild.buildroot.net/results/c8e/c8e30a1eb6e77ae76292b6a01bede5abb2b1b7b7/

Cc: Phil Eichinger <phil.eichinger@gmail.com>
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

Revert "package/kodi-inputstream-rtmp: bump version to 1.0.6"

This reverts commit eec04a7b7dc32749f9c1f5e0ac1087230c542f28.

The original commit bumped the package to a version which is at runtime
incompatible with Kodi.

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

package/opencore-amr: fix license, add license hash

Fixes
http://autobuild.buildroot.net/results/656/65667eb8fa1d447251b707e8a201d811cdaf7590/

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

ejabberd: Replace $(HOST_DIR) to /usr in ERL path

Previously, it was working by luck. Buildroot has fixed its definition
of HOST_DIR and pkg-autotools.mk uses the classical /usr prefix. So,
fix this sed expression to correctly replace $(HOST_DIR) by /usr in ERL
path.

Signed-off-by: Johan Oudinet <johan.oudinet@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

iproute2: fix static build with uClibc

Version 4.13 added a local definition of strlcpy(). This clashes with
the uClibc provided one when linking statically. Add a patch that
detects libc provided strlcpy(), and avoids the multiple definition.

Fixes:
http://autobuild.buildroot.net/results/f34/f34684dcdb47938a3da8b00c8b29000cc23b4e3d/
http://autobuild.buildroot.net/results/a01/a01c5775945f5ffe75451722b676fc93333a0f14/
http://autobuild.buildroot.net/results/0ee/0eeab1a0ca943f9a035a0d458ddf9cc210bc87f4/

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

package/openobex: disable doc build

Openobex would need the host variant of xmlint (provided by host-
libxml2) to build the documentation.
To avoid this dependency we disable the build of the documentation.

Fixes
http://autobuild.buildroot.net/results/668/668ffb15f8c4faf0f4c462488dbb70b61386d465/

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

evemu: bump version to 2.7.0

See [1] for details.

[1] https://lists.freedesktop.org/archives/input-tools/2017-September/001416.html

Signed-off-by: Peter Seiderer <ps.report@gmx.net>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

orangepi_pc: bump linux to 4.13.3

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

orangepi_pc: bump u-boot to 2017.09

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

orangepi_zero: bump u-boot to 2017.09

Cc: Sergey Matyukevich <geomatsi@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

DEVELOPERS: Update imx6-sabresd config entry

After the conversion to SPL the imx6sabresd defconfig targets have
been renamed,so update the entry accordingly.

Suggested-by: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: Fabio Estevam <festevam@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

.gitlab-ci.yml: Update imx6-sabresd defconfig entries

After the conversion to SPL the imx6sabresd defconfig targets have
been renamed, so update their entries accordingly.

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

perl: bump to version 5.26.1

Signed-off-by: Francois Perrad <francois.perrad@gadz.org>
Tested-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

configs/mx25pdk: Bump U-Boot and kernel versions

Bump to U-Boot 2017.09 and kernel to 4.13.3.

Signed-off-by: Fabio Estevam <fabio.estevam@nxp.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/libmad: don't install .pc in target

We explicitly remove pkg-config files from target in the target-finalize
step, and our pkg-config looks for .pc files from staging, not target.

So, explicitly installing the .pc in target is useless.

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

engicam: imx6qdl_icore: Bump to latest Linux and uboot releases

- Linux: 4.13
- Uboot: 2017.07

Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

flex: Fix segfault of stage1flex when host has glibc 2.26 or newer

When the host uses glibc 2.26 or newer, Flex will try to use the
newly-introduced reallocarray() function, but as it would not define
_GNU_SOURCE a segmentation fault would occur later on due to the
compiler assumming that the function is implicitly defined.

This issue manifests itself due to a crash of "stage1flex" during the
Flex bootstrap:

  ./stage1flex   -o stage1scan.c ./scan.l
  make[2]: *** [Makefile:1725: stage1scan.c] Segmentation fault (core dumped)

This imports the patch from the upstream Git repository, and adds flags
in the .mk file to rebuild the Autotools scripts and support files. Due
to the latter, the patch to disable the documentation is changed so it
modifies the .am file instead.

Signed-off-by: Adrian Perez de Castro <aperez@igalia.com>
[Arnout:
 - Add Adrian's Sob and upstream ref to new patch;
 - Keep patch 1 as patch 1;
 - Keep Vicente as author of path 1;
 - Add reason for autoreconf in a comment.]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/glibc: bump to 2.26

Drop upstream patch:

0001-sh-Fix-building-with-gcc5-6.patch
https://sourceware.org/git/?p=glibc.git;a=commit;h=d40dbe722f004f999b589de776f7e57e564dda01

0002-CVE-2017-1000366-Ignore-LD_LIBRARY_PATH-for-AT_SECUR.patch
https://sourceware.org/git/?p=glibc.git;a=commit;h=f6110a8fee2ca36f8e2d2abecf3cba9fa7b8ea7d

0003-ld.so-Reject-overly-long-LD_PRELOAD-path-elements.patch
https://sourceware.org/git/?p=glibc.git;a=commit;h=6d0ba622891bed9d8394eef1935add53003b12e8

0004-ld.so-Reject-overly-long-LD_AUDIT-path-elements.patch

https://sourceware.org/git/?p=glibc.git;a=commit;h=81b82fb966ffbd94353f793ad17116c6088dedd9

0005-fix-binutils-2-29-build.patch
https://sourceware.org/git/gitweb.cgi?p=glibc.git;a=commitdiff_plain;h=388b4f1a02f3a801965028bbfcd48d905638b797

0006-i686-Add-missing-IS_IN-libc-guards-to-vectorized-strcspn.patch
https://sourceware.org/git/gitweb.cgi?p=glibc.git;a=commitdiff_plain;h=69052a3a95da37169a08f9e59b2cc1808312753c

0006-sh4-trap.patch
https://sourceware.org/git/gitweb.cgi?p=glibc.git;a=commitdiff_plain;h=db3d848e154b00071f4a5e729d5884efad410109

But 2.26 version introduced quite a few regressions on x86, x86_64 and
powerpc64le when float128 support was added. All these issues are now
fixed in the glibc 2.26 stable branch.

Backport patches from glibc 2.26 stable branch related to bug 21930.
Take all patches listed in this bug report in order to have all fix
for this issue [1] [2].

Fixes gnuradio build:
[...]/sysroot/usr/include/boost/functional/hash/detail/hash_float.hpp: In function ‘std::size_t boost::hash_detail::float_hash_value(T)’:
[...]/sysroot/usr/include/boost/functional/hash/detail/hash_float.hpp:210:21: error: expected primary-expression before ‘float’
             switch (fpclassify(v))
                     ^
[...]/sysroot/usr/include/boost/functional/hash/detail/hash_float.hpp:210:21: error: expected primary-expression before ‘long’
             switch (fpclassify(v))
                     ^
[...]/sysroot/usr/include/boost/functional/hash/detail/hash_float.hpp:210:21: error: found ‘:’ in nested-name-specifier, expected ‘::’
             switch (fpclassify(v))
                     ^
[...]/sysroot/usr/include/boost/functional/hash/detail/hash_float.hpp:210:21: error: ‘_Float128’ is not a class or namespace
             switch (fpclassify(v))
                     ^

[...]/sysroot/usr/include/boost/functional/hash/detail/hash_float.hpp:210:21: error: ‘_Generic’ was not declared in this scope
             switch (fpclassify(v))
                     ^

Also backport fix for glibc bug 22146.
Without this patch some C++ applications (jsoncpp, mesa3d and kodi) fail
to build due to an issue while building the toolchain.

A test in the libstdc++ configure script fail when -Os and float128 are
used.

See the bug report for details [3].

[1] https://sourceware.org/ml/libc-alpha/2017-08/msg00586.html
[2] https://sourceware.org/bugzilla/show_bug.cgi?id=21930
[3] https://sourceware.org/bugzilla/show_bug.cgi?id=22146

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Tested-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/flex: disable reallocarray

This doesn't compile with glibc 2.26 where reallocarray() as been introduced.

It's a nasty issue, when reallocarray() is available for the target, flex will
build a small tool called stage1flex for the host (using _FOR_BUILD) but with
the config.h generated for the target.

When the host doesn't have a glibc >= 2.26, reallocarray() is never defined
while building stage1flex:

misc.c:147:8: warning : implicit declaration of function « reallocarray »
[-Wimplicit-function-declaration]
  mem = reallocarray(NULL, (size_t) size, element_size);
        ^~~~~~~~~~~~
misc.c:147:6: warning : assignment makes pointer from integer without a cast
[-Wint-conversion]
  mem = reallocarray(NULL, (size_t) size, element_size);
      ^

Disable reallocarray for now, reallocarray() support may be
enabled in a followup patch.

Fixes:
stage1flex-misc.o: In function `allocate_array':
misc.c:(.text+0x38f): undefined reference to `reallocarray'
stage1flex-misc.o: In function `reallocate_array':
misc.c:(.text+0xc8a): undefined reference to `reallocarray'
collect2: error: ld returned 1 exit status

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/diffutils: disable getopt_long replacement with glibc

Since glibc >= 2.26, don't try to use getopt_long replacement bundled
with diffutils. It will conflict with the one from glibc.
The getopt copy bundled in glibc has been updated with the one from gnulib.

After a lot of warnings...
Fixes:
xstrtol-error.c: In function 'xstrtol_error':
xstrtol-error.c:84:26: error: invalid use of undefined type 'struct rpl_option'
     option = long_options[opt_idx].name;
                          ^
xstrtol-error.c:84:26: error: dereferencing pointer to incomplete type 'const struct rpl_option'

[...]

xstrtol-error.c:92:1: error: conflicting types for 'xstrtol_fatal'
 xstrtol_fatal (enum strtol_error err,
 ^~~~~~~~~~~~~
In file included from xstrtol-error.c:20:0:
xstrtol.h:69:16: note: previous declaration of 'xstrtol_fatal' was here
 _Noreturn void xstrtol_fatal (enum strtol_error,
                ^~~~~~~~~~~~~

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/gcc: enable float128 on powerpc64le with glibc >= 2.26 toolchains

float128 is now required on powerpc64le with glibc >= 2.26 toolchains.

Fixes:
running configure fragment for sysdeps/gnu
running configure fragment for sysdeps/powerpc/powerpc64le
checking if [...]powerpc64le-buildroot-linux-gnu-gcc supports binary128 floating point type... no
checking if the target machine is at least POWER8... yes
configure: error: ***  binary128 floating point type (GCC >= 6.2) is required on powerpc64le.

In order to prepare the glibc bump to this version, we enable
binary128 floating point in gcc. This require at least gcc >= 6.2.

Disable gcc 4.x and 5.x for internal toolchain.

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>