support/testing: fix ATF Vexpress test case

This test case currently fails to build with:

./build/juno/release/bl1/context_mgmt.o: In function `cm_prepare_el3_exit':
context_mgmt.c:(.text.cm_prepare_el3_exit+0x54): undefined reference to `cm_set_next_context'
context_mgmt.c:(.text.cm_prepare_el3_exit+0x54): relocation truncated to fit: R_AARCH64_JUMP26 against undefined symbol `cm_set_next_context'

This issue has been fixed upstream in commit
10c252c14b7f446c0b49ef1aafbd5d37804577dd, available since v1.3. So
while we bump, let's bump to the latest version of ATF, v1.5.

Fixes:

  https://gitlab.com/buildroot.org/buildroot/-/jobs/64360659

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

linux-headers: bump 4.{4, 9, 14, 16}.x series

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

linux: bump default to version 4.16.4

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

configs/warp7: Bump kernel and U-Boot versions

Bump kernel to version 4.16.3 and U-Boot to 2018.03.

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

support/scripts/check-uniq-files: remove csv module import

Since commit 5563a1c6a48716debe2983869ddb757318094dce
("support/check-uniq-files: support weird locales and filenames"), the
'csv' Python module is no longer used by the check-uniq-files.

Due to this, flake8 complains with:

support/scripts/check-uniq-files:4:1: F401 'csv' imported but unused

Fix this by dropping the useless csv import.

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

mbedtls: security bump to version 2.7.2

The release announcement mentions these security fixes:

  Defend against Bellcore glitch attacks by verifying the results of RSA
  private key operations.

  Fix implementation of the truncated HMAC extension. The previous
  implementation allowed an offline 2^80 brute force attack on the HMAC
  key of a single, uninterrupted connection (with no resumption of the
  session).

  Reject CRLs containing unsupported critical extensions.

  Fix a buffer overread in ssl_parse_server_key_exchange() that could
  cause a crash on invalid input. (CVE-2018-9988)

  Fix a buffer overread in ssl_parse_server_psk_hint() that could cause
  a crash on invalid input. (CVE-2018-9989)

Drop upstream patch.

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

lz4: disable static library on shared only

Don't build the static library when BR2_SHARED_LIBS=y, to reduce build
time.

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

modem-manager: support a defaults file in the init script

Sometimes it is useful to pass some parameters to ModemManager when it
starts (e.g. --log-level). Allow the user add a file with such flags in
a MODEMMANAGER_ARGS variable. This is simpler than overriding the whole
startup script (e.g. by means of a rootfs overlay).

Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Acked-by: Petr Vorel <petr.vorel@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

lz4: don't build library and tools in parallel

Since commit e93fec6673e (lz4: bump to version 1.8.1.2) we use two make
targets, 'lib' and 'lz4'. Both targets depend on the static library.
When these targets run in parallel the two make instances that build the
static library race against each other. Split these targets to ensure
build order.

Should fix:
http://autobuild.buildroot.net/results/a8d/a8d956ff420f6a265c5c00b33646dbbc24ce2d48/

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

configs: add openssl for kernel where necessary

Add host-openssl to those configs, which need it for the Linux kernel build.

Signed-off-by: Vincent Stehlé <vincent.stehle@laposte.net>
Cc: Matthew Weber <matthew.weber@rockwellcollins.com>
Cc: Marcin Niestroj <m.niestroj@grinn-global.com>
Cc: Fabio Estevam <fabio.estevam@nxp.com>
Reviewed-by: Fabio Estevam <festevam@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

strace: fix build for m68k

Add upstream patch fixing build for m68k target.

Fixes:
http://autobuild.buildroot.net/results/bef/befdcb3f87c0490cd623520403b9467ec7359950/
http://autobuild.buildroot.net/results/ddb/ddb7616991468bc45d85aa1b56bec2cb5eb897c4/

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

package/unixodbc: bump version to 2.3.6

Added license hashes.

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

package/ccid: bump version to 1.4.29

Upstream moved to https://ccid.apdu.fr/ according to
http://pcsclite.alioth.debian.org/ccid.html, updated _SITE and
Config.in.

Added license hash.

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

package/pcsc-lite: bump version to 1.8.23

Upstream moved to https://muscle.apdu.fr/ according to
http://pcsclite.alioth.debian.org/, updated _SITE and Config.in.

Added license hash.

Added optional systemd support available since upstream commit
https://github.com/LudovicRousseau/PCSC/commit/30e10951f81b9480e788965f89d0d4d0aee909c0#diff-67e997bcfdac55191033d57a16d1408a

We need to add systemd support in this patch because omitting it will
cause an error during configure:

checking for LIBSYSTEMD... no
configure: error: install libsystemd-dev or use --disable-libsystemd

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

support/config-fragments/autobuild: update ARC pre-built toolchain

The toolchain has been rebuilt to fix the RPATH of the binutils programs.

Fixes:

  http://autobuild.buildroot.net/results/b2562b05d397d4e1ffe0f8d2f4ce4c84ab6feae1/

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

support/config-fragments/autobuild: update OpenRISC pre-built toolchain

The toolchain has been rebuilt after commit
7fb6fe883a6ee360aa861fdae93d7441f9f5d52d, which adds more fcntl()
definitions in uClibc for OpenRISC.

Fixes:

  http://autobuild.buildroot.net/results/664e7ee968e46b7ac0054ce4d5d4f9a82201550e/

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

bluez5_utils: add patch to fix readline issue

Since bluez5_utils 5.49, some code using readline was compiled even if
readline was not available. After this issue was reported upstream, a
patch was proposed by an upstream developer to address the issue. This
commit integrates this patch (under review upstream), which fixes the
problem.

Fixes:

  http://autobuild.buildroot.net/results/3e266a79acab8b8eb33360f7afbc1cd6db46f7cb/

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

lxc: fix offsetof() related musl build issue

This commit adds a patch that fixes a build issue with musl due to the
fact that offsetof() is used without including <stddef.h>.

Fixes:

  http://autobuild.buildroot.net/results/3c6e34cbcd43aba4dd2b18687bd89fc369c9dab7/

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

uclibc: backport upstream patch adding fcntl definitions on OpenRISC

A number of fcntl.h definitions were missing on OpenRISC in uClibc-ng,
causing build failures such as:

gfile.c: In function 'splice_stream_with_progress':
gfile.c:3017:35: error: 'F_SETPIPE_SZ' undeclared (first use in this function)
   buffer_size = fcntl (buffer[1], F_SETPIPE_SZ, 1024 * 1024);

This comit backports a patch that was merged in upstream uClibc-ng,
which addresses this problem.

Fixes:

  http://autobuild.buildroot.net/results/182ebbba6c7466b07e965e5120d919cbcf5e9da6/

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

support/config-fragments/autobuild: update all Buildroot pre-built toolchains

This commit updates all the pre-built Buildroot toolchains, which have
all been rebuilt with Buildroot as of commit 046c5e2. The initial
motivation for this update is that an upcoming bump of procps-ng uses
fopencookie(), which has only been introduced in musl 1.1.19, which
itself started being used in Buildroot after the 2018.02 release.

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

package/{mesa3d, mesa3d-headers}: bump version to 18.0.1

Removed patch 0002, upstream committed a different fix:
https://cgit.freedesktop.org/mesa/mesa/commit/?h=18.1&id=cbee1bfb34274668a05995b9d4c78ddec9e5ea4c
https://cgit.freedesktop.org/mesa/mesa/commit/?h=18.1&id=5a75019ad0270a974788a9b8648ba98ff4203768

Renumbered patch 0003 -> 0002.

Tested using test-pkg (removed skipped toolchains from log):

                armv5-ctng-linux-gnueabi [ 1/46]: OK
              armv7-ctng-linux-gnueabihf [ 2/46]: OK
                        br-aarch64-glibc [ 3/46]: OK
                           br-arcle-hs38 [ 4/46]: OK
                  br-arm-cortex-a9-glibc [ 6/46]: OK
                   br-arm-cortex-a9-musl [ 7/46]: OK
                             br-arm-full [ 9/46]: OK
                   br-i386-pentium4-full [12/46]: OK
                br-i386-pentium-mmx-musl [13/46]: OK

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

linux-headers: bump 4.9.x series

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

libuv: bump to version 1.20.1

Signed-off-by: Asaf Kahlon <asafka7@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

linux-headers: bump 4.{14, 15, 16}.x series

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

linux: bump default to version 4.16.3

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

bootstrap: bump to version 4.1.0

Also install bootstrap.min.css.map as it is referenced from
bootstrap.min.css.

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

package/tk: new package

Use the same version as tcl package.

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

package/tcllib: update license to SPDX format

While at it add the license file hash.

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

package/tcl: update license to SPDX format

While at it add the license file hash.

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

download/git: fix transform regexp for older tar versions

Older versions of tar (e.g. 1.27.1) incorrectly interpret the escaping
of the regexp separator, and generate broken tarballs.

For example, given the following transform expression:
    --transform="s/^\.\//squashfs-e38956b92f738518c29734399629e7cdb33072d3\//"

the resulting paths in the generated tarball would be:
    squashfs-e38956b92f738518c29734399629e7cdb33072d3\/

i.e. a directory which last character is indeed a '\'.

We fix that by using a separator which is very unlikely to occur in a
filename.

Fixes:
    http://autobuild.buildroot.org/results/742/7427f34e5c9f6d043b0fe6ad2c66cc0f31d2b24f/

and probably a slew of others as well...

Take this opportunity to fix indentation on the following line
(leading spaces, not TABs).

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

lz4: improve static only build support

The current method of supporting static only build, removal of all lines
that match the SHARED regex from lib/Makefile, is crude and fragile.
Instead, patch lib/Makefile to allow disable of shared libraries.

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

dhcpcd: fix legal info

The latest version bump to 7.0.3 moved dhcpcd.c to under src/. Since the
tarball now has a LICENSE file, use that instead.

While at it, add a hash for that file.

Fixes:
http://autobuild.buildroot.net/results/f57/f5790e0a82a3b814d828d18eb71a3f6c422bb2b8/
http://autobuild.buildroot.net/results/6b9/6b930827913f9b4d2602d5777f4243ba7b295ca7/
http://autobuild.buildroot.net/results/603/603eda004a502f613e9b92cea4fb8cc9e4068255/

Cc: John Stile <johns@msli.com>
Cc: Yegor Yefremov <yegorslists@googlemail.com>
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Reviewed-by: Yegor Yefremov <yegorslists@googlemail.com>
Reviewed-by: Rahul Bedarkar <rahulbedarkar89@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

lz4: fix host build with glibc older than 2.17

glibc before 2.17 require the -lrt linker option for clock_gettime(). We
do not support pre 2.17 glibc for the target anymore, but there are
still host platforms with these versions. Add upstream patch to skip use
of clock_gettime() with older glibc versions.

Should fix:
http://autobuild.buildroot.net/results/bc5/bc5d6447ab16a61d9dcf56723106f0b107826ef4/
http://autobuild.buildroot.net/results/7dd/7dd47bef902845a0f7803f1a6c702ec154855858/
http://autobuild.buildroot.net/results/952/9523dbc767dc325befabeb9240c4009e78779c7d/

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

zeromq: needs NPTL

The recent zeromq version bump to 4.2.5 added a call to
pthread_condattr_setclock() which is only provided by the NPTL
implementation of pthreads. Add this dependency to zeromq and all
reverse dependencies recursively.

Fixes:
http://autobuild.buildroot.net/results/2d0/2d070074097270a1e6973bc19e2bded2a7023c49/
http://autobuild.buildroot.net/results/1ce/1ce6d6f0f6390f552954a09f7dc753e0baa86675/
http://autobuild.buildroot.net/results/97c/97c5b9f3bc2c94f8476e884b9a3163bbf2dad8d4/

Cc: Asaf Kahlon <asafka7@gmail.com>
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

prosody: add dependency on shared libraries

The prosody package uses several Lua packages that require shared libraries.
Therefore, prosody must require shared libraries too.

Fixes: http://autobuild.buildroot.net/results/b126bce0bbd111f021ed5be7c8773fef2a8f8579
Signed-off-by: Danomi Manchego <danomimanchego123@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

DEVELOPERS: remove Philipp Claves' entry

The email address of Philipp Claves bounces.

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

package/libcec: disable build infos

Build infos can break the build due to missing terminating " character.

Fixes:
http://autobuild.buildroot.net/results/3b1/3b1182783fc958dfed96c6b1c097993662fc7308

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Cc: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

grub2: force -fno-stack-protector in CPPFLAGS

In commit 2a27294e9ade6130a12ced9a1f152c51431a870e ("grub2: force
-fno-stack-protector in CFLAGS"), a fix was made to the grub2 package
to make it build properly even when SSP support is enabled.

However, commit 20a4583ebf7fe97ea22a1ea11621dd44a8114ca5 ("security
hardening: add RELFO, FORTIFY options") reworked how SSP options are
passed, and they are now passed in CPPFLAGS instead of CFLAGS, making
the fix introduced by 2a27294e9ade6130a12ced9a1f152c51431a870e no
longer operating.

This commit will force no-stack-protector in CPPFLAGS instead of
CFLAGS.

Fixes bug #10961.

Signed-off-by: Tarek El-Sherbiny <tarek_el-sherbiny@waters.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

support/testing: add test for syslog-ng

Check:
 - the daemon is started;
 - a message is collected;
 - the daemon does not issue a warning message on startup.

When the .conf file version does not match the package version a warning
message shows up on serial on every boot. This message is generated by
syslog-ng before it is running, so it is not logged to
/var/log/messages. So in order to test the message is generated, restart
the server. It makes the message appears on /var/log/messages (since the
server is already running) where its existence can be easily tested
using grep.

Signed-off-by: Ricardo Martincoski <ricardo.martincoski@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

syslog-ng: bump version header in conf file to 3.10

Remove a runtime warning message about configuration file being too old.

Do the same as commit 3dad25466d "syslog-ng: Bump version header in conf
file to 3.9". Package version of syslog-ng is 3.10.1, so bump version
number in syslog-ng.conf to 3.10.

Also add a comment to avoid the same warning message reappears when the
package is bumped.

Signed-off-by: Ricardo Martincoski <ricardo.martincoski@datacom.ind.br>
Cc: Chris Packham <judge.packham@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

download/git: be sure we do fetch something

The different versions of git will behave in different ways when
fetching remote references, as summarised by the table below:

              |     ancient git         | new git
--------------------------------------------------------------------
git fetch     | fetch all refs but tags | fetches all refs but tags
git fetch -t  | fetches only tags       | fetch all refs and tags

(git-fetch may still fetch tags, but only if reachable from a branch)

So, to cover all the bases, we do a simple fetch, to be sure we have
branches, followed by the existing fetch -t, to get extra tags.

Fixes:
    http://autobuild.buildroot.net/results/0a2/0a238a7f55ea56c33b639ad03ed5796143426889/build-end.log

Reported-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

toolchain-external-custom: allow specifying relative path to binaries

There are cases where a downloaded toolchain doesn't have its binaries
placed directly in a "bin" subfolder (where BuildRoot currently looks
for them).

A common example is the official Raspberry Pi Toolchain
(https://github.com/raspberrypi/tools), which has its binaries in
"arm-bcm2708/arm-linux-gnueabihf/bin".

This commit introduces BR2_TOOLCHAIN_EXTERNAL_REL_BIN_PATH that defaults
to "bin" and can be changed as needed.

Signed-off-by: Calin Crisan <ccrisan@gmail.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
[Thomas: rework a bit how TOOLCHAIN_EXTERNAL_REL_BIN_PATH is defined.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

lz4: bump to version 1.8.1.2

Use lib and lz4 make targets to avoid build of the manual generator, and
the examples. Both of them use the target compiler to build host tools.

Add license files hashes.

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

multicat: bump to version 2.2

Drop upstream patches.

Add upstream patch fixing build with musl libc.

Add license file hash.

Set LDLIBS to TARGET_LDFLAGS for TARGET_LDFLAGS to be added to the link
commands. Makefile does not use the LDFLAGS environment variable. This
fixes static linking.

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

dhcpc: bump to version 7.0.3

Use sha256 checksum from *.distinfo.asc file.

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

tslib: update to 1.16

For those who wonder what happend, as usual, there's a changelog summary:
https://github.com/kergoth/tslib/releases

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

download/git: ensure we have a sane repository

There are cases where a repository might be broken, e.g. when a previous
operation was killed or otherwise failed unexpectedly.

We fix that by always initialising the repository, as suggested by
Ricardo. git-init is safe on an otherwise-healthy repository:

    Running git init in an existing repository is safe. It will not
    overwrite things that are already there. [...]

Using git-init will just ensure that we have the strictly required files
to form a sane tree. Any blob that is still missing would get fetched
later on.

Reported-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Reported-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Acked-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Tested-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

download/git: ensure we always work in the expected repository

git always look directories up until it finds a repository. In case
the git cache is broken, it may no longer be identified as a repository,
and git will look higher in the directories until it finds one.

In the default conditions, this would be Buildroot's own git tree
(because DL_DIR is a subdir of Buildroot), but in some situations may
very well be any repository the user has Buildroot in, like a
br2-external tree...

So, we force git to use our git cache and never look elsewhere, as
Suggested by Ricardo.

Use GIT_DIR, as it has been there for ages now, while --git-dir was
only introduced later (even if most distros ship an later version),
as suggested by Arnout.

Also fix the one call to git that was not using the wrapper.

Reported-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Acked-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Tested-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

pixiewps: add missing dependency on threads

Fixes:

  http://autobuild.buildroot.net/results/0c8d0eeb693ac29739845f489caffc2f9fa3b262/

Signed-off-by: Guillaume W. Bres <guillaume.bressaix@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

scanpypi: add support for the new PyPI infrastructure

https://pypi.python.org URL has been changed to https://pypi.org.

Package's JSON object now contains sha256 checksum, so use it
instead of locally computed one. Change comments in the hash
file accordingly.

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

package/wmctrl: x-includes and x-libraries must be set for cross-compiling

set x-includes and x-libraries configure option for cross-compiling.
wmctrl can use poisoned paths if these options are not passed to
configure script.

Fixes:
http://autobuild.buildroot.net/results/7e5/7e5cba8a5650a00e10d1a5528f38cb2bf772aee1

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Cc: Jérôme Pouiller <jezz@sysmic.org>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

nmap: fix check-package warning

Fixes:

package/nmap/nmap.mk:88: expected indent with tabs
93 lines processed
1 warnings generated

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

package/python-xlrd: fix check-package warnings on Config.in.host

Fixes:

package/python-xlrd/Config.in.host:4: help text: <tab><2 spaces><62 chars> (http://nightly.buildroot.org/#writing-rules-config-in)

https://gitlab.com/buildroot.org/buildroot/-/jobs/63334884

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

Config.in: fix check-package warning

Fixes:

Config.in:203: help text: <tab><2 spaces><62 chars> (http://nightly.buildroot.org/#writing-rules-config-in)
Config.in:204: help text: <tab><2 spaces><62 chars> (http://nightly.buildroot.org/#writing-rules-config-in)

https://gitlab.com/buildroot.org/buildroot/-/jobs/63334884

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

mariadb: bump version to 10.1.32

Release notes: https://mariadb.com/kb/en/mariadb-10132-release-notes/
Changelog: https://mariadb.com/kb/en/mariadb-10132-changelog/

Signed-off-by: Ryan Coe <bluemrp9@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/wayland: Bump version to 1.15.0

All hashes provided by upstream.

Updated Yann's patch:
 - Took v2 from https://patchwork.freedesktop.org/patch/143038/
 - Rebased on top of v1.15

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/weston: Bump version to 4.0.0

Added all hashes provided by upstream.

Yann's patch was integrated in upstream, see:
https://cgit.freedesktop.org/wayland/weston/commit/?id=fa41bdfbc

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

valgrind: bump version to 3.13.0

- update site
- add license hash
- remove 0004-Fixes-for-musl-libc.patch (applied upstream, see [1]) and
  corresponding autoreconf option

[1] https://sourceware.org/git/?p=valgrind.git;a=commit;h=f8f1e9cb434cc97ce16586437407a6ac7e6a74a7

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

usb_modeswitch: set CXX to false when C++ is missing

Similar to the openocd fix in commit 5966e2dc54 (package/openocd: fix
fallout after no-C++ fixups) the jimctl that is bundled with
usb_modeswitch also wants to find a binary. This broke with commit
4cd1ab158 (core: alternate solution to disable C++). Revert to 'false'
instead of 'no' here as well.

Fixes:
http://autobuild.buildroot.net/results/b4d/b4d4e925763cb6558af915f9781afe07fc557ebc/
http://autobuild.buildroot.net/results/61b/61b9dc2148df2e8fd0b30e62aedbfd30bb755e19/
http://autobuild.buildroot.net/results/468/468a7e6b049f159fbb4e79d3a12b53ca890f6933/

Cc: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

perl: bump to version 5.26.2

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

qt5: qt5websockets: install missing qml modules

QML modules for QtWebSockets are located in /usr/qml/QtWebSockets since
Qt 5.5.

/usr/qml/Qt/WebSockets still exists for compatibility reasons, but
it contains only a qmldir file which points to ../../QtWebSockets/.

see also: http://doc.qt.io/qt-5.6/qtwebsockets-qmlmodule.html

Signed-off-by: Manfred Schlaegl <manfred.schlaegl@ginzinger.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

qt53d: install missing QML modules, plugins and examples

Some files were missing on the first build of qt53d but added later:
- by qt5base for the plugins because it copies the whole /usr/lib/qt/plugins
  directory
- by qt5declarative for the QML modules because it copies the whole
  /usr/qml directory

Also, the qt53d examples were not installed if
BR2_PACKAGE_QT5BASE_EXAMPLES was set.

Signed-off-by: Romain Reignier <rom.reignier@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

micropython: bump version to v1.9.3

(With correct formatting now)

The build step now requires building mpy-cross first.

Signed-off-by: Blomme, Maarten <Maarten.Blomme@flir.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

legacy: remove options made legacy 5 years ago

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

package/swupdate: default website have a new API

2018.03 introduce a new website with Websocket asynchronous
communication[1]

[1] https://github.com/sbabic/swupdate/blob/master/doc/source/mongoose.rst

Signed-off-by: Julien Corjon <corjon.j@ecagroup.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

linux-firmware: support paths with subdirs in LINUX_FIRMWARE_DIRS

Following commit cdb7876f63a253254b86637aaac8b255a426b60a
("linux-firmware: add Qualcomm video accelerator and GPU firmware
options"), LINUX_FIRMWARE_DIRS can now contain entries with subdirs,
such as "qcom/venus-1.8". Such entries are currently not supported by
the logic in LINUX_FIRMWARE_INSTALL_DIRS.

Indeed the current logic consists in:

 - Removing the destination directory
 - Copying to the destination directory

Unfortunately, the last step fails if a component of the path leading
to the destination directory does not exist. To overcome this, we
simply add the appropriate mkdir before the copy.

Fixes:

  http://autobuild.buildroot.net/results/54cbca90742faf486e59d95b95a0b0494e47b385/

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

alsa-lib: update license file hash

FSF address update.

Fixes:
http://autobuild.buildroot.net/results/1b9/1b9ec9cc0f22a58ad37f21be1b6eb4cc064b6dbc/
http://autobuild.buildroot.net/results/d32/d32ae8b67c1aade7225c42a4685dc62fdaa0b0e3/
http://autobuild.buildroot.net/results/d9b/d9b8e918fa4a1f7a9470473f82bb21b3dce1acc4/

Cc: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

upmpdcli: don't test if the binary exists in the init script

The test doesn't make sense. It just exits without any error if the
binary doesn't exist, which is silly.

Replace the DAEMON variable, which was used only once, by the full path
of the binary file.

Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

tvheadend: don't test if the binary exists in the init script

The test doesn't make sense. It just exits without any error if the
binary doesn't exist, which is silly.

Replace the DAEMON variable, which was used only once, by the full path
of the binary file.

Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

triggerhappy: don't test if the binary exists in the init script

The test doesn't make sense. It just exits without any error if the
binary doesn't exist, which is silly.

Replace the DAEMON variable, which was used only once, by the full path
of the binary file.

Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

transmission: don't test if the binary exists in the init script

The test doesn't make sense. It just exits without any error if the
binary doesn't exist, which is silly.

Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

tpm2-abrmd: don't test if the binary exists in the init script

The test doesn't make sense. Buildroot installs both abrmd and its init
script as part of the same package. But if it ever happens for some
reason, the error message from start-stop-daemon should be pretty clear.

Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

tinyhttpd: don't test if the binary exists in the init script

The test doesn't make sense. It just exits without any error if the
binary doesn't exist, which is silly.

Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

squid: don't test if the binary exists in the init script

The test doesn't make sense. It just exits without any error if the
binary doesn't exist, which is silly.

Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

prosody: don't test if the binary exists in the init script

The test doesn't make sense. It just exits without any error if the
binary doesn't exist, which is silly.

Replace the DAEMON variable, which was used only once, by the full path
of the binary file.

Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

proftpd: don't test if the binary exists in the init script

The test doesn't make sense. It just exits without any error if the
binary doesn't exist, which is silly.

Replace the DAEMON variable, which was used only once, by the full path
of the binary file.

Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

openvpn: don't test if the binary exists in the init script

The test doesn't make sense. It just exits without any error if the
binary doesn't exist, which is silly.

Replace the DAEMON variable, which was used only once, by the full path
of the binary file.

Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

openntpd: don't test if the binary exists in the init script

The test doesn't make sense. It just exits without any error if the
binary doesn't exist, which is silly.

Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

ntp: don't test if the binary exists in the init script

The test doesn't make sense. It just exits without any error if the
binary doesn't exist, which is silly.

Replace the DAEMON variable, which was used only once, by the full path
of the binary file.

Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

nfs-utils: don't test if the binary exists in the init script

The test doesn't make sense. It just exits without any error if the
binary doesn't exist, which is silly.

Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

network-manager: don't test if the binary exists in the init script

The test doesn't make sense. It just exits without any error if the
binary doesn't exist, which is silly.

Replace the NETWORKMANAGER_BIN variable, which was used only once, by
the full path of the binary file.

Drop the now useless variables prefix, exec_prefix and sbindir.

Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

netsnmp: don't test if the binaries exist in the init script

The test doesn't make sense. It just exits without any error if the
binary doesn't exist, which is silly.

Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

netplug: don't test if the binary exists in the init script

The test doesn't make sense. It just exits without any error if the
binary doesn't exist, which is silly.

Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

neard: don't test if the binary exists in the init script

The test doesn't make sense. It just exits without any error if the
binary doesn't exist, which is silly.

Replace the DAEMON variable, which was used only once, by the full path
of the binary file.

Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

mpd: don't test if the binary exists in the init script

The test doesn't make sense. It just exits without any error if the
binary doesn't exist, which is silly.

Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

modem-manager: don't test if the binary exists in the init script

The test doesn't make sense. It just exits without any error if the
binary doesn't exist, which is silly.

Drop the MODEMMANAGER_BIN variable, which was never initialized.

Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

haveged: don't test if the binary exists in the init script

The test doesn't make sense. It just exits without any error if the
binary doesn't exist, which is silly.

Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

eudev: don't test if the binary exists in the init script

The test doesn't make sense. It just exits with error code 5 if the
binary doesn't exist, which is silly. Buildroot installs both udevd and
its init script as part of the same package. But if it ever happens for
some reason, the error message "/sbin/udevd: No such file or directory"
in the start case should be pretty clear.

Replace the UDEV_BIN variable, which was used only once, by the full
path of the binary file.

Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

dnsmasq: don't test if the binary exists in the init script

The test doesn't make sense. It just exits without any error if the
binary doesn't exist, which is silly.

Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

dmraid: don't test if the binary exists in the init script

The test doesn't make sense. It just exits without any error if the
binary doesn't exist, which is silly.

Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

dhcpcd: don't test if the binary exists in the init script

The test doesn't make sense. It just exits without any error if the
binary doesn't exist, which is silly.

Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

dbus: don't test if the binary exists in the init script

The test doesn't make sense. It just exits without any error if the
binary doesn't exist, which is silly.

Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

bind: don't test if the binary exists in the init script

The test doesn't make sense. It just exits without any error if the
binary doesn't exist, which is silly.

Signed-off-by: Carlos Santos <casantos@datacom.ind.br>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

sqlite: avoid use of unsupported -Ofast CFLAGS

sqlite3 refuses to be built with -ffast-math (a side effect of -Ofast) when it
falls back to implementing its own isnan() function.

sqlite3.c: In function ‘sqlite3IsNaN’:
sqlite3.c:28554:3: error: #error SQLite will not work correctly with the -ffast-math option of GCC.

To work around this, when -Ofast is used replace with -O3.

Signed-off-by: Joshua Henderson <joshua.henderson@microchip.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

php: remove Blackfin patch

And renumber remaining patches.

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

wayland: remove Blackfin patch

And renumber remaining patch.

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

p7zip: remove Blackfin patch

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

libsepol: remove Blackfin patch

And renumber remaining patches.

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

libsemanage: remove Blackfin patch

And renumber the remaining patch.

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

libpng: remove Blackfin patch

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