Merge branch 'next'

Kickoff 2021.11 cycle

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

docs/website/news.html: add 2021.08 announcement link

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

Update for 2021.08

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

package/xvisor: fix build with gcc 11

Fix the following build failures with gcc 11:

/home/buildroot/autobuild/instance-3/output-1/host/lib/gcc/arm-buildroot-linux-gnueabihf/11.1.0/../../../../arm-buildroot-linux-gnueabihf/bin/ld: /home/buildroot/autobuild/instance-3/output-1/build/xvisor-0.3.0/build/drivers/drivers.o: in function `psmouse_extensions':
/home/buildroot/autobuild/instance-3/output-1/build/xvisor-0.3.0/drivers/input/mouse/psmouse-base.c:783: undefined reference to `lifebook_detect'
/home/buildroot/autobuild/instance-3/output-1/host/lib/gcc/arm-buildroot-linux-gnueabihf/11.1.0/../../../../arm-buildroot-linux-gnueabihf/bin/ld: /home/buildroot/autobuild/instance-3/output-1/build/xvisor-0.3.0/drivers/input/mouse/psmouse-base.c:783: undefined reference to `alps_detect'
/home/buildroot/autobuild/instance-3/output-1/host/lib/gcc/arm-buildroot-linux-gnueabihf/11.1.0/../../../../arm-buildroot-linux-gnueabihf/bin/ld: /home/buildroot/autobuild/instance-3/output-1/build/xvisor-0.3.0/drivers/input/mouse/psmouse-base.c:783: undefined reference to `ps2pp_init'
/home/buildroot/autobuild/instance-3/output-1/host/lib/gcc/arm-buildroot-linux-gnueabihf/11.1.0/../../../../arm-buildroot-linux-gnueabihf/bin/ld: /home/buildroot/autobuild/instance-3/output-1/build/xvisor-0.3.0/drivers/input/mouse/psmouse-base.c:783: undefined reference to `trackpoint_detect'
/home/buildroot/autobuild/instance-3/output-1/host/lib/gcc/arm-buildroot-linux-gnueabihf/11.1.0/../../../../arm-buildroot-linux-gnueabihf/bin/ld: /home/buildroot/autobuild/instance-3/output-1/build/xvisor-0.3.0/drivers/input/mouse/psmouse-base.c:783: undefined reference to `fsp_detect'

Fixes:
 - http://autobuild.buildroot.org/results/69062b9c80567d135edd48890165e69881cf7295

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

package/python-pyudev: fix build with python 2

Build with python 2 is broken since bump to version 0.22.0 in commit
0adb141d342707ca9a478f57f187e38d5bb716f2:

error:   File "/usr/lib/python2.7/site-packages/pyudev/_ctypeslib/utils.py", line 54
    lib = cdll.LoadLibrary(f'lib{name}.so')
                                         ^
SyntaxError: invalid syntax

Fixes:
 - http://autobuild.buildroot.org/results/8b35ca6910dfd881953968f8d88ac842d57c9262

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

package/wlroots: fix uclibc build

Fix the following uclibc build failure raised since at least bump to
version 0.11.0 in commit 0bc9c89612cc1f41f9a64f6e889f8bcd8a871e30:

In file included from ../include/wlr/types/wlr_data_device.h:13,
                 from ../types/data_device/wlr_drag.c:7:
../include/wlr/types/wlr_seat.h:221:18: error: field 'last_event' has incomplete type
  221 |  struct timespec last_event;
      |                  ^~~~~~~~~~

Fixes:
 - http://autobuild.buildroot.org/results/3501ceb4290638b2f6d70aaa4d8ce74feec3a525

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>

package/qt5/qt5base: fix build on riscv32

Fix the following build failure on riscv32:

In file included from thread/qmutex_linux.cpp:45,
                 from thread/qmutex.cpp:804:
thread/qfutex_p.h: In function 'int QtLinuxFutex::_q_futex(int*, int, int, quintptr, int*, int)':
thread/qfutex_p.h:116:30: error: '__NR_futex' was not declared in this scope; did you mean '_q_futex'?
  116 |         int result = syscall(__NR_futex, addr, op | FUTEX_PRIVATE_FLAG, val, val2, addr2, val3);
      |                              ^~~~~~~~~~
      |                              _q_futex

Fixes:
 - http://autobuild.buildroot.org/results/ffedfc000029072d5d724e98ab4551fe973658ce

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>

package/gstreamer1/gstreamer1-mm: fix build with libglib 2.68

Fix the following build failure raised since bump of libglib2 to version
2.68.1 in commit c72524fb1b1c6b7dd1b1f511eea0fdf5a39dd3a0:

../gstreamer/gstreamermm/register.h: In function 'GType Gst::register_mm_type(const gchar*)':
/home/buildroot/autobuild/run/instance-2/output-1/host/bin/../arm-buildroot-linux-gnueabihf/sysroot/usr/include/glib-2.0/glib/gatomic.h:117:19: error: argument 2 of '__atomic_load' must not be a pointer to a 'volatile' type
  117 |     __atomic_load (gapg_temp_atomic, &gapg_temp_newval, __ATOMIC_SEQ_CST); \
      |     ~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Fixes:
 - http://autobuild.buildroot.org/results/1c75cdcc183642fd4c15d56825848b83f2ad11a5

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>

package/netsniff-ng: fix static build of mausezahn

The configure script fails to detect libpcap in static build because it
does not take into account the libnl dependency on link. As a result the
configure script silently disables mausezahn build even when
BR2_PACKAGE_NETSNIFF_NG_MAUSEZAHN is enabled. Add upstream patch to use
pkg-config for libpcap link flags.

Cc: Joris Lijssens <joris.lijssens@gmail.com>
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/python3: security bump to version 3.9.7

Fixes the following security issues:

- bpo-42278: Replaced usage of tempfile.mktemp() with TemporaryDirectory to
  avoid a potential race condition.

- bpo-41180: Add auditing events to the marshal module, and stop raising
  code.__init__ events for every unmarshalled code object.  Directly
  instantiated code objects will continue to raise an event, and audit event
  handlers should inspect or collect the raw marshal data.  This reduces a
  significant performance overhead when loading from .pyc files.

- bpo-44394: Update the vendored copy of libexpat to 2.4.1 (from 2.2.8) to
  get the fix for the CVE-2013-0340 “Billion Laughs” vulnerability.  This
  copy is most used on Windows and macOS.

- bpo-43124: Made the internal putcmd function in smtplib sanitize input for
  presence of \r and \n characters to avoid (unlikely) command injection.

https://www.python.org/downloads/release/python-397/

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

package/openmpi: fix fortran build

Fix the following build failure with fortran raised since bump to
version 4.0.0 in commit 366e7f1ecb4343ef8b18aef2a11c49055cc89abe and
https://github.com/open-mpi/ompi/commit/99730f798b7aef4c87d282aa24b98916fea45bb6:

checking size of Fortran type(test_mpi_handle)... (cached) 4
checking alignment of Fortran type(test_mpi_handle)... configure: error: Can not determine alignment of type(test_mpi_handle) when cross-compiling

Fixes:
 - http://autobuild.buildroot.org/results/86ffde2f67ffc0bfaeebe72fe742a5c241bc580b

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

package/nginx: override endianness

Though several cross-compilation patches exist in buildroot's nginx
package dir they do not seem to address endianness.

The test program generated by the configure script compiles but fails
to run (as it is built for another architecture) but the script does
not distinguish between the failure to run the program and an
indication of certain endianness.  As such the fallback of big-endian
is used.  This setting then causes http2 headers (anything not in the
static dictionary) to come out as undecipherable trash on 64bit
targets (see ngx_http_v2_huff_encode_buf()).

This commit includes a patch to the configure script to allow a
`--force-endianness=big|little` flag as well as setting that flag in
buildroot's package makefile.

Signed-off-by: Nevo Hed <nhed+buildroot@starry.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

Update for 2021.08-rc3

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

package/linux-pam: find libxcrypt through pkg-config

Find libxcryt through pkg-config to avoid the following build failure:

/home/buildroot/autobuild/run/instance-3/output-1/host/opt/ext-toolchain/bin/../lib/gcc/riscv64-buildroot-linux-musl/10.2.0/../../../../riscv64-buildroot-linux-musl/bin/ld: .libs/passverify.o: in function `.L30':
passverify.c:(.text+0x368): undefined reference to `crypt_checksalt'

Fixes:
 - http://autobuild.buildroot.org/results/20b14e222b35c2d1269960075832b784ba81aa1a

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/fetchmail: bump version to 6.4.21

Release notes:
https://sourceforge.net/p/fetchmail/mailman/message/37333073/

"It contains the security fix for CVE-2021-36386 of 6.4.20, and fixes
a regression/a bug that causes log message truncation/run-together
prominently visible with --logfile that was introduced into 6.4.20."

Updated note for CVE-2021-36386:
https://sourceforge.net/p/fetchmail/mailman/message/37333078/

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

package/php: security bump version to 8.0.10

Quoting https://www.php.net/
"This is a security fix release."

Changelog: https://www.php.net/ChangeLog-8.php#8.0.10

CVE-ID were not mentioned in any of the fixed bugs.

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

package/samba4: bump version to 4.14.7

Release notes of this bugfix release:
https://www.samba.org/samba/history/samba-4.14.7.html

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

package/c-ares: add C_ARES_CPE_ID_VENDOR

cpe:2.3:a:c-ares_project:c-ares is a valid CPE identifier for this
package:

  https://nvd.nist.gov/products/cpe/search/results?namingFormat=2.3&keyword=cpe%3A2.3%3Aa%3Ac-ares_project%3Ac-ares

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/c-ares: fix build with latest nodejs

Fix the following build failure with nodejs raised since bump to version
12.22.5 in commit 7038b029d8c8774eca6d7888d6642d7e84ff5165:

../src/cares_wrap.cc:42:11: fatal error: ares_nameser.h: No such file or directory
   42 | # include <ares_nameser.h>
      |           ^~~~~~~~~~~~~~~~

Fixes:
 - http://autobuild.buildroot.org/results/a0f867d5e765fc1aa052de5e53ed350b3b20743f

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/c-ares: security bump to version 1.17.2

- NodeJS passes NULL for addr and 0 for addrlen to
  ares_parse_ptr_reply() on systems where malloc(0) returns NULL. This
  would cause a crash.
- If ares_getaddrinfo() was terminated by an ares_destroy(), it would
  cause a crash
- Crash in sortaddrinfo() if the list size equals 0 due to an unexpected
  DNS response
- Expand number of escaped characters in DNS replies as per RFC1035 5.1
  to prevent spoofing follow-up
- Perform validation on hostnames to prevent possible XSS due to
  applications not performing valiation themselves

https://c-ares.haxx.se/changelog.html#1_17_2

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/json-for-modern-cpp: bump version to 3.10.2

License file year changed to 2021.

Signed-off-by: Michael Nosthoff <buildroot@heine.tech>
[Arnout: update license hash]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

toolchain: handle toolchains with multiple ld*.so.* files

Some 3rd party vendor toolchains have multiple files which match
these glob patterns.  In this case, the shell script failed.
Switching to use find and xargs solves the issue.

Signed-off-by: Jonah Petri <jonah@petri.us>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/python-aioprocessing: new package

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
[Arnout: add to DEVELOPERS]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/rwmem: new package

rwmem is small tool to read & write device registers. Some of the
features include:

- support mmaped and i2c devices
- addressing with 8/16/32/64 bit addresses
- accessing 8/16/32/64 bit memory locations
- little and big endian addressess and accesses
- bitfields
- address ranges
- register description database

Python bindings are disabled for now.

Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/qt5/qt5opcua: new package

The Qt OPC UA module implements a Qt API to interact with OPC UA on
top of a 3rd party OPC UA stack.
The default is open62541, which is bundled by qt5opcua in version 1.0,
so we dont need to provide/depend on br's own open62541 package.

Another dependency is mbedtls, but it's optional.

Signed-off-by: Andreas Naumann <anaumann@ultratronik.de>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/vuejs-router: new package

Vue Router is the official router for Vue.js.

Signed-off-by: Thomas Claveirole <thomas.claveirole@green-communications.fr>
[Arnout: use comment instead of submenu]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

support/testing/tests: add test for check_bin_arch

This tests build a bogus package that installs a binary built for the
host architecture into $(TARGET_DIR), which should cause a build
failure, at least as long as the host architecture isn't ARM.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Herve Codina <herve.codina@bootlin.com>
[yann.morin.1998@free.fr: drop uneeded subprocess import to fix flake8]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

support/testing/infra: add log_file_path() function

Some tests will need to grep through the build log to verify that some
features are working are expected. In order to allow them to open the
build log, we provide a new function called log_file_path(), which
returns the path to the log file if available.

We also use this function in open_log_file().

Note that open_log_file() cannot be used directly to grep through the
log file at the end of a build: because it opens in "a+" mode, it
greps starting from the end of the file.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

docs/manual: fix typo in name of a qmake-package variable

Since commit 39d334faa5c1 (package/pkg-qmake: add <pkg>_SYNC_QT_HEADERS
support), the qmake-package infra recognises said variable but the
manual has the wrong variable name, which is missing the "_QT" part.

We fix that by amending the manual to document the proper variable name.

Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/pkg-qmake.mk: Move QT5_QT_CONF_FIXUP to post-prepare hook

QT5_QT_CONF_FIXUP tweaks files for per-package directory build.
This is typically the kind of operation expected to be in
post-prepare hook.

Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

qt5: Fix sporadic build failure during top-level parallel build

When using top level parallel build, independent qt5 packages may be
built in parallel. Because of their staging dirs being hardlinked, they
all use the same qt.conf file to manipulate during configure, while
another qt5 package might already use it. This leads to weird build failures
because the folders qmake is using are diverted in erratic ways.
Fix this by actually recreating a non-shared qt.conf file for every package.

Signed-off-by: Andreas Naumann <anaumann@ultratronik.de>
Reviewed-by: Herve Codina <herve.codina@bootlin.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/stunnel: bump to version 5.60

This release includes the following changes:

- New features
  - New 'sessionResume' service-level option to allow or disallow session resumption
  - Added support for the new SSL_set_options() values.
  - Download fresh ca-certs.pem for each new release.
- Bugfixes
  - Fixed 'redirect' with 'protocol'. This combination is not supported by 'smtp', 'pop3' and 'imap' protocols.
  - Enforced minimum WIN32 log window size.
  - Fixed support for password-protected private keys with OpenSSL 3.0 (thx to Dmitry Belyavskiy).
  - Added missing TLS options supported in OpenSSL 1.1.1k.

Signed-off-by: Pierre-Jean Texier <texier.pj2@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/libqmi: bump to version 1.30.0

This release introduces meson build system support.

As meson files are only available in git, switch the location to
https://gitlab.freedesktop.org.

Add support for bash-completion.

Add an upstream patch fixing the compilation without gobject-intorspection.

Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/mesa3d-demos: use BR2_PACKAGE_MESA3D_GBM

Use BR2_PACKAGE_MESA3D_GBM introduced by commit
b6c051acf787c804e732bc58ba8d7e440701a168

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/mpv: use BR2_PACKAGE_MESA3D_GBM

Use BR2_PACKAGE_MESA3D_GBM instead of BR2_PACKAGE_MESA3D_OPENGL_EGL as
GBM can also be provided by a DRI driver

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/sdl2: use BR2_PACKAGE_MESA3D_GBM

BR2_PACKAGE_MESA3D_GBM should be used instead of
BR2_PACKAGE_MESA3D_OPENGL_EGL as spotted by Arnout in review of
https://patchwork.ozlabs.org/project/buildroot/patch/20210821093908.677218-1-fontaine.fabrice@gmail.com/

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/luv: bump to version 1.41.1-0

Signed-off-by: Jörg Krause <joerg.krause@embedded.rocks>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

utils/test-pkg: add gitlab-ci support

The gitlab-ci support in test-pkg allows to parallelize the test-pkg
work into several gitlab jobs. It's much faster than local serialized
testing.

To trigger this, a developer will have to add, in the latest commit of
their branch, a token on its own line, followed by a configuration
fragment, e.g.:

    test-pkg config:
    SOME_OPTION=y
    # OTHER_OPTION is not set
    SOME_VARIABLE="some value"

This configuration fragment is used as input to test-pkg.

To be able to generate one job per test to run, we need the list of
tests in the parent pipeline, and the individual .config files (one per
test) in the child pipeline. We use the newly-introduced --prepare-only
mode to test-pkg, and collect all the generated .config files as
artefacts; those are inherited in the child pipeline via the
"needs::pipeline" and "needs::job" directives. This is a bit tricky,
and is best described by the Gitlab-CI documentation [0].

We also list those .config files to generate the actual list of jobs to
run in the child pipeline.

Notes:
  - if the user provides an empty fragment, this is considered an error:
    indeed, without a fragment (and the package name), there is no way
    to know what to test;
  - if that fragment yields an empty list of tests, then there is
    nothing to test either, so that is also considered an error.

[0] https://docs.gitlab.com/ee/ci/yaml/README.html#artifact-downloads-to-child-pipelines

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Cc: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
[yann.morin.1998@free.fr:
  - split the change to test-pkg to its own patch
  - generate the actual yml snippet in support/scripts/generate-gitlab-ci-yml,
    listing the .config files created by test-pkg
  - some code-style-candies...
]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

utils/test-pkg: add mode to only prepare .config files

Currently, running test-pkg is only done locally on the developers
machine.

In a follow up commit, we'll add the possibility to run test-pkg in a
gitlab-ci pipeline and, to speed things up, with one job per buildable
configuration.

As such, we will need to make sure that test-pkg only prepares the
configurations, and that it does not build them.

Add such a mode, with a new option, --prepare-only

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

package/log4cxx: add libesmtp optional dependency

libesmtp is an optional dependency (disabled by default) since version
0.10.0 and
https://github.com/apache/logging-log4cxx/commit/8a35b429c6b1bc73635d9b6db482523619199a90

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

configs/stm32f469_disco_xip_defconfig: alternative defconfig for XIP

Update STM32F469-disco configuration files to operate with new kernel.

Result of make tinyconfig was taken as a starting point to fit kernel
into flash memory.
Current setup kernel + rootfs fits in 1.6MB on-chip flash memory

Fixes:
    - Move kernel to new flash bank due to growth of dtb size
    - Fix kernel start address in bootloader
    - Remove outdated path which doesn't affect normal operation mode

For better binary size optimization gcc LTO is turned on.

Signed-off-by: Yauheni Saldatsenka <eugentoo@gmail.com>
[Arnout:
 - squash 3 patches into 1;
 - remove unused dts file;
 - move linux/linux.config to linux-xip.config;
 - add a sentence to readme to say SD card is not needed.
]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/owfs: remove Python sysconfigdata fixup

This fixup is done at pkg-generic level for all packages.
So, it is no more needed in owfs package.

Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Reviewed-by: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/pkg-generic.mk: move python fixup to generic package infrastructure

Fixing _sysconfigdata*.{py,pyc} was previously done by python package
infrastructure. Some packages use python stuff without using python
package infrastructure.
These packages perform overwrites and need the specific python fixup
to fix them.

In order to be sure to fix all of these packages, the python fixup
is moved to the generic package infrastructure and applied to all
packages.
This follows the same principle as for the .la libtool files fixup.

Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/pkg-python: invalidate precompiled _sysconfigdata*.pyc

For per-package directories, we fixup the _sysconfigdata*.py files, so
that they get proper path pointing to the current package's direcotry
structure.

However, the corresponding, pre-compiled blobs _sysconfigdata*.pyc were
left around, and thus are inconsistent with their source. They might
also be regenerated when a package would install a python module; this
regeneration would trigger the soon-to-be-introduced overwrite
detection.

This commit simply removes _sysconfigdata*.pyc files; they will anyway
be regenerated by the PYTHON{,3}_CREATE_PYC_FILES target finalize hooks.
This is an efficient way to guarantee the consistency between the source
and precompiled versions, and to not trigger the overwrite detection.

Signed-off-by: Herve Codina <herve.codina@bootlin.com>
[yann.morin.1998@free.frs: reword the commit log]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/pkg-python: fix PKG_PYTHON_FIXUP_SYSCONFIGDATA

Find used by PKG_PYTHON_FIXUP_SYSCONFIGDATA can fail if directories
are not present. This failure is silently ignored because find is
on the LHS of a pipe.

This commit fixes the find failure. HOST_DIR is used as the starting
point and the search is filtered on the expected directories.

This commit also adds -print0 and the $(Q) verbosity flag as minor
changes.

Signed-off-by: Herve Codina <herve.codina@bootlin.com>
[yann.morin.1998@free.fr:
  - split long line with the two -path options
  - move "| xargs ..." onto its own line
]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/fakedate: avoid recursion if not first in the PATH

With BR2_REPRODUCIBLE, our 'fakedate' wrapper script will end up in
host/bin/date. Currently, it iterates over all the 'date' found in PATH,
until it finds one that is not the script itself, in an attempt to avoid
infinite recursion by calling itself again and again.

This heuristic works OK in Buildroot, because host/bin/ is first in the
PATH, and so that means the first entry in the PATH is skipped.

However, this is going to fail as soon as our wrapper is not the first
in the PATH. Indeed, in that situation, the current heuristic will stop
on the first 'date' in the PATH, as it is not the script itself, and
since our script was executed, that probably means the first 'date' was
itself a wrapper that ended up calling us. So, calling it again will
eventually trickle to calling us again, and thus creating the loop our
heuristic was made to avoid.

This situation currently does not occur in Buildroot, because host/bin/
is first, *and* we have no package that provide their own 'date' wrapper
during their build steps.

But when we generate an SDK with BR2_REPRODUCIBLE, then our wrapper
script will be in sdk/bin/, and there is no longer any guarantee this
comes first in the PATH, thus opening the possibility that another
buildsystem based on our SDK, but which has its own 'date' wrapper, will
trigger this infinite recursion.

We fix that by iterating, in reverse order, over all the 'date' we can
find in PATH, and when we find ourselves, then we know the one we found
in the iteration just before is the one that we should call.

'which -a' is old enough that we can expect it to be always available;
it has been present at least since Debian Squeeze, released 2011.

Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/xen: fix build with 64 bits time_t

Fix build of xen with 64 bites time_t:

/tmp/instance-0/output-1/build/xen-4.14.2/tools/qemu-xen/hw/input/virtio-input-host.c: In function 'virtio_input_host_handle_status':
/tmp/instance-0/output-1/build/xen-4.14.2/tools/qemu-xen/hw/input/virtio-input-host.c:198:28: error: 'struct input_event' has no member named 'time'
  198 |     if (gettimeofday(&evdev.time, NULL)) {
      |                            ^

Fixes:
 - http://autobuild.buildroot.org/results/136ce42f44bf48d3db4eda7b1548bf7ac1b97d51

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/ushare: fix NLS build

Commit c4e1a075101914f8e0d47f8bb9c06bea13ba0fd2 forgot to add
--enable-nls to patch resulting in the following build failure:

Unknown option "--enable-nls".

Fixes:
 - http://autobuild.buildroot.org/results/6ab2555b419355f01310f230fe612f2a3699bbfd

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/sdl2: drop patch

Drop patch (and so SDL2_RUN_AUTOGEN) as patch is already in version
2.0.16 which was applied to next by commit
c80fbf269c020c53a6140ffbf328d64674a8ada0

Fixes:
 - http://autobuild.buildroot.org/results/4f9026e94373fb31ef7e45f8b4fd55adc178013a

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/boost: fix sparc build

Fix the following build failure on sparc raised since bump to version
1.77.0 in commit d39d8f7cee9b4634e6f490020204f63fae419e8e:

./boost/predef/architecture/sparc.h:37:38: error: missing ')' in expression
   37 | #   if !defined(BOOST_ARCH_SPARC) && (defined(__sparcv9) || defined(__sparc_v9__)
      |                                      ^

Fixes:
 - http://autobuild.buildroot.org/results/c1f15e3a0cefb7e65e5288fb4564d59a3b06a0bd

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

support/testing: test_atf: bump the custom version to v2.5

This version bump is needed to pass the ATF test with
hardening option enabled (-fstack-protector-strong)

With the version v2.2, ATF fail due to undefined references:

./build/juno/release/bl2u/arm_tzc400.o: In function `arm_tzc400_setup':
arm_tzc400.c:(.text.arm_tzc400_setup+0x10): undefined reference to `__stack_chk_guard'
arm_tzc400.c:(.text.arm_tzc400_setup+0x18): undefined reference to `__stack_chk_guard'
arm_tzc400.c:(.text.arm_tzc400_setup+0xb8): undefined reference to `__stack_chk_guard'
arm_tzc400.c:(.text.arm_tzc400_setup+0xcc): undefined reference to `__stack_chk_fail'

Since commit ccac9a5bbbd7374187a0f0017101ece0c202851d, Buildroot no
longer forces ENABLE_STACK_PROTECTOR. However, we rely on the ATF build
system to handle it correctly, and this wasn't the case in v2.2.

Fixes: https://gitlab.com/buildroot.org/buildroot/-/jobs/1524842591
Signed-off-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/coreutils: Prevent overwriting of fakedate

When BR2_REPRODUCIBLE is set and host-coreutils needs to be built, the
fakedate script installed to 'host/bin/date' will be overwritten by
host-coreutils.

Besides, we do not need our host-coreutils for 'date' at all; we really
rely on the host system to provide it.

Unconditionally disable installing the 'date' binary in host-coreutils.

Note that we explicitly request only ln and realpath to be installed,
but the coreutils buildsystem does not strictly obey to that, as was
already noticed in 885e6fdb8a40 (package/coreutils: introduce a host
variant), which added that comment above HOST_COREUTILS_CONF_OPTS:

    # Explicitly install ln and realpath, which we *are* insterested in.
    # A lot of other programs still get installed, however, but disabling
    # them does not gain much at build time, and is a loooong list that is
    # difficult to maintain...

So, we also update that comment to explain why we still anyway disable
installation of 'date'.

Signed-off-by: Conrad Ratschan <conrad.ratschan@collins.com>
[yann.morin.1998@free.fr:
  - unconditionally disable installing date
  - extend comment and commit log to explain why we need
    --enable-no-install-program=date despite the existing
    --enable-install-program=ln,realpath
]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/mtd: add ubihealthd missing comment

Commit 42a3fee35e67d8d946f52a82bfa88ee6fd9f04b6 forgot to add comment on
headers 3.17+

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/busybox: add systemd service for telnetd

Signed-off-by: Sam Voss <sam.voss@collins.com>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
[yann.morin.1998@free.fr:
  - introduce BUSYBOX_INSTALL_TELNET_SERVICE
  - move _INSTALL_INIT_SYSTEMD alphabetically between openrc and sysv
  - drop the comment about Type=simple  (Arnout)
]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/mtd: make ubihealthd independent of ubifs

Since version 2.1.3 ubihealthd can be enabled without of ubifs-utils.

This also fixes usability of enabling BR2_PACKAGE_MTD_UBIHEALTHD.
BR2_PACKAGE_MTD_UBIFS_UTILS is a blind option. The only way to enable it
is to enable BR2_PACKAGE_MTD_MKFSUBIFS that selects it. ubihealthd
dependency on BR2_PACKAGE_MTD_UBIFS_UTILS makes enabling it unintuitive.

Cc: Markus Mayer <mmayer@broadcom.com>
Cc: Matt Weber <matthew.weber@collins.com>
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/ipmiutil: avoid the need for autoreconf

Build fails since bump to version 3.1.7 in commit
011f31ee244b7fe4621b3f69ca3d7e7b1cfdb5bb because config.h.in is older
than aclocal.m4:

make[1]: Entering directory '/tmp/instance-4/output-1/build/ipmiutil-3.1.7'
(CDPATH="${ZSH_VERSION+.}:" && cd . && autoheader)
/bin/bash: autoheader: command not found

Fixes:
 - http://autobuild.buildroot.org/results/2005af881726473f2cda176e90c1e41e4baea67c

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/nodejs: security bump to version 12.22.5

Fix CVE-2021-22931, CVE-2021-22940 and CVE-2021-22939:
https://nodejs.org/en/blog/vulnerability/aug-2021-security-releases

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

board/acmesystems/{aria, arietta}-g25: fix genimage.cfg

On the Aria and the Arietta AT91Bootstrap builds, the file name of the
bootloader embeds its version number, and the genimage configuration
needs this filename in order to build the boot filesystem image. Commit
0614f435a09e99c33f418a646eaa7647944c2f4c bumped the AT91Bootstrap
version of all acmesystems' boards but failed to update genimage.cfg
accordingly, which broke the builds. The Acqua board is not affected
by this issue.

Update the affected genimage.cfg with the correct filenames.

Fixes:
https://gitlab.com/buildroot.org/buildroot/-/jobs/1515521690
https://gitlab.com/buildroot.org/buildroot/-/jobs/1515521691
https://gitlab.com/buildroot.org/buildroot/-/jobs/1515521692
https://gitlab.com/buildroot.org/buildroot/-/jobs/1515521694

Signed-off-by: Edgar Bonet <bonet@grenoble.cnrs.fr>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/localedef: bump to version 2.34-9-g9acab0bba6a5a57323b1f94bf95b21618a9e5aa4

resync the version with glibc package.

Dropped 2 patches (fixes backported from previous releases),
rebased 2 which are kept (only line numbers changed).

Suggested-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Petr Vorel <petr.vorel@gmail.com>
[Arnout: resolve conflicts]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/glibc: bump version to 2.34-9-g9acab0bba6a5a57323b1f94bf95b21618a9e5aa4

Signed-off-by: Petr Vorel <petr.vorel@gmail.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

Merge branch 'master' into next

Conflicts:
package/go/go.hash
package/go/go.mk

Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/bubblewrap: bump to version 0.5.0

Not much new, but has a few interesting bug fixes. Release notes:

  https://github.com/containers/bubblewrap/releases/tag/v0.5.0

Signed-off-by: Adrian Perez de Castro <aperez@igalia.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/boost: bump version to 1.77.0

* drop 0001-fenv.patch, issue [0] was marked fixed since boost 1.64 (commit [1])
* drop upstreamed patch 0002
* filesystem now depends on boost-atomic
* math broke the build without always lockfree atomic ints, disable for now. reported at [2].

[0] https://svn.boost.org/trac/boost/ticket/11756
[1] https://github.com/boostorg/test/commit/cb2a1c2488dc77b9baf1b05781208a9639dfcd69
[2] https://github.com/boostorg/math/issues/673

Signed-off-by: Michael Nosthoff <buildroot@heine.tech>
[Arnout: improve propagation of reverse dependencies:
 - also update comment of cc-tool;
 - add boost-filesystem to reason of gnuradio, libcpprestsdk, uhd;
 - move dependency to arch-deps of mongodb]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

docs/website: use lore for ml search

the nabble.com link is dead and lore has a good search.
So use lore for the search form.

Signed-off-by: Michael Nosthoff <buildroot@heine.tech>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/mtd: ubihealthd needs kernel 3.17+

ubihealthd requires getrandom(2) that was introduced in kernel version
3.17. ubihealthd does not build when getrandom(2) is not detected, so
the following installation step fails.

Technically the dependency should also be on glibc version 2.25+. But we
have no way to depend on glibc versions of external toolchains.
Toolchain built with kernel headers older than 3.17 can build
ubihealthd, but it will fail at run-time. So this is a pretty close
approximation of the actual dependency.

Fixes:
http://autobuild.buildroot.net/results/2d42b0a626367e4051d0e2aadcce39e974fe09d4/
http://autobuild.buildroot.net/results/a2b6dbf707275e3f8262479c0650cfc7cb9abc8d/

Cc: Matt Weber <matthew.weber@collins.com>
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/ushare: fix build without msgfmt

Fix the following build failure without /usr/bin/msgfmt raised since the
addition of ushare in commit 74097fd659154499612f21fabeda4e3e7c8fdbfc:

make[3]: Entering directory `/home/buildroot/autobuild/run/instance-3/output-1/build/ushare-2.1/po'
/usr/bin/msgfmt -c --statistics -o fr.gmo fr.po
make[3]: /usr/bin/msgfmt: Command not found

To fix this build failure, set GMSGFMT to $(HOST_DIR)/bin/msgfmt and
don't build po files if NLS is disabled

Fixes:
 - http://autobuild.buildroot.org/results/9f6b5b8f38386135bacd2d8f6e97c1fea77bbe69

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/cjson: bump to version 1.7.15

Fixes:
- Fix potential core dumped for strrchr
- Fix null pointer crash in cJSON_CreateXxArray
- Fix several null pointer problems on allocation failure
- Fix a possible dereference of null pointer

https://github.com/DaveGamble/cJSON/releases/tag/v1.7.15

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

Revert "toolchain: handle toolchains with multiple ld*.so.* files"

This breaks some existing external toolchains. Since we're very close to
a release, don't try to fix it, but instead simply revert.

This reverts commit 6f911a17257d945eb3cb5c9c9a966f0e5fca0842.

Fixes: http://autobuild.buildroot.net/results/afe/afe44f4b6a3c53e5864cfb10b04529011e72cf5c/
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

configs/pc_x86_64_{efi, bios}_defconfig: update kernel to 4.19.204

gcc 10.x is now used by default but the kernel 4.18.10 used by
pc_x86_64_{efi,bios}_defconfig doesn't build with it.

Bump the kernel to 4.19.204 release that contains a lot of
fixes for newer gcc.

Fixes:
https://gitlab.com/kubu93/buildroot/-/jobs/1525741062
https://gitlab.com/kubu93/buildroot/-/jobs/1525741060

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

configs/beaglev_defconfig: update linux-headers to 5.13.

The beaglev kernel is based on the 5.13 branch, update
the expected linux-headers version to 5.13.

This has been wrong ever since the bump of the kernel version in commit
9a1bd7cc1ce512672a5d76baa86c449e36137052: the headers were bumped to
5.12 instead of 5.13.

Fixes:
https://gitlab.com/kubu93/buildroot/-/jobs/1525740895

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

package/alsa-utils: Fix alsa-plugins compatibility

Previously, alsa-plugins would not work if alsa-utils was installed
after it. This happened because:

1. alsa-plugins copies some files $(TARGET_DIR)/usr/share/alsa/alsa.conf.d
2. alsa-utils removes these files during installation ( rm -rf $(TARGET_DIR)/usr/share/alsa/;)

The `rm -rf` command was originally added as part of the fix for
https://bugs.buildroot.org/show_bug.cgi?id=1573 11 years ago.

The intention might have been to allow for unconfiguring some options
and then rebuilding alsa-utils. However, this is a scenario that does
not work anyway.

The simplest fix for the `alsa-plugins` compatibility issue appears to
be to remove the `rm -rf` command.

Signed-off-by: Gleb Mazovetskiy <glex.spb@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/libopenssl: security bump version to 1.1.1l

Signed-off-by: Michael Fischer <mf@go-sys.de>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

DEVELOPERS: drop Maxime Hadjinlian

Maxime has not been contributing to Buildroot for several years, so it
doesn't make sense to keep him in the DEVELOPERS file and make us
think that those packages are being maintained and to Cc: him on
patches affecting those packages.

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

toolchain: handle toolchains with multiple ld*.so.* files

Some 3rd party vendor toolchains have multiple files which match
these glob patterns.  In this case, the shell script failed.
Switching to use find and xargs solves the issue.

Signed-off-by: Jonah Petri <jonah@petri.us>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/libarchive: security bump to version 3.5.2

Fix CVE-2021-36976: libarchive 3.4.1 through 3.5.1 has a use-after-free
in copy_string (called from do_uncompress_block and process_block).

https://github.com/libarchive/libarchive/releases/tag/v3.5.2

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/belle-sip: add BELLE_SIP_CPE_ID_VENDOR

cpe:2.3:a:linphone:belle-sip is a valid CPE identifier for this package:

  https://nvd.nist.gov/products/cpe/search/results?namingFormat=2.3&keyword=cpe%3A2.3%3Aa%3Alinphone%3Abelle-sip

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/libffi: disable use of static exec trampolines

TestGst1Python test segfault since the libffi bump to 3.4.2.

Apply the same fix from Yocto [1] disabling static exec trampolines.

Fixes:
https://gitlab.com/kubu93/buildroot/-/jobs/1522848331

[1] http://git.yoctoproject.org/cgit/cgit.cgi/poky/commit/?id=dadfef3950fae4e93ce4c13ab91a2a7f41b3702e

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

package/linux-pam: add libxcrypt optional dependency

Add libxcrypt optional dependency and fix the following build failure
with libxcrypt and uclibc-ng raised since the addition of libxcrypt in
commit 464bbe26ff5fb9e5bfe26a26ea65c700b90598f5:

/home/buildroot/autobuild/instance-1/output-1/host/opt/ext-toolchain/bin/../lib/gcc/arm-buildroot-linux-uclibcgnueabihf/9.3.0/../../../../arm-buildroot-linux-uclibcgnueabihf/bin/ld: unix_chkpwd-passverify.o: in function `verify_pwd_hash':
passverify.c:(.text+0xab4): undefined reference to `crypt_checksalt'

Fixes:
 - http://autobuild.buildroot.org/results/65d68b7c9c7de1c7cb0f941ff9982f93a49a56f8

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/localedef: bump to version 2.33-46-gedfd11197ecf3629bbb4b66c5814da09a61a7f9f

resync the version with glibc package.

Remove upstream patches.

Rebase remaining patches for glibc 2.33.

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

package/python{3}-requests: allow idna 3.x to be installed on python 3.x

The tests.package.test_docker_compose.TestDockerCompose is broken
since the python-idna version bump to 3.0 because python-requests needs
python-idna < 3.0.

 # docker-compose up -d
 Traceback (most recent call last):
   File "/usr/lib/python3.9/site-packages/pkg_resources/__init__.py", line 583, in _build_master
   File "/usr/lib/python3.9/site-packages/pkg_resources/__init__.py", line 900, in require
   File "/usr/lib/python3.9/site-packages/pkg_resources/__init__.py", line 791, in resolve
 pkg_resources.ContextualVersionConflict: (idna 3.2 (/usr/lib/python3.9/site-packages), Requirement.parse('idna<3,>=2.5'), {'requests'})

 During handling of the above exception, another exception occurred:

 Traceback (most recent call last):
   File "/usr/bin/docker-compose", line 6, in <module>
     from pkg_resources import load_entry_point
   File "/usr/lib/python3.9/site-packages/pkg_resources/__init__.py", line 3252, in <module>
   File "/usr/lib/python3.9/site-packages/pkg_resources/__init__.py", line 3235, in _call_aside
   File "/usr/lib/python3.9/site-packages/pkg_resources/__init__.py", line 3264, in _initialize_master_working_set
   File "/usr/lib/python3.9/site-packages/pkg_resources/__init__.py", line 585, in _build_master
   File "/usr/lib/python3.9/site-packages/pkg_resources/__init__.py", line 598, in _build_from_requirements
   File "/usr/lib/python3.9/site-packages/pkg_resources/__init__.py", line 786, in resolve
 pkg_resources.DistributionNotFound: The 'idna<3,>=2.5' distribution was not found and is required by requests

Fixes:
https://gitlab.com/kubu93/buildroot/-/jobs/1522848327

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

support/testing: test_docker_compose: bump the kernel to 4.19.204

gcc 10.x is now used by default but the kernel 4.19 used by
test_docker_compose doesn't build with it.

Bump the kernel to 4.19.204 release that contains a lot of
fixes for newer gcc.

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

package/mpd: bump to version 0.22.10

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

package/kodi-pvr-hts: bump version to 8.4.0-Matrix

Changelog:
https://github.com/kodi-pvr/pvr.hts/blob/Matrix/pvr.hts/changelog.txt

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

support/testing: revert the last change of check_network()

check_network() must check the error code of the command
used to check the network configuration with the value
passed as argument "exitCode".

But this argument is ignored since this commit [1].

Revert the last change of check_network().

Fixes:
https://gitlab.com/kubu93/buildroot/-/jobs/1522848308
https://gitlab.com/kubu93/buildroot/-/jobs/1522848306

[1] afc1ed4d5152e0f6e724e6986a1d12c8001b94fe

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

package/fluidsynth: bump to version 2.2.2

For change log since v2.2.1, see:
- https://github.com/FluidSynth/fluidsynth/releases/tag/v2.2.2

./utils/test-pkg --package fluidsynth
6 builds, 2 skipped, 0 build failed, 0 legal-info failed

Signed-off-by: Julien Olivain <ju.o@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/libiio: bump to version 0.23

In versions > 0.21, the CMake script doesn't auto-enable features
according to what it detects in the environment; options have to be
explicitely enabled. Update the libiio.mk script accordingly.

Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

toolchain: improve conditions for gcc bug 99140

Gcc bug 99140 has been fixed on gcc 8.x but reappeared on gcc 9.x while
it's been fixed on gcc 10.x+. So let's update
BR2_TOOLCHAIN_HAS_GCC_BUG_99140 accordingly.

Fixes:
http://autobuild.buildroot.net/results/c55/c55f50a8d657695f0d5492c32efa666254cd7f99/

Signed-off-by: Giulio Benetti <giulio.benetti@benettiengineering.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

package/libcap: don't overwrite 'empty' when generating loader.txt

Upstream commit [1] introduced an invocation of objcopy to generat
loader.txt. However, objcopy, if not provided with an output file, will
overwrite the input file. This is usually harmless because it will be
identical, but the timestamp is updated. This may cause 'empty' to be
newer than 'loader.txt', which causes 'loader.txt' and its dependencies
to be rebuilt during 'make install'

We provide a different set of parameters during 'make install'. In
particular, we no longer pass in HOST_CONFIGURE_OPTS, so we no longer
set LDFLAGS. Thus, there is no -Wl,rpath option that is passed in, which
causes the resulting binaries to have an incorrect RPATH.

Fix this by adding /dev/null as the output file in the objcopy
invocation.

Patch was sent upstream, but there's no mailing list, just a single
person.

Fixes: http://autobuild.buildroot.net/results/600/600aff5b839b48db80751cace5fa9670b7a3d698
(hopefully)

[1] https://git.kernel.org/pub/scm/libs/libcap/libcap.git/commit/?id=efd293947f940180eedd8d0915b124f4aedccc08

Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

Revert "package/libcap: don't overwrite 'empty' when generating loader.txt"

This was for the next branch, not master...

This reverts commit 6fb0dbe4038ccde96e4600ca566d6269911077bc.

Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/libcap: don't overwrite 'empty' when generating loader.txt

Upstream commit [1] introduced an invocation of objcopy to generat
loader.txt. However, objcopy, if not provided with an output file, will
overwrite the input file. This is usually harmless because it will be
identical, but the timestamp is updated. This may cause 'empty' to be
newer than 'loader.txt', which causes 'loader.txt' and its dependencies
to be rebuilt during 'make install'

We provide a different set of parameters during 'make install'. In
particular, we no longer pass in HOST_CONFIGURE_OPTS, so we no longer
set LDFLAGS. Thus, there is no -Wl,rpath option that is passed in, which
causes the resulting binaries to have an incorrect RPATH.

Fix this by adding /dev/null as the output file in the objcopy
invocation.

Patch was sent upstream, but there's no mailing list, just a single
person.

Fixes: http://autobuild.buildroot.net/results/600/600aff5b839b48db80751cace5fa9670b7a3d698
(hopefully)

[1] https://git.kernel.org/pub/scm/libs/libcap/libcap.git/commit/?id=efd293947f940180eedd8d0915b124f4aedccc08

Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

support/testing: remove TestPythonPy2{Cryptography, ServiceIdentity, Treq, Twisted, Txtorcon}

The python2 support has been removed since the python-idna bump to version 3.2 [1]

[1] 0c7e30b43a5e98abfc8db521a6415b5a5c1c267f

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

package/libmodsecurity: disable -fPIC on m68k_cf

This package has -fPIC gcc option set by default but we can't use it on
m68k_cf since it doesn't support it throwing a gcc build failure. So let's
disable it by passing -fno-PIC.

Fixes:
http://autobuild.buildroot.net/results/b92980a563fe7ee331e70f288ce041be0bf29d40/

Signed-off-by: Giulio Benetti <giulio.benetti@benettiengineering.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/mesa3d: fix build on riscv32

Fix the following build failure on riscv32:

../src/util/futex.h: In function 'sys_futex':
../src/util/futex.h:39:19: error: 'SYS_futex' undeclared (first use in this function); did you mean 'sys_futex'?
   39 |    return syscall(SYS_futex, addr1, op, val1, timeout, addr2, val3);
      |                   ^~~~~~~~~
      |                   sys_futex

Fixes:
 - http://autobuild.buildroot.org/results/692700a5f967760a0b8cd358b1712f1d5a7b681e

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/snort3: link with -latomic if needed

Fix the following build failure raised since bump to version 3.1.6
in commit e66f2fd310374779d415fa683813cc5f5ccf6be9 and
https://github.com/snort3/snort3/commit/3e518d86040e74c328bcc0d5bbd49c1da6b7a3ec:

/home/buildroot/autobuild/instance-1/output-1/host/opt/ext-toolchain/bin/../lib/gcc/powerpc-buildroot-linux-uclibc/9.3.0/../../../../powerpc-buildroot-linux-uclibc/bin/ld: service_inspectors/dce_rpc/CMakeFiles/dce_rpc.dir/dce_smb2_file.cc.o: undefined reference to symbol '__atomic_load_8@@LIBATOMIC_1.0'

Fixes:
 - http://autobuild.buildroot.org/results/df34a69175e61bc7b180d89c738747e19aaf13bf

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/sdl2: fix build with kmsdrm

Build with kmsdrm is broken since bump to version 2.0.14 in commit
5e0da5c40da82d90c3f6ca037170838a6689b65b. Indeed, first patch was
already applied in this version:
https://github.com/libsdl-org/SDL/commit/9354aea19834ada7ffb90d379600a242a7aa820f
but upstream made other changes that requires EGL so add an upstream
patch to fix the build failure

Moreover, run autogen.sh instead of autoreconf as it breaks the build
and is not recommended by upstream:
https://github.com/libsdl-org/SDL/pull/4214

Fixes:
 - http://autobuild.buildroot.org/results/355c7e5092e7641d8b04ecb550e2671d70720bd2

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
[Arnout: add dependency on host-autoconf]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/sdl2: kmsdrm needs GBM

kmsdrm needs GBM (and so mesa3d) since its addition in version 2.0.6:
https://github.com/libsdl-org/SDL/commit/56363ebf6124b345e1cfbd14fb6c0e654837910c

If libgbm is not found, kmsdrm will be silently disabled

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/lua-lunix: fix sys/sysctl.h detection

sysctl.h has been removed from glibc since version 2.32.

Fixes: http://autobuild.buildroot.net/results/749a11d5289c6fec3b2f236b9073fc1ab730d090/
Signed-off-by: Francois Perrad <francois.perrad@gadz.org>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

Config.in: disable Fortify Source for microblaze

As reported by Toolchain-builder project [1], the microblaze glibc
toolchain creates a system that doesn't boot when FORTIFY_SOURCE is
enabled: the init process hangs.

Also, hardening features may not be wanted or possible for such
slow soft-core cpus [2].

Note: for completeness, BR2_RELRO_PARTIAL was manually tested and it
does boot.

[1] https://gitlab.com/bootlin/toolchains-builder/-/jobs/1467624500
[2] http://lists.busybox.net/pipermail/buildroot/2021-June/312416.html

Signed-off-by: Romain Naour <romain.naour@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Giulio Benetti <giulio.benetti@benettiengineering.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/freescale-imx/imx-gpu-viv: use make loops instead of shell loops

We generally prefer using make loops instead of shell loops. They
bring automatic error handling, as they abort the loop when there is
an error, without the need for "|| exit 1".

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Reviewed-by: Gary Bisson <gary.bisson@boundarydevices.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>