package/zmqpp: use BR2_ENABLE_RUNTIME_DEBUG iso BR2_ENABLE_DEBUG

BR2_ENABLE_DEBUG should just steer the availability of debug symbols and
should have no negative effect on performance.

Introduction of 'assert' statements, 'debug'-type builds with additional
logging, etc. should be steered by BR2_ENABLE_RUNTIME_DEBUG instead.

Signed-off-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/zmqpp: don't set CONFIG=debug

Below are the flags set by zmqpp depending on the specified CONFIG variable:

CONFIG_FLAGS =
ifeq ($(CONFIG),debug)
        CONFIG_FLAGS = -g -fno-inline -ftemplate-depth-1000
endif
ifeq ($(CONFIG),valgrind)
        CONFIG_FLAGS = -g -O1 -DNO_DEBUG_LOG -DNO_TRACE_LOG
endif
ifeq ($(CONFIG),max)
        CONFIG_FLAGS = -O3 -funroll-loops -ffast-math -finline-functions \
            -fomit-frame-pointer -DNDEBUG
endif
ifneq (,$(findstring $(CONFIG),release loadtest))
        CONFIG_FLAGS = -O3 -funroll-loops -ffast-math -finline-functions \
            -fomit-frame-pointer -DNO_DEBUG_LOG -DNO_TRACE_LOG -DNDEBUG
endif

For the flags added with CONFIG=debug, '-g' is to be steered by the core
infrastructure (could be '-g1', '-g2' etc.)
The flag '-ftemplate-depth' is only a protection against incorrect code and
not really needed in Buildroot context.
Finally, the flag '-fno-inline' may be useful when really stepping through
zmqpp code, but is a very specific use case.

With the above in mind, not passing CONFIG=debug may actually be better.
Use 'CONFIG=buildroot' instead.

Note that we don't pass an empty 'CONFIG' to avoid confusion, as this
variable is also passed through to the variable BUILD_ENV, even though it is
currently unused.

Signed-off-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/sofia-sip: don't set 'NDEBUG' explicitly

BR2_ENABLE_DEBUG should just steer the availability of debug symbols and
should have no negative effect on performance.

Introduction of 'assert' statements, 'debug'-type builds with additional
logging, etc. should be steered by BR2_ENABLE_RUNTIME_DEBUG instead.

The sofia-sip package was setting '--enable-ndebug' conditionally based on
BR2_ENABLE_DEBUG, and this would have to be updated to be based on
BR2_ENABLE_RUNTIME_DEBUG.

However, the sofia-sip option '--enable-ndebug' only sets the 'NDEBUG'
preprocessor macro, and the core package infrastructure already sets this
macro correctly based on BR2_ENABLE_RUNTIME_DEBUG.

This means that the explicit '--enable-ndebug' flag can be removed.

Suggested-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/sofia-sip: correct passing of '--enable-ndebug'

The preprocessor option NDEBUG, triggered by the configure option
'--enable-ndebug', should be read as 'no-debug'. When NDEBUG is set, asserts
are _disabled_.

The sofia-sip package had inverted logic, and set '--enable-ndebug' when
BR2_ENABLE_DEBUG was enabled, while it should be the other way around.

Reported-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/ripgrep: use BR2_ENABLE_RUNTIME_DEBUG iso BR2_ENABLE_DEBUG

BR2_ENABLE_DEBUG should just steer the availability of debug symbols and
should have no negative effect on performance.

Introduction of 'assert' statements, 'debug'-type builds with additional
logging, etc. should be steered by BR2_ENABLE_RUNTIME_DEBUG instead.

Signed-off-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/qt5: use BR2_ENABLE_RUNTIME_DEBUG iso BR2_ENABLE_DEBUG

BR2_ENABLE_DEBUG should just steer the availability of debug symbols and
should have no negative effect on performance.

Introduction of 'assert' statements, 'debug'-type builds with additional
logging, etc. should be steered by BR2_ENABLE_RUNTIME_DEBUG instead.

Signed-off-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/oracle-mysql: use BR2_ENABLE_RUNTIME_DEBUG iso BR2_ENABLE_DEBUG

BR2_ENABLE_DEBUG should just steer the availability of debug symbols and
should have no negative effect on performance.

Introduction of 'assert' statements, 'debug'-type builds with additional
logging, etc. should be steered by BR2_ENABLE_RUNTIME_DEBUG instead.

Signed-off-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/boost: use BR2_ENABLE_RUNTIME_DEBUG iso BR2_ENABLE_DEBUG

BR2_ENABLE_DEBUG should just steer the availability of debug symbols and
should have no negative effect on performance.

Introduction of 'assert' statements, 'debug'-type builds with additional
logging, etc. should be steered by BR2_ENABLE_RUNTIME_DEBUG instead.

Signed-off-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/sysrepo: remove explicit setting of CMAKE_BUILD_TYPE

sysrepo explicitly sets CMAKE_BUILD_TYPE=Release, ignoring any possible
value of BR2_ENABLE_DEBUG (previously) or BR2_ENABLE_RUNTIME_DEBUG (now).

With the introduction of BR2_ENABLE_RUNTIME_DEBUG, this change should no
longer be necessary. Users that do not wish to have additional runtime
debugging just keep BR2_ENABLE_RUNTIME_DEBUG disabled (default value).

As the 'Debug' build type enables tests, disable them explicitly.
As the 'Debug' build type uses a custom REPO_PATH which does not exist on
target, force /etc/sysrepo like in the 'Release' build type.

Signed-off-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Acked-by: Heiko Thiery <heiko.thiery@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/sysrepo: use default CMAKE_BUILD_TYPE for host package

Use the default build type for host-sysrepo.
In the current version of sysrepo, this happens to be 'Debug'.
As 'Debug' also enables tests, explicitly disable them.

Suggested-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Acked-by: Heiko Thiery <heiko.thiery@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/flare-engine: disable effect of CMAKE_BUILD_TYPE

flare-engine enables profiling when CMAKE_BUILD_TYPE is 'Debug'. The
Buildroot package explicitly avoided that by forcing CMAKE_BUILD_TYPE to
'RelWithDebInfo' when pkg-cmake.mk would normally set it to 'Debug'. Until
recently, this was the case when BR2_ENABLE_DEBUG was enabled.

A previous commit changed the condition under which CMAKE_BUILD_TYPE=Debug
was set, from BR2_ENABLE_DEBUG=y to BR2_ENABLE_RUNTIME_DEBUG=y, so logically
the flare-engine package would have to be updated accordingly.

However, apart from the profiling flag, the flare-engine package only uses
CMAKE_BUILD_TYPE to determine flags that Buildroot wants to control itself,
like optimization and debugging flags.

This means we can fake CMAKE_BUILD_TYPE to a value that has no meaning for
flare-engine itself, without needing to check BR2_ENABLE_DEBUG nor
BR2_ENABLE_RUNTIME_DEBUG.

Incidentally, this trick was already done in case
BR2_TOOLCHAIN_HAS_GCC_BUG_85180 was true, so move that line out of this
condition.

Signed-off-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/pkg-cmake.mk: determine CMAKE_BUILD_TYPE depending on BR2_ENABLE_RUNTIME_DEBUG

The CMAKE_BUILD_TYPE is currently set as 'Debug' in case BR2_ENABLE_DEBUG is
set, and as 'Release' in other cases. However, while the description of
BR2_ENABLE_DEBUG is to enable debug symbols (no runtime impact), the 'Debug'
build type in CMake can actually have runtime impact. For one, because it
does not set -DNDEBUG like is done for 'Release', but also because packages
may do custom things based on it.

The question of which CMAKE_BUILD_TYPE Buildroot should set, be it 'Debug',
'Release', 'RelWithDebInfo' or others, has come up several times in the
past. See some references below:

- July 2016: switch from Debug to RelWithDebInfo:
  https://git.buildroot.org/buildroot/commit/?id=4b0120183404913f7f7788ef4f0f6b51498ef363

- October 2016: switch from RelWithDebInfo back to Debug:
  https://git.buildroot.org/buildroot/commit/?id=104bb29e0490bfb487e2e665448dd3ca07fcc2b5
  and changes to make sure Buildroot's flags are respected:
  https://git.buildroot.org/buildroot/commit/?id=12494ef48f893684d0800e7f6fe39a2ceaed0451

- August 2017: bug #10246 - "BR2_ENABLE_DEBUG does not have the expected
  effect for cmake packages"
  https://bugs.busybox.net/show_bug.cgi?id=10246

- August 2017: mail thread following bug #10246:
  http://lists.busybox.net/pipermail/buildroot/2017-August/200778.html

In the last mail thread, Samuel Martin confirmed that the 'Release' build
type could be used in all cases, because Buildroot is actually making sure
that the optimization flags are those determined by Buildroot, not the
defaults of cmake, thanks to commit 12494ef48f.
But Arnout Vandecappelle objected to using always 'Release', stating that
users may actually want the extra assertions.

With the introduction of BR2_ENABLE_RUNTIME_DEBUG, Buildroot can now cater
for all cases:

- use CMAKE_BUILD_TYPE=Release by default. This makes sure that there is no
  unexpected performance degradation triggered by enabling BR2_ENABLE_DEBUG.

- users can optionally enable BR2_ENABLE_RUNTIME_DEBUG if they want runtime
  debug info like assertions, at the risk of introducing performance
  degradation. In this case, we switch to CMAKE_BUILD_TYPE=Debug.

- orthogonally to the above, BR2_ENABLE_DEBUG still determines passing the
  '-g' flag to enable debug symbols, and BR2_OPTIMIZE_X still determines the
  used optimization flags.

Signed-off-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Reviewed-by: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

core: enable 'NDEBUG' unless BR2_ENABLE_RUNTIME_DEBUG is set

The 'assert' statement in glibc honors the 'NDEBUG' preprocessor macro: if
it is set, then the assert statement is compiled away.

Define this 'NDEBUG' macro when BR2_ENABLE_RUNTIME_DEBUG is disabled (the
default case).

Signed-off-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Reviewed-by: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

core: introduce BR2_ENABLE_RUNTIME_DEBUG

Some packages have optional runtime assertions, extra traces, or other
elements that can help in debugging problems. However, such runtime elements
can negatively influence performance.

In a test program performing 100K gRPC calls from a client to a local server
and receiving the returned response, we see following execution time:

    - runtime debug enabled: 1065 seconds
    - runtime debug disabled:  48 seconds

This is more than a factor 20 (!) difference. Analysis shows that the
problem mostly stems from libabseil-cpp (a dependency of gRPC) which enables
mutex deadlock analysis when the preprocessor flag 'NDEBUG' is not set,
which adds a 'backtrace()' call on every lock/unlock.  Potentially worse,
when libunwind is enabled and linked with the test program, 'backtrace()' is
not provided by glibc but by libunwind itself.

For production systems, users expect good performance out-of-the-box. In the
example above, the difference is huge and unless explicitly tested and
analyzed, users may not realize that the performance could be much better.

Address this problem by introducing a new option BR2_ENABLE_RUNTIME_DEBUG,
which can be used by packages or package infrastructures to set the
necessary flags.

Note that BR2_ENABLE_RUNTIME_DEBUG is orthogonal to BR2_ENABLE_DEBUG: the
former changes runtime behavior, while the latter is only expected to add
debug symbols to the build. Today, the cmake build system does introduce a
runtime impact when BR2_ENABLE_DEBUG is set, but that will be rectified in a
subsequent commit.

Signed-off-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Reviewed-by: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/{hostapd,wpa_supplicant}: alphabetical order to select openssl features

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

package/wget: enable MD4 and DES support in openssl

NTLM support in wget requires MD4 and DES. Make sure that openssl
supports these algorithms.

Fixes:
http://autobuild.buildroot.net/results/a342a4547dce1551e141c06c454f6e5f0830b671/
http://autobuild.buildroot.net/results/12b1f6f701b46a1dd2de5e32c86beecf590b6530/

Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/wlroots: check for BR2_PACKAGE_SEATD

Fix the wlroots packaging by checking for the correct variable to
determine whether seatd is being built.

Signed-off-by: Adrian Perez de Castro <aperez@igalia.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/hostapd: enable DES in openssl

Hostapd/wpa_supplicant crypto backend assumes that openssl always
provides DES support. However DES support in openssl has become
optional since  commit a83d41867c8d ("package/libopenssl: add
option to enable some features").

Select openssl DES support in hostapd Kconfig to avoid
build failures.

Signed-off-by: Sergey Matyukevich <geomatsi@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/wpa_supplicant: enable DES in openssl

Hostapd/wpa_supplicant crypto backend assumes that openssl always
provides DES support. However DES support in openssl has become
optional since  commit a83d41867c8d ("package/libopenssl: add
option to enable some features").

Select openssl DES support in wpa_supplicant Kconfig to avoid
build failures.

Fixes:
http://autobuild.buildroot.net/results/00d46eb765f23a1f6e168338f2cc863ae348045d/

Signed-off-by: Sergey Matyukevich <geomatsi@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/python-pymupdf: new package.

Python bindings for the PDF rendering library MuPDF.

Signed-off-by: Raphaël Mélotte <raphael.melotte@essensium.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
[Arnout:
 - propagate harfbuzz dependencies;
 - add comment to mupdf to keep versions in sync;
 - clarify license situation + lack of license file]

package/mupdf: new package.

MuPDF is a lightweight PDF, XPS, and E-book viewer.

Note: some packages, like PyMuPDF, depend on mupdf's libraries, hence
the patch.

Signed-off-by: Raphaël Mélotte <raphael.melotte@essensium.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
[Arnout:
 - propagate harfbuzz dependencies;
 - correct version number in hash file;
 - patch on post-patch hook instead of post-extract;
 - nicer line splitting]

package/Makefile.in: pass '-g0' explicitly if !BR2_ENABLE_DEBUG

If BR2_ENABLE_DEBUG is not set, Buildroot did not pass any flag
to control debug level. This means that the build system of the package
itself would control it.

Instead, provide an explicit '-g0' (no debugging symbols) to get consistent
behavior across packages.

Suggested-by: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/mtd: add BR2_PACKAGE_MTD_UBIHEALTHD option

mtd-utils-2.1.2 add ubihealthd daemon

Signed-off-by: Xiaobing Luo <luoxiaobing0926@gmail.com>
Reviewed-by: Matthew Weber <matthew.weber@collins.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/libgit2: disable NTLM support unconditionally.

NTLM is an old authentication protocol depending on DES and MD4 and
its creator, Microsoft, discourage its use since 2010, because it is
weak and vulnerable to many attacks.

libgit2 has support for NTLM and it is enabled by default if openssl
is used, however, since commit a83d41867c8d ("package/libopenssl: add
option to enable some features"), support for DES and MD4 can be
disabled in openssl.

Since openssl is an optional dependency to libgit2 and getting rid of
NTLM is a goal in itself for many windows administrators, disable NTLM
by default.

Fixes: http://autobuild.buildroot.net/results/efb4ce2174cb6944558e895ce852182ba2738a70
Signed-off-by: Nicolas Cavallari <nicolas.cavallari@green-communications.fr>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/libepoxy: bump to version 1.5.7

This update includes a build fix for the RaspberryPi 4 (from 1.5.7)
and a fix for the GL library loader when GLX is involved (from 1.5.6)
Release notes:

  https://github.com/anholt/libepoxy/releases/tag/1.5.6
  https://github.com/anholt/libepoxy/releases/tag/1.5.7

Signed-off-by: Adrian Perez de Castro <aperez@igalia.com>
[francois.perrad@gadz.org: fix indent in hash file]
Signed-off-by: Francois Perrad <francois.perrad@gadz.org>
[yann.morin.1998@free.fr: meld Adrian and François' patches together]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/libsoup: bump to version 2.72.0

Among other fixes, switching to the current stable version plugs a few
memory leaks; solves many WebSocket bugs; makes NTLM authentication work;
allows building gobject-introspection data when cross-building (to be
enabled in a separate patch); fixes message cancellation; adds support
for HTTP 308 permanent redirects, same-site cookies, secure cookies,
HSTS, and Brotli compression.

This also removes the unneeded LIBSOUP_CONF_ENV which defined an
autoconf variable, and updates the Meson build configuration options
to follow the changes done by upstream.

Signed-off-by: Adrian Perez de Castro <aperez@igalia.com>
[yann.morin.1998@free.fr:
  - reorder options and move them one per line
  - while at it, do the same for dependencies
]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/{protobuf, python-protobuf}: bump to version 3.17

Signed-off-by: Michael Nosthoff <buildroot@heine.tech>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/libsoup: bump to version 2.66.5

Change libsoup to version 2.66.5, which introduces many fixes. This
needs adding the libpsl dependency, and switching over to Meson as the
autotools based build system is no more. The existing patch is not
needed anymore.

Signed-off-by: Adrian Perez de Castro <aperez@igalia.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/libpsl: new package

Package libpsl is a new non-optional dependency of future libsoup
versions.

Signed-off-by: Adrian Perez de Castro <aperez@igalia.com>
[yann.morin.1998@free.fr: reorder some variables]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/sqlite: bump to version 3350500

Release notes: https://www.sqlite.org/releaselog/3_35_5.html

Signed-off-by: Francois Perrad <francois.perrad@gadz.org>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/kodi-inputstream-adaptive: bump version to 2.6.16-Matrix

Changelog:
https://github.com/xbmc/inputstream.adaptive/blob/Matrix/inputstream.adaptive/addon.xml.in#L22

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/xmrig: new package

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
[me:
  - introduce XMRIG_SUFFIX
  - extend comment about installation commands
]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/libsndfile: add external libs support

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
[yann.morin.1998@free.fr: dependencies fit on a single line]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/libsndfile: bump to version 1.0.31

- Switch site to github to get latest release
- Drop all patches (already in version)
- Update indentation in hash file (two spaces)

https://github.com/libsndfile/libsndfile/releases/tag/1.0.31

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

package/pifmrds: use pkg-config

Use pkg-config to retrieve libsndfile dependencies

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
(cherry picked from commit 888546e5273d77d49bec564a515e85d7acee6bdd)
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/makedumpfile: bump to version 1.6.9

Signed-off-by: Alexander Egorenkov <egorenar-dev@posteo.net>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/fluidsynth: bump to version 2.2.1

For change log since v2.2.0, see:
- https://github.com/FluidSynth/fluidsynth/releases/tag/v2.2.1

./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: Yann E. MORIN <yann.morin.1998@free.fr>

package/wpa_supplicant: enable MD4 in openssl

Hostapd/wpa_supplicant crypto backend assumes that openssl always
provides MD4 support. However MD4 support in openssl has become
optional since  commit a83d41867c8d ("package/libopenssl: add
option to enable some features").

Select openssl MD4 support in wpa_supplicant Kconfig to avoid
build failures.

Signed-off-by: Sergey Matyukevich <geomatsi@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/hostapd: enable MD4 in openssl

Hostapd assumes that openssl always provides MD4 support. However MD4
support in openssl has become optional since  commit a83d41867c8d
("package/libopenssl: add option to enable some features").

Select openssl MD4 support in hostapd Kconfig to avoid build failures.

Fixes:
http://autobuild.buildroot.net/results/09ce1d8009da98cd2a3f2c8488abaa9f060a9d6b/

Signed-off-by: Sergey Matyukevich <geomatsi@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/python-bleak: bump to version 0.11.0

Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/openocd: add optional dependency on libgpiod

To make the Linux GPIOD bitbang adapter driver through libgpiod
available.

Signed-off-by: Sébastien Szymanski <sebastien.szymanski@armadeus.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/openocd: bump to 0.11.0

Remove upstream patches.

Signed-off-by: Sébastien Szymanski <sebastien.szymanski@armadeus.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/webkitgtk: make gst-gl depend on _HAS_PLATFORM

The GStreamer-GL usage in WebKitGTK is usable as long as GStreamer
is configured with a valid platform API (GLX+OpenGL, EGL+OpenGL,
EGL+GLES, etc.), which is exactly what the symbol
BR2_PACKAGE_GST1_PLUGINS_BASE_LIB_OPENGL_HAS_PLATFORM represents.

Signed-off-by: Adrian Perez de Castro <aperez@igalia.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/wpewebkit: make gst-gl depend on _HAS_PLATFORM

The GStreamer-GL usage in WPE WebKit is usable as long as GStreamer
is configured with a valid platform API (GLX+OpenGL, EGL+OpenGL,
EGL+GLES, etc.), which is exactly what the symbol
BR2_PACKAGE_GST1_PLUGINS_BASE_LIB_OPENGL_HAS_PLATFORM represents.

Signed-off-by: Adrian Perez de Castro <aperez@igalia.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/webkitgtk: select missing multimedia deps

Select a few missing multimedia related dependencies:

- BR2_PACKAGE_GST1_PLUGINS_GOOD_PLUGIN_AUTODETECT is needed for
  "autoaudiosink"; not having this element can cause a crash as
  it is used unconditionally.
- BR2_PACKAGE_GST1_PLUGINS_GOOD_PLUGIN_MATROSKA and
  BR2_PACKAGE_GST1_PLUGINS_GOOD_PLUGIN_VPX are needed for
  WebM video playback.

Note that BR2_PACKAGE_GST1_PLUGINS_GOOD_PLUGIN_VPX depends on THREADS
(because of libvpx), but webktigtk already depend on libgtk3, which
depends on THREADS. The probability that GTK3 drops the dependency on
THREADS is sufficiently close to zero that we need not account for that.
So we do not propagate the THREADS dependency for this option.

Signed-off-by: Adrian Perez de Castro <aperez@igalia.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/wpewebkit: select gstreamer autoaudiosink

Select BR2_PACKAGE_GST1_PLUGINS_GOOD_PLUGIN_AUTODETECT when multimedia
support is enabled. This is needed at runtime to automatically select
a suitable audio output element, otherwise WebKit will crash at an
assertion due to the missing "autoaudiosink" element. More here:

  https://wpewebkit.org/about/faq.html#why-does-the-browser%2Flauncher-(e.g.-cog)-crash-when-trying-to-play-audio%3F

Signed-off-by: Adrian Perez de Castro <aperez@igalia.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/uclibc: enable linuxthreads for all supported architectures

Synchronize the list of architectures that have LinuxThreads support
with the ones from uClibc-ng.

Signed-off-by: Mircea GLIGA <mgliga@bitdefender.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/audit: bump to version 3.0.1

Drop all upstream patches as they no longer are applicable.

Signed-off-by: Adam Duskett <aduskett@gmail.com>
Tested-by: Matthew Weber <matthew.weber@collins.com>
Tested-by: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/setools: bump to version 4.4.0

- Remove upstream patch 0001-remove-werror-flag-from-setup.patch
- Refactor Do-not-export-use-setools.InfoFlowAnalysis-and-setoo.patch to apply
  to 4.4.0

Signed-off-by: Adam Duskett <aduskett@gmail.com>
Tested-by: Matthew Weber <matthew.weber@collins.com>
Tested-by: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/selinux-python: bump to version 3.2

Signed-off-by: Adam Duskett <aduskett@gmail.com>
Tested-by: Matthew Weber <matthew.weber@collins.com>
Tested-by: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/semodule-utils: bump to version 3.2

Signed-off-by: Adam Duskett <aduskett@gmail.com>
Tested-by: Matthew Weber <matthew.weber@collins.com>
Tested-by: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/restorecond: bump to version 3.2

Signed-off-by: Adam Duskett <aduskett@gmail.com>
Tested-by: Matthew Weber <matthew.weber@collins.com>
Tested-by: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/checkpolicy: bump to version 3.2

Signed-off-by: Adam Duskett <aduskett@gmail.com>
Tested-by: Matthew Weber <matthew.weber@collins.com>
Tested-by: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/policycoreutils: bump to version 3.2

Also refactor patches to apply cleanly.

Signed-off-by: Adam Duskett <aduskett@gmail.com>
Tested-by: Matthew Weber <matthew.weber@collins.com>
Tested-by: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/libselinux: bump to version 3.2

Also drop upstream patch
0003-libselinux-rename-gettid-to-something-which-never-conflicts.patch

Signed-off-by: Adam Duskett <aduskett@gmail.com>
Tested-by: Matthew Weber <matthew.weber@collins.com>
Tested-by: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/libsemanage: bump to version 3.2

Signed-off-by: Adam Duskett <aduskett@gmail.com>
Tested-by: Matthew Weber <matthew.weber@collins.com>
Tested-by: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/libsepol/Config.in: Add policy version 33

Policy version 33 was added starting with kernel 5.8.

Signed-off-by: Adam Duskett <aduskett@gmail.com>
Tested-by: Matthew Weber <matthew.weber@collins.com>
Tested-by: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/libsepol: bump to version 3.2

Signed-off-by: Adam Duskett <aduskett@gmail.com>
Tested-by: Matthew Weber <matthew.weber@collins.com>
Tested-by: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

boot/arm-trusted-firmware: bump to version 2.5

Release notes:
https://trustedfirmware-a.readthedocs.io/en/latest/change-log.html#version-2-5

The existing qemu_aarch64_sbsa_defconfig was using an unpinned version
of ATF, so to avoid any regression, it is pinned to the previous
version, 2.4.

Signed-off-by: Heiko Thiery <heiko.thiery@gmail.com>
[yann.morin.1998@free.fr: pin ATF version in qemu_aarch64_sbsa_defconfig]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/libbpf: new package

A mirror of bpf-next linux tree bpf-next/tools/lib/bpf
directory plus its supporting header files. The version
of the package reflects the version of ABI.

Signed-off-by: Jugurtha BELKALEM <jugurtha.belkalem@smile.fr>
Signed-off-by: Romain Naour <romain.naour@gmail.com>
Cc: Qais Yousef <qais.yousef@arm.com>
Reviewed-by: Qais Yousef <qais.yousef@arm.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
[Arnout: add hash file]

package/llvm: add BPF backend

bcc is a front-end tool for eBPF :
https://github.com/iovisor/bcc/blob/master/README.md.
eBPF is the most powerful Linux tracer, and bcc
allows to write eBPF scripts in C and PYTHON3.

BCC requires enabling LLVM backend BPF support.

Signed-off-by: Jugurtha BELKALEM <jugurtha.belkalem@smile.fr>
Signed-off-by: Romain Naour <romain.naour@gmail.com>
Tested-by: Qais Yousef <qais.yousef@arm.com>
Reviewed-by: Qais Yousef <qais.yousef@arm.com>
Cc: Qais Yousef <qais.yousef@arm.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/libglib2: bump version to 2.68.2

Includes an important bugfix: #2387 json-glib does not build with glib 2.68.1

Signed-off-by: Adam Duskett <aduskett@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/gobject-introspection: bump to version 1.68.0

Other changes:
  - Change -Dgi_cross_use_host_gi=true to -Dgi_cross_use_prebuilt_gi=true as the
    option has changed.
  - Explicitly define python3 in both the host and target builds of
    g-ir-tool-template.in

Signed-off-by: Adam Duskett <aduskett@gmail.com>
[yann.morin.1998@free.fr:
  - add hash for giscanner/scannerlexer.l used as license file
  - two spaces in hash file
]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/libglib2: add -Dxattr=true when libselinux is selected

Xattrs is required for SELinux. Explicitly enable the feature when the
libselinux package is selected.

Signed-off-by: Adam Duskett <aduskett@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/libglib2: bump to version 2.68.1

Other changes:
  - Drop 0002-allow-explicit-disabling-of-tests.patch as this patch has been
    upstreamed and the option is now simply "tests."
  - Add -Dglib_debug=disabled and -Dlibelf=disabled as defaults to both
    HOST_LIBGLIB2_CONF_OPTS and LIBGLIB2_CONF_OPTS
  - Refactor existing patches to apply to 2.68.1

Tested with test-pkg -p libglib2 -a.

Signed-off-by: Adam Duskett <aduskett@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

configs/acmesystems_acqua_a5_*: use tested bootloader version

The new branch "4.x" of at91bootstrap has dropped the contrib directory,
which holds the defconfigs used by the Acmesystems Acqua A5 boards. We
then cannot use the latest at91bootstrap anymore.

As commit e009816c67fc1335a9b2da8687b55c5ff6aa27cf introduced support
for at91bootstrap 4.x in the "next" branch, it had to work around this
by pinning these defconfigs to the "latest 3.x" version.

Avoid this and any future incompatibility problems by explicitly
requesting a tested version of at91bootstrap.

Signed-off-by: Edgar Bonet <bonet@grenoble.cnrs.fr>
[yann.morin.1998@free.fr: use a custom git, like other acme defconfigs]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

board/raspberrypi: add Compute Module 4/IO Board support (64 Bit)

Signed-off-by: Peter Seiderer <ps.report@gmx.net>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

board/raspberrypi: add Compute Module 4/IO Board support

Signed-off-by: Peter Seiderer <ps.report@gmx.net>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

board/raspberrypi: add rpi4 64 bit hint

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

board/raspberrypi: add rpi3 64 bit hint

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

package/raspberrypi-usbboot: enable for BR2_aarch64

Signed-off-by: Peter Seiderer <ps.report@gmx.net>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/raspberrypi-usbboot: bump version to 9324fd7

- add license description/file/hash

- rebased 0001-Makefile-allow-passing-CFLAGS-LDFLAGS.patch

- removed 0002-Makefile-add-DESTDIR-support.patch
  (Makefile install target removed since [1])

- removed 0003-main.c-rework-logic-to-find-def1-def2-and-def3-files.patch
  (bootcode.bin/bootcode4.bin and start.elf/start4.elf compiled in since [2])

- change host install command

[1] https://github.com/raspberrypi/usbboot/commit/9e6ff777bbf271b75aa0b1aa24e6f12d7b768375
[2] https://github.com/raspberrypi/usbboot/commit/1bb4c2da4775329bcf5811616fc8f0ce78ee2121

Signed-off-by: Peter Seiderer <ps.report@gmx.net>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/{mesa3d, mesa3d-headers}: bump version to 21.1.1

Release notes:
https://lists.freedesktop.org/archives/mesa-announce/2021-May/000631.html

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/libopenssl: md5 is not optional

Commit a83d41867c8d (package/libopenssl: add option to enable some
features) made md5 an optional feature.

However, md5 has not been optional since 1.1.0, released in 2016.

Drop that option.

Fixes:
   - http://autobuild.buildroot.net/results/9d79c9f313ff989449e9b47c1ff0afd3a167fd2d/

Reported-by: "Weber, Matthew L Collins" <Matthew.Weber@collins.com>
Reviewed-by: "Matthew Weber" <Matthew.Weber@collins.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/postgresql: select mandatory libopenssl features

Select mandatory libopenssl features (which are selectable since
commit a83d41867c8d69a77d5cd0a665aa216af5340359):

- BR2_PACKAGE_LIBOPENSSL_ENABLE_BLOWFISH
- BR2_PACKAGE_LIBOPENSSL_ENABLE_CAST
- BR2_PACKAGE_LIBOPENSSL_ENABLE_DES

Fixes:

  - http://autobuild.buildroot.net/results/7fdcaa337a7369673ac4580ff7a2bbccc895dca2

  openssl.c:449:35: warning: implicit declaration of function 'EVP_bf_ecb'; did you mean 'EVP_sm4_ecb'? [-Wimplicit-function-declaration]
    449 |  if (!EVP_EncryptInit_ex(evp_ctx, EVP_bf_ecb(), NULL, NULL, NULL))
        |                                   ^~~~~~~~~~
        |                                   EVP_sm4_ecb
  openssl.c:449:35: warning: passing argument 2 of 'EVP_EncryptInit_ex' makes pointer from integer without a cast [-Wint-conversion]
    449 |  if (!EVP_EncryptInit_ex(evp_ctx, EVP_bf_ecb(), NULL, NULL, NULL))
        |                                   ^~~~~~~~~~~~
        |                                   |
        |                                   int
  In file included from openssl.c:34:
  .../host/arm-buildroot-linux-uclibcgnueabihf/sysroot/usr/include/openssl/evp.h:583:53: note: expected 'const EVP_CIPHER *' {aka 'const struct evp_cipher_st *'} but argument is of type 'int'
    583 |                                   const EVP_CIPHER *cipher, ENGINE *impl,
        |                                   ~~~~~~~~~~~~~~~~~~^~~~~~

Signed-off-by: Peter Seiderer <ps.report@gmx.net>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/libcurl: make sure openssl supports DES

DES support in openssl is optional since commit a83d41867c8d
("package/libopenssl: add option to enable some features"). curl uses
DES support in openssl when enabled. Select DES support to avoid build
failure.

Fixes:
http://autobuild.buildroot.net/results/8931fd2b71eb2490cdc041b47eff58377ac823da/
http://autobuild.buildroot.net/results/f159195adf05c375d7d314a270afd3a55d43ebdd/
http://autobuild.buildroot.net/results/e143c75bccd7bcd4216954b76099404a38cf9bc1/

Cc: Erwan GAUTRON <erwan.gautron@bertin.fr>
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Reviewed-by: Matthew Weber <matthew.weber@collins.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

support/testing: add python-dbus-next test

Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/python-dbus-next: new package

Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

configs/microchip_sama5d2_icp: new defconfigs

Add the new SAMA5D2 ICP (Industrial Connectivity Board)
with linux4sam_2020.10 components.
Update README file with new defconfigs.

https://www.linux4sam.org/bin/view/Linux4SAM/Sama5d2IcpMainPage

Signed-off-by: Eugen Hristev <eugen.hristev@microchip.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

boot/at91bootstrap3: move BR2_TARGET_AT91BOOTSTRAP3_NEEDS_PYTHON3

As spotted by Eugen, BR2_TARGET_AT91BOOTSTRAP3_NEEDS_PYTHON3 currently
is outside the at91bootstrap section, because it was inccorectly added
after the 'endif' statement rather than before, which makes the
menuconfig layout weird.

Move it around.

Reported-by: Eugen Hristev <eugen.hristev@microchip.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

configs/uevm5432: new defconfig

The OMAP5432 uEVM[1] is a development board from Texas Instruments.

It is similar to the OMAP4 Panda boards, from which this configuration is
inspired.

[1]: https://svtronics.com/5432

Signed-off-by: Vincent Stehlé <vincent.stehle@laposte.net>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/libmdbx: bump version to 0.10.0

Release notes: https://github.com/erthink/libmdbx/releases/tag/v0.10.0

Signed-off-by: Leonid Yuriev <leo@yuriev.ru>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/dt-utils: bump version to 2021.03.0

Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/webkitgtk: bump to version 2.32.1

Update to a new major release which bring in improvements and a few new
features. Release notes:

  https://webkitgtk.org/2021/03/26/webkitgtk2.32.0-released.html
  https://webkitgtk.org/2021/05/10/webkitgtk2.32.1-released.html

None of the new features in WebKitGTK 2.32.x need additional
dependencies.

Signed-off-by: Adrian Perez de Castro <aperez@igalia.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

boot/at91bootstrap3: add BR2_TARGET_AT91BOOTSTRAP3_NEEDS_PYTHON3

Some at91bootstrap3 configurations now use Python scripts on the host
for NAND/PMEC related utilities. In order to be able to use those
scripts, this commit adds a new
BR2_TARGET_AT91BOOTSTRAP3_NEEDS_PYTHON3 which allows to express the
need for host-python3 as a dependency to build at91bootstrap3.

Signed-off-by: Eugen Hristev <eugen.hristev@microchip.com>
[Thomas: this was extracted from a patch from Eugen adding
at91bootstrap 4.x support.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

boot/at91bootstrap3: bump 3.x series to 3.10.2

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

boot/at91bootstrap3: add support for at91bootstrap 4.x series

The project at https://github.com/linux4sam/at91bootstrap was until
now releasing 3.x versions, which were packaged using
boot/at91bootstrap3/ in Buildroot. Microchip has now started a new
branch of at91bootstrap, called 4.x, which will only support the
following devices: sam9x60, sama5d2, sama5d3, sama5d4, sama7g5. A
number of older devices from Microchip will only be supported by the
existing 3.x series.

Therefore, we cannot simply remove support for the 3.x series, and
allow using only the 4.x series.

So what this commit does is extend the boot/at91bootstrap3 package to
support building both 3.x and 4.x versions. In detail, this implies:

 * Having the BR2_TARGET_AT91BOOTSTRAP3_LATEST_VERSION symbol point to
   the latest 4.x version. Indeed, we want
   BR2_TARGET_AT91BOOTSTRAP3_LATEST_VERSION to really point to the
   latest upstream version, even if that means potential breakage for
   users. Users who want to use a fixed version of at91bootstrap
   should anyway not be using
   BR2_TARGET_AT91BOOTSTRAP3_LATEST_VERSION.

 * Introduce BR2_TARGET_AT91BOOTSTRAP3_LATEST_VERSION_3X for users who
   would like to use the latest 3.x series.

 * Adjust the installation logic, as images to install are now in
   build/binaries/*.bin instead of binaries/*.bin. In order to not
   have to differentiate 3.x and 4.x, we simply use $(wildcard ...) to
   expand the list of files to install.

 * To make it clear that boot/at91bootstrap3 supports both 3.x and
   4.x, we also update the prompt of the package.

at911bootstrap does not carry a license file; so far we were using
main.c as the license file, as it carries the license blurb. Now that we
have a known alternate version, we would need a per-version hash for
that file. However, this is a bit too cumbersome to handle, so just drop
using main.c as the license file. When upstream introduces a proper
license file, we can revisit the situation.

Update the two defconfigs that were using the upstream 3.9.3 version;
all other defconfigs are using custom tarballs or custom git trees.

Signed-off-by: Eugen Hristev <eugen.hristev@microchip.com>
[Thomas: while this patch is based on previous work by Eugen, it was
reworked quite significantly.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
[yann.morin.1998@free.fr:
  - drop main.c as license file, explain why
  - update the two defconfigs
]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/grpc: bump to version 1.37.1

Fixes from Changelog [0]:
 - Fix #25897 to avoid crashes when certificates are not yet updated
 - [Backport] Fix use-after-unref bug in fault_injection_filter.

[0] https://github.com/grpc/grpc/releases/tag/v1.37.1

Signed-off-by: Michael Nosthoff <buildroot@heine.tech>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/kodi-peripheral-joystick: bump version to 1.7.2-Matrix

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

support/dependencies, scripts: accept patches with renames

Currently, patches with renames are refused, as they reqire patch 2.7
or newer. So far, we did not require that version because it was too
recent to be widely available.

But patch 2.7 has been released in 2012, almost 9 years ago now; it is
old enough that we can start relying on it.

Add a check that patch is GNU patch 2.7 or newer, and so drop the common
check for patch, and drop the check about renames in apply-patches.sh.

Signed-off-by: Ryota Kinukawa <pojiro.jp@gmail.com>
[yann.morin.1998@free.fr:
  - drop common check
  - shorten variable names
  - drop now-incorrect comment about busybox w/desktop
]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/rustc: add support for Tier 1 and Tier 2 platforms

This commit reworks how BR2_PACKAGE_HOST_RUSTC_TARGET_ARCH_SUPPORTS is
defined to more clearly map with the list of platforms supported by
Rust as listed at
https://doc.rust-lang.org/nightly/rustc/platform-support.html. Indeed,
the situation is not as simple as a list of architectures, all
supported for both glibc and musl.

So instead, we take the approach of directly mapping with what's
described at
https://doc.rust-lang.org/nightly/rustc/platform-support.html, which
means:

 * A list of Tier 1 platforms (in fact just 3 platforms)

 * A list of Tier 2 platforms with host tools (i.e where rustc and
   cargo themselves are available for the target, something that isn't
   relevant for Buildroot)

 * A list of Tier 2 platforms with no host tools support.

For each platform, we add as a comment its Rust tuple name, as listed
at https://doc.rust-lang.org/nightly/rustc/platform-support.html, and
then the corresponding Buildroot architecture/libc dependency.

This is obviously more verbose than it was, but it is also a lot
easier to maintain.

With this, a total of 16 new platforms are supported, 13 of which are
musl-based. The additional non-musl platforms are ARMv5TE, RISC-V
64-bit and Sparc64.

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

package/rustc: prepare for musl-based platforms support

The current rustc package only supports configurations based on glibc
and hardcodes this requirement. This patch prepares the addition of
support for musl-based platforms by using $(LIBC) instead of
hardcoding "gnu" as the C library specifier when defining
RUSTC_TARGET_NAME.

Signed-off-by: Nathaniel Husted <nathaniel.husted@outlook.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/abootimg: Update to 1ebeb39, switch to ajs124

ajs124 [1] has 2 fixes over than original [2].

[1] https://gitlab.com/ajs124/abootimg
[2] https://github.com/ggrandou/abootimg.

Signed-off-by: Petr Vorel <petr.vorel@gmail.com>
[yann.morin.1998@free.fr: also change homepage in Config.in]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/ltp-testsuite: enable ebizzy

ebizzy has been fixed in 20200930 (commit 967612c45).

Signed-off-by: Petr Vorel <petr.vorel@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/imlib2: bump version to 1.7.1

Signed-off-by: Petr Vorel <petr.vorel@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

package/numactl: add -latomic to numa.pc

Instead of passing -latomic through LDFLAGS as done in commit
4ed540ddf59bec4b389be44d7f42820d2466904f, properly check for it in
configure.ac and add it to numa.pc or numactl users won't be able to get
it

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

package/gobject-introspection/Config.in: Allow RISCV_32 builds

Qemu 6.0.0 no longer segfaults when running qemu-riscv32. As such, it's now
possible to allow riscv32 to compile gobject-introspection.

This partially reverts commit c94a212390e33c122cf9db4c772caf8bf3804565.
The symbol BR2_PACKAGE_GOBJECT_INTROSPECTION_ARCH_SUPPORTS is kept, even
though it is now equivalent to BR2_PACKAGE_HOST_QEMU_ARCH_SUPPORTS,
because it's likely that architecture dependencies will pop up again in
the future.

The configuration of a previously failing autobuilder [1] now passes.

[1] http://autobuild.buildroot.org/results/668397b1df42297505e6fc8353c4752290a5628d

Signed-off-by: Adam Duskett <aduskett@rivian.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/qemu: bump version to 6.0.0

Other changes:
  - Drop patch 0003-hw-usb-host-libusb.c-fix-build-with-kernel-5.0.patch as it
    is part of 6.0.0.
  - Rename 0004-meson-add-tests-option.patch to
    0003-meson-add-tests-option.patch and refresh it for 6.0.0

Signed-off-by: Adam Duskett <aduskett@rivian.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

package/x11r7/xlib_libXfixes: bump version to 6.0.0

Release notes:
https://lists.x.org/archives/xorg-announce/2021-May/003086.html

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

package/mesa3d: bump version to 21.1.0

Removed kmsro-related options, these are now handled by the build
system:
https://cgit.freedesktop.org/mesa/mesa/commit/?id=70813c1c13b99cb029c8fa3537163650bdd17b6d

Updated license hash due to typo fix in upstream commit:
https://cgit.freedesktop.org/mesa/mesa/commit/?id=2f42c21221cf89f738225e3ac2aa2e85953c1dc1

Rebased patch 0003 due to upstream commit:
https://cgit.freedesktop.org/mesa/mesa/commit/?id=a9618e7c4214e29c97009870ad8904623746ffe9

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