rust: Enable building rust from stable, beta and nightly channels
Rust follows the train release model via the stable, beta and nightly channels,
by default we build rust from the stable channel, however there are certain
features which are only available in the beta or nightly channels.
Make these channels available by setting a RUST_CHANNEL variable which defaults
to stable making this change transparent to the user.
The snapshot version used by rust during its compilation wont necessarily match
the version being built, specially if were building from an unstable channel,
to avoid confusion rename this to SNAPSHOT_VERSION and use RUST_VERSION for the
version to be built, which is automatically defined to PV.
Append -beta or -nightly to rusts PV for signature awareness.
It is important to note that this does not build rust from the beta/nightly
published tarball (which today build rust v1.67.0 and v1.68.0 respectively),
instead this builds rust from the current selected version (1.66.0) and enables
the beta/nightly features for that version.
Setting the variable RUST_CHANNEL=nightly results in the following:
Mikko Rapeli [Wed, 14 Dec 2022 08:54:25 +0000 (10:54 +0200)]
runqemu: add QB_SETUP_CMD and QB_CLEANUP_CMD
These enable running custom shell setup and cleanup commands
before and after qemu. Enables machine configurations to for
example run qemu with swtpm to emulate TPM devices.
Example config with meta-tpm2 based swtpm in machine config:
* image recipe depens on swtpm-native to get the native tools to PATH,
same handling as qemu itself
* in this case, swtpm daemon stops automatically with qemu machine, but
QB_CLEANUP_CMD could be used to kill a specific process and wipe
temporary files
Now runqemu and testimage.bbclass can be used with swtpm.
Signed-off-by: Mikko Rapeli <mikko.rapeli@linaro.org> Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Khem Raj [Thu, 15 Dec 2022 08:45:42 +0000 (00:45 -0800)]
acl: Enable largefile support by default
This ensures right flags are passed so enable it. Musl and decoupled
these APIs from _GNU_SOURCE and now are only available when
_LARGEFILE64_SOURCE macro is enabled
ptest-packagelists.inc: correctly assign fast and slow tests
The execution times were checked from
https://autobuilder.yocto.io/pub/non-release/20221127-2/testresults/qemux86-64-ptest/core-image-ptest-all/log.do_testimage.40537.20221127010849
and ptests were moved both ways according to the same criteria:
more than 30 seconds or less.
openssl is dropped from fast tests, as it is also listed in slow
tests (which is correct).
With the rearrangement fast ptests altogether take about 420 seconds
on qemux86_64 with kvm.
m4 ptest was missing a depedency.
Signed-off-by: Alexander Kanavin <alex@linutronix.de> Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
devtool/upgrade: correctly handle recipes where S is a subdir of upstream tree
'devtool modify' writes additional settings to workspace .bbappend so that this
can be handled correctly, but 'devtool upgrade' does not. This adds the missing
settings.
In particular, local files should not anymore mysteriously disappear from
SRC_URIs on upgrades.
Signed-off-by: Alexander Kanavin <alex@linutronix.de> Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Khem Raj [Wed, 14 Dec 2022 03:20:29 +0000 (19:20 -0800)]
libpcre2: Upgrade to 10.42
10.41 regressed apps which were using pcre2posix.h without pcre2.h
mariadb is one such application in meta-oe which broke, this immediate
upgrade helps fixing these sort of apps. See [1] and [2]
Ross Burton [Tue, 13 Dec 2022 16:14:07 +0000 (16:14 +0000)]
meson: no need to rebuild on install
On install, Meson will rebuild targets which don't define their
dependencies, as it can't know if they need to be rebuilt or not.
This includes gtk-doc, which can be slow to run. As we control the
execution of meson we know that a rebuild isn't required, so we can pass
--no-rebuild and speed up the install task.
Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Marta Rybczynska [Wed, 14 Dec 2022 18:45:19 +0000 (19:45 +0100)]
efibootmgr: update compilation with musl
Since the commit 005b6aba89eaf1b79fdd7565dd028fdd9bbfcc7d
(efivar: add musl libc compatibility) efibootmgr compiles with
musl too. Update the variable to take that into account.
Signed-off-by: Marta Rybczynska <marta.rybczynska@linaro.org> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
License-Update: Upstream has added Unicode Terms of Use license
(Unicode-TOU). Signed-off-by: Alex Kiernan <alex.kiernan@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Alex Kiernan [Fri, 16 Dec 2022 12:06:31 +0000 (12:06 +0000)]
rust: Install target.json for target rustc
As part of the merging builds, installation of target.json was lost
causing `rustc` to fail on startup with:
error: Error loading target specification: Could not find specification for target "aarch64-poky-linux-gnu". Run `rustc --print target-list` for a list of built-in targets
We know the full target specification, so just install it directly
rather than using the glob approach from previously.
Fixes: b9b0cd99cdc7 ("rust: Enable nativesdk and target builds + replace rust-tools-cross-canadian") Signed-off-by: Alex Kiernan <alex.kiernan@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
LICENSE for curl is set to MIT-open-group which is wrong and proper
license is 'curl'
I check below link and the line "Curl and libcurl are licensed under
the license below, which is inspired by MIT/X, but not identical." says
that the license is identical to MIT but actual license is identical
from the file "meta/files/common-licenses/curl" Link: https://curl.se/docs/copyright.html
Also, I do not find the MIT-open-group license text in the entire
source-code
Marek Vasut [Sat, 17 Dec 2022 04:57:38 +0000 (05:57 +0100)]
package_rpm: Fix Linux 6.1.0 perf 1.0 version mistranslation
With Linux 6.1.0 and perf 1.0-r9, a build which includes perf-dev fails due
to perf-dev depending on perf 6.6.1.0-r9 . This is because translate_vers()
operates on perf-dev and mangles its version. The following scenario occurs:
ver=6.1.0-r9
pv=1.0
pkgv=6.1.0
reppv=6.1.0
With Linux 6.1.0, a corner case is hit where pv is a substring of ver, which
yields this corrupted version 6.6.1.0-r9 . Example in python3:
Richard Purdie [Fri, 16 Dec 2022 11:15:26 +0000 (11:15 +0000)]
bitbake.conf: Add module function vardepsexclude entries
Now that bitbake can look into python module code variable and function
dependencies, there are a few extra basic variables we need to exclude.
This is done per function since it keeps the main exclusion list cleaner
and is also slightly faster.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Patrick Williams [Mon, 12 Dec 2022 22:18:08 +0000 (16:18 -0600)]
kernel-fitimage: reduce dependency to the cpio
In order to build a fitimage, we don't need the image to be "complete"
but we need the cpio portion of it built since the cpio is what ends
up inside the FIT. By reducing the dependency to `do_image_cpio` we
are able to include an image's rootfs as the ramdisk for a FIT and
then bundle that FIT into a larger SPI flash layout.
Signed-off-by: Patrick Williams <patrick@stwcx.xyz> Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Ross Burton [Mon, 12 Dec 2022 17:35:47 +0000 (17:35 +0000)]
gtk+3: upgrade 3.24.34 -> 3.24.35
The gtktypefuncs.c removal is obsolete as that file isn't in the tarball
anymore, so remove it.
However the wayland-scanner files are still in the tarball despite
efforts to stop this, so generalise the removal to ensure it covers all
files which may be affected. Further investigation will be needed to
see if this is still an issue.
Rebase 0003-Add-disable-opengl-configure-option.patch.
Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Ross Burton [Mon, 12 Dec 2022 15:25:41 +0000 (15:25 +0000)]
meson: drop redundant is_debianlike() patch
This patch disables is_debianlike() so that it always returns False in the
name of deterministic builds. However, the caller default_libdir() then
looks at the host filesystem to decide if /usr/lib64 exists and will
return either lib or lib64, so it still isn't deterministic.
This is only used for the default libdir value and meson.bbclass
always passes an explicit --libdir, so this is moot.
Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Wang Mingyu [Mon, 12 Dec 2022 07:29:54 +0000 (15:29 +0800)]
libpng: upgrade 1.6.38 -> 1.6.39
Changelog:
==========
* Changed the error handler of oversized chunks (i.e. larger than
PNG_USER_CHUNK_MALLOC_MAX) from png_chunk_error to png_benign_error.
* Fixed a buffer overflow error in contrib/tools/pngfix.
* Fixed a memory leak (CVE-2019-6129) in contrib/tools/pngcp.
* Disabled the ARM Neon optimizations by default in the CMake file,
following the default behavior of the configure script.
* Allowed configure.ac to work with the trunk version of autoconf.
* Removed the support for "install" targets from the legacy makefiles;
removed the obsolete makefile.cegcc.
* Cleaned up the code and updated the internal documentation.
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com> Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Robert Andersson [Sun, 11 Dec 2022 17:02:47 +0000 (18:02 +0100)]
go-crosssdk: avoid host contamination by GOCACHE
By default GOCACHE is set to $HOME/.cache.
Same issue for all other go recipes had been fixed by commit 9a6d208b:
[ go: avoid host contamination by GOCACHE ]
but that commit missed go-crosssdk recipe.
Signed-off-by: Robert Andersson <robert.m.andersson@atlascopco.com> Signed-off-by: Ming Liu <liu.ming50@gmail.com> Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Richard Purdie [Fri, 9 Dec 2022 16:05:24 +0000 (16:05 +0000)]
lib/sstatesig: Drop OEBasic siggen
We're now used to using hashes as part of the task hashes and the sstate code
relies on this. The older OEBasic hash approach therefore wouldn't work and
can be removed.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Xiangyu Chen [Fri, 9 Dec 2022 07:05:38 +0000 (15:05 +0800)]
openssh: remove RRECOMMENDS to rng-tools for sshd package
It appears that rngd is not needed as of linux-5.6 and later[1]
and should not be installed by default since the purpose of rngd
is to provide additional trusted sources of entropy.
We did some testing on real hardware, the result seems to support that
we no longer need rngd by default on kernel v5.6 and later.
Testing result as below:
1. observing the crng init stage.
the "random: crng init done" always available before fs being mounted.
2. generating random number without rngd.
testing command: dd if=/dev/random of=/dev/null status=progress
on Marvell CN96xx RDB board, speed almost 20.4 MB/s without block
on NXP i.mx6q board, speed almost 31.9 MB/s without block
on qemu x86-64, speed almost 2.6MB/s without block
3. using rngtest command without rngd
testing command: rngtest -c 1000 </dev/random
on Marvell CN96xx RDB board:
rngtest: input channel speed: (min=4.340; avg=135.364; max=146.719)Mibits/s
rngtest: FIPS tests speed: (min=8.197; avg=69.020; max=72.800)Mibits/s
rngtest: Program run time: 418771 microseconds
on NXP i.mx6q board:
rngtest: input channel speed: (min=96.820; avg=326.769; max=340.598)Mibits/s
rngtest: FIPS tests speed: (min=15.090; avg=37.543; max=40.324)Mibits/s
rngtest: Program run time: 570229 microseconds
on qemu x86-64:
rngtest: input channel speed: (min=37.769; avg=101.136; max=136.239)Mibits/s
rngtest: FIPS tests speed: (min=10.288; avg=30.682; max=40.155)Mibits/s
rngtest: Program run time: 836800 microseconds
4. observing sshd service.
using "systemctl disable rng-tools" disable service and reboot system.
system boot up normal, sshd service also start in normal time without
block.
baremetal-image: Avoid overriding qemu variables from IMAGE_CLASSES
Since qemuboot is part of IMAGE_CLASSES via qemu.inc it is being
inherited before we set the QB_FOO variables.
Since our variables have conditional definitions and at that point
they've already been defined by qemuboot, we can no longer define
them in our class.
Move the IMAGE_CLASSES inherit to execute it after we set the
QB_FOO variables to fix booting via runqemu.
Richard Purdie [Thu, 8 Dec 2022 11:54:32 +0000 (11:54 +0000)]
yocto-check-layer: Allow OE-Core to be tested
For unknown reasons we've never seemingly run the check layer script
against OE-Core itself. This isn't entirely straightforward as the core
layer is a bit of a special case, we can't for example compare signatures
against ourselve and we can't remove core from bblayers.conf.
Core does have distro, machine and software components too, in the case
of distro, our fallback default settings. Whilst the qemu machines could
be split into a seperate layer directory, core wouldn't then parse at all
standalone due to the lack of any machine so it seems a bit pointless to
do that.
These changes tweak the script to handle core's special cases, specifically
to allow distro and machine directories and to account for the README placed
a directory level higher than other layers.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org> Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Alex Stewart [Wed, 7 Dec 2022 22:33:33 +0000 (16:33 -0600)]
opkg: upgrade to version 0.6.1
Opkg 0.6.1 Changes:
- Opkg will no longer complain when trying to clean up the temporary
directory, if the directory does not exist.
- Fixed a SEGFAULT when parsing package indexes with invalid `Size` or
`Installed-Size` fields. These indexes will now produce a
comprehensible error.
- Fixed an inconsistecy in .list generation where files would sometimes
be entered with/without a trailing slash. The trailng slash should now
always be removed.
- Fixed [a bug](https://bugzilla.yoctoproject.org/show_bug.cgi?id=10461)
in package removal, where empty common directories would be left on
disk, even after all owning packages were removed.
Signed-off-by: Alex Stewart <alex.stewart@ni.com> Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Richard Purdie [Thu, 1 Dec 2022 17:11:07 +0000 (17:11 +0000)]
devtool/friends: Use LAYERSERIES_CORENAMES when generating LAYERSERIES_COMPAT entries
It seems some layers want to subvert the intent of LAYERSERIES_COMPAT
so bitbake is going to have to become stricter about the values there.
To work with this, use LAYERSERIES_CORENAMES to generate the entries in
LAYERSERIES_COMPAT instead of the current magic LAYERSERIES_COMPAT_core
value which may not continue to work.
The downside to this is when migating between releases, people would
need to update devtool workspace layer.conf files. I guess you could
argue this is a feature!
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Wed, 7 Dec 2022 16:57:55 +0000 (16:57 +0000)]
libxml2: upgrade 2.9.14 -> 2.10.3
Change ptest away from using the upstream Makefiles to manually running
the tests: they're not actually integrated with automake anyway so this
didn't gain us anything apart from patches we can't send upstream. Drops
the following patches:
- 0001-Make-ptest-run-the-python-tests-if-python-is-enabled.patch
- 0001-Port-gentest.py-to-Python-3.patch
- fix-execution-of-ptests.patch
- remove-fuzz-from-ptests.patch
- runtest.patch
Add a simple patch to install the test binaries via the Makefile:
- install-tests.patch
The Python module is built differently and a number of patches no longer
apply and appear to be redundant, remove:
- python-sitepackages-dir.patch
libxml-m4-use-pkgconfig.patch has been sent upstream now, mark as
backport.
Remove obsolete --without-docbook option.
Remove obsolete xml2Conf.sh packaging.
Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Peter Marko [Thu, 1 Dec 2022 14:04:40 +0000 (15:04 +0100)]
externalsrc: fix lookup for .gitmodules
Commit 0533edac277080e1bd130c14df0cbac61ba01a0c broke
bitbake parsing when bitbake is executed from directory with existing .gitmodules
and the recipe in externalsrc does not have .gitmodules
The check needs to search for .gitmodules in sources path, not cwd.
iParsing recipes...ERROR: ExpansionError during parsing <path to recipe>
...
bb.data_smart.ExpansionError: Failure expanding variable do_compile[file-checksums], expression was ${@srctree_hash_files(d)} which triggered exception CalledProcessError: Command '['git', 'config', '--file', '.gitmodules', '--get-regexp', 'path']' returned non-zero exit status 1.
Signed-off-by: Peter Marko <peter.marko@siemens.com> Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Ross Burton [Mon, 5 Dec 2022 14:59:32 +0000 (14:59 +0000)]
lib/buildstats: fix parsing of trees with reduced_proc_pressure directories
The /proc/pressure support in buildstats is creating directories in the
buildstats tree called reduced_proc_pressure, which confuses the parsing
logic as that cannot be parsed as a name-epoc-version-revision tuple.
Explicitly skip this directory to solve the problem.
Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Ovidiu Panait [Mon, 5 Dec 2022 14:56:49 +0000 (16:56 +0200)]
kernel.bbclass: remove empty module directories to prevent QA issues
Currently, allyesconfig test runs for x86_64 fail with:
ERROR: linux-yocto-5.19.17+gitAUTOINC+0cba9aa404_aaf4490d18-r0 do_package:
QA Issue: linux-yocto: Files/directories were installed but not shipped in any
package:
/lib/modules/5.19.17/kernel/drivers/nvdimm
With CONFIG_NVDIMM_TEST_BUILD=m, an empty nvdimm directory is created during
modules_install, which triggers the QA issue.
Extend kernel_do_install() to also remove inner empty directories that might
get created by modules_install.
projects / thirdparty / openembedded / openembedded-core.git / log
