As with the kmod version of depmod, exclude .debug from being
searched. Since busybox does not use the depmod.d and any
configuration file option is ignored we just hardcode it.
(From OE-Core rev: c082752c06d5723433886cbf7ce2d88a51fb64f1)
Signed-off-by: Saul Wold <saul.wold@windriver.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
meson would fail to detect compiler type in some rare care where
specific substring '-xt' is detected in compiler --version output and
it so happens that this string can be generated by clang --version if
clang is installed into a directory containing 'xt-' in its name. with
recipe specific sysroots, this is quite likely to happen in OE build
system as we are seeing the issue with newly proposed gnome-text-editor
recipe
https://lists.openembedded.org/g/openembedded-devel/topic/90150031#96301
(From OE-Core rev: ff75909f2a9e970aaf389e0012888c29f02376e3)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Generalise this test so that it works on more than qemux86-64:
- Don't edit a file in arch/x86 to cause a rebuild, instead use init/
- Look for the edits in the build tree, as the deployed kernel could be
of any type (zImage/bzImage/etc) and edits may be in the compressed
part.
Also remove redundant checks on the result of runCmd(), as this will
raise AssertionError exceptions itself so the explicit asserts will
never trigger.
(From OE-Core rev: f1d2b2ec51f748a28d1bca6615558d553876e5c3)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When a test case starts, self.td is populated with all the variables in
the data store. Typically this can be used instead of get_bb_var(),
which saves a bitbake call per variable lookup.
The only catch is that in parallel runs the build directory is moved
after td is populated, so paths in the build directory are wrong: these
still need to be fetched in the test.
(From OE-Core rev: 884201c6c1bbf7c1b958bab7d7c91e27577eeaac)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Split the tests into further classes: one which exercises the CLI and
doesn't need to build images at all, and another which is just the
tests that manipulate existing images.
(From OE-Core rev: c9bc4def71325dba7b7ad93001f7fe1acced0bea)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
OEScriptTests currently skips if cairo isn't present, and does a build
of core-image-minimal. This is only required for the
OEPybootchartguyTests tests, so move that logic there so that the
OEListPackageconfigTests run even if cairo isn't available.
This leaves OEScriptTests as a simple class containing the scripts_dir
assignment, which can then be reused by other tests to avoid code
duplication.
(From OE-Core rev: fe3ee517772ad8396fcac21ae7eb8d1c8b68fbdb)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
There may be environments or machines which don't have working runqemu,
so tag all of the tests which use runqemu() so that they can be skipped.
(From OE-Core rev: 3f45ce6d2b1dfde8bc3d554397d55f81846c52d5)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
has_feature() should be splitting the feature string into substrings and
then looking for membership instead of looking for simple substrings.
has_machine() should be using equality instead of substrings.
(From OE-Core rev: a4c63819234e252c58e040af8bbdbfb96b6feccf)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This file dates back to 2016. Half of the functions have never been used,
the rest are used in one place and have now been replaced.
(From OE-Core rev: 5a053b2a84e7a671925fb4a78005145786c57f6a)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
There's no need to use a series of over-generalised functions to just
wrap a string in a tuple.
(From OE-Core rev: 080854fe346a76f5fbe25058ba1b2425a0459b5e)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
test_arch_work_dir_and_export_source uses the archiver to generate SRPMS,
so explicitly set PACKAGE_CLASSES to ensure that package_rpm is used.
(From OE-Core rev: 030157fa8dd405be60c55523e074b2ec950d36ad)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Some tests which are marked as x86-specific will actually work on
aarch64 (e.g. use EFI), whilst some other tests really are x86-specific
(e.g. use syslinux).
(From OE-Core rev: 1285bdaa4f472519083c03946ee34c7d8c204e27)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
There's no need to pass a recipe name when determining the target
architecture, there's no need to cap the size of the lru_cache as it
will only have one entry, and __name__ is set by @wraps.
(From OE-Core rev: e8e6c679f6eb74cb25c124a18af88dd5c2e2c833)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
By default bitbake() will raise an assertion if it fails, so there's no
need to wrap it in a further assert.
(From OE-Core rev: de3c8994dc482cf5e9f3317c4762fe4ac35d9f31)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
bb.utils.rename() only exists to handle moves across filesystems. As
these moves are within the same directory we can just use os.rename().
(From OE-Core rev: 450e09b6d4ca019848aec4c62fce280a09395b97)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The 'devtool upgrade' tests fail if Git doesn't know the user's name or
email, so verify this before the tests start and skip if it is not.
(From OE-Core rev: b4d8aca97a1ead38ce55f3bb5859d3d05d9dd84f)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
These image types use syslinux which is only available on x86, so only
add them to IMAGE_TYPES on x86.
(From OE-Core rev: 2ea047a026dd61a8e0a24c6bbe278849485a2c27)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
BitBake can optionally 'import yaml' if BB_LOGCONFIG specifies a yaml
file. This is a 3rd party module, so that this works out of the box
when buildtools is used -- either explicitly via buildtools-tarball or
implicitly via eSDK -- we can add pyyaml to the buildtools.
(From OE-Core rev: 6be90f884bb3fc87d9aa21cb882a835d6bc583a9)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We can re-trigger this code and there is little point in stacking a ton of duplicate
paths which just waste time during searches for modules. This could in theory alter
layer module search order but that seems unlikely in common use.
(From OE-Core rev: dadce8468db1c0fd0e04801cdc6cf287c2808477)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The patch creates ${MULTIMACH_TARGET_SYS}-toolchain.cmake file
at ${SDK_INSTALL_DIR}/sysroots/${SDK_SYS}/usr/share/cmake/, which is
per-toolchain CMake toolchain file containing arch-specific values
and independent of OE environment variables.
The file gets created after installing SDK toolchain installer
ined by running "bitbake -c populate_sdk <image>".
The changes are similar to meson-setup.py which is used to
create arch-specific
${SDK_INSTALL_DIR}/sysroots/${SDK_SYS}/usr/share/meson/*-meson.cross
[YOCTO #14644]
Tested-by: Jan Dorniak <jaskij@gmail.com>
(From OE-Core rev: 42e68397ec74b3cd8ae5df45355c8f6254b48cd8)
Signed-off-by: Jagadeesh Krishnanjanappa <workjagadeesh@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Some graphics driver implementations ( e.g sgx ) do not yet support APIs from mesa
21.x, and some portions of weston simple clients depend on these APIs,
therefore introduce a way to specify all or a selection fo clients to
build
Fixes
clients/weston-simple-dmabuf-feedback.p/simple-dmabuf-
feedback.c.o: in function `create_dmabuf_buffer':
| simple-dmabuf-feedback.c:(.text+0x1076): undefined reference to `gbm_bo_get_fd_for_plane'
(From OE-Core rev: 400aae43d08f0b9f787ac0d21cb3c97058d76748)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
It can be useful to use git on target (e.g. with some wrapper like
etckeeper for keeping track of changes to /etc), and for such cases,
it is likely one has no need for pulling from/pushing to http[s]
repositories. From the INSTALL file:
- "libcurl" library ... If you do not use http:// or https://
repositories, and do not want to put patches into an IMAP
mailbox, you do not have to have them (use NO_CURL).
- "expat" library; git-http-push uses it for remote lock
management over DAV. Similar to "curl" above, this is
optional (with NO_EXPAT).
Setting --without-expat and --without-curl reduces the size of the
installed "git" package from 18M to 12M, in addition to avoiding
pulling those libraries into the rootfs.
(From OE-Core rev: 49f81198c5d233a9a2612c3b8366681dd85bea59)
Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The missing gitsm:// mappings looks like an oversight, add them.
(From OE-Core rev: 6600b9fca7888fb41647cd000b9efb7f0762dfde)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When running g-ir-scanner, we get the following error:
Couldn't find include 'GObject-2.0.gir' (search path:
'['/usr/lib64', 'gir-1.0', '/usr/local/share/gir-1.0', '/usr/share/gir-1.0',
'/usr/share/gir-1.0', '/usr/share/gir-1.0']')
This is because g-ir-tool-template.in is not setting girdir correctly.
It's using the prefix instead of the actual girdir.
Note that we don't get such error at do_rootfs time because the extra
directories are speicified. But we will get such error at runtime when
multilib is enabled.
(From OE-Core rev: 4bc68b6a8187947516615e47f22a6b16d71c3b37)
Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Until now the CVE checker was giving information about CVEs found for
a product (or more products) contained in a recipe. However, there was
no easy way to find out which products or recipes have no CVEs. Having
no reported CVEs might mean there are simply none, but can also mean
a product name (CPE) mismatch.
This patch adds CVE_CHECK_COVERAGE option enabling a new type of
statistics. Then we use the new JSON format to report the information.
The legacy text mode report does not contain it.
This option is expected to help with an identification of recipes with
mismatched CPEs, issues in the database and more.
This work is based on [1], but adding the JSON format makes it easier
to implement, without additional result files.
[1] https://lists.openembedded.org/g/openembedded-core/message/159873
(From OE-Core rev: d1849a1facd64fa0bcf8336a0ed5fbf71b2e3cb5)
Signed-off-by: Marta Rybczynska <marta.rybczynska@huawei.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Add an option to output the CVE check in a JSON-based format.
This format is easier to parse in software than the original
text-based one and allows post-processing by other tools.
Output formats are now handed by CVE_CHECK_FORMAT_TEXT and
CVE_CHECK_FORMAT_JSON. Both of them are enabled by default.
The JSON output format gets generated in a similar way to the
text format with the exception of the manifest: appending to
JSON arrays requires parsing the file. Because of that we
first write JSON fragments and then assemble them in one pass
at the end.
(From OE-Core rev: df567de36ae5964bee433ebb97e8bf702034994a)
Signed-off-by: Marta Rybczynska <marta.rybczynska@huawei.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Changelog:
==========
- Changed urllib3[brotli] extra to favor installing Brotli libraries that are
still receiving updates like brotli and brotlicffi instead of brotlipy.
This change does not impact behavior of urllib3, only which dependencies are
installed.
- Fixed a socket leaking when HTTPSConnection.connect() raises an exception.
- Fixed server_hostname being forwarded from PoolManager to HTTPConnectionPool
when requesting an HTTP URL. Should only be forwarded when requesting an HTTPS URL.
(From OE-Core rev: 1c44078db4e8fc3ed992ede38708bea0dcf87f11)
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Changelog:
=========
Fixed support for pytest 7.0, and pytest>=7.0 is now required.
(From OE-Core rev: 34f6bc8ca0cfc310fd6ba494b995fa86d28b5a6e)
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Changelog:
=========
Improve error detection and message when Hypothesis is
run on a Python implementation without support for "-0.0", which is
required for the "floats()" strategy but can be disabled by unsafe
compiler options (issue #3265).
If the "shrink" phase is disabled, stop the "generate" phase as
soon as an error is found regardless of the value of the
"report_multiple_examples" setting, since that's probably what you
wanted (issue #3244).
(From OE-Core rev: 56702a6c8e066d3730dd336eeb98d10534226601)
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
License-Update: year updated to 2022
Changelog:
=========
- Handle RSASSA-PSS in keys.PrivateKeyInfo.bit_size and
keys.PublicKeyInfo.bit_size
- Handle RSASSA-PSS in keys.PrivateKeyInfo.wrap and keys.PublicKeyInfo.wrap
- Updated docs for keys.PrivateKeyInfo.algorithm and
keys.PublicKeyInfo.algorithm to reflect that they can return "rsassa_pss"
(From OE-Core rev: 8fbe3bd4aca7a8906e342bcc9f27e205398919c3)
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Use the new MOUNT_COPYBIND_AVOID_OVERLAYFS flag provided by mount-copybind.
When SELinux is enabled, processes accessing OverlayFS mounts will get a denial
if the process setting up the mount doesn't have all the permissions that
the accessor has.
(From OE-Core rev: 6002bdc77643c363a8326bf163baecba8b36e3e0)
Signed-off-by: Luca Boccassi <luca.boccassi@microsoft.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
It was found when a end user wants to build a squashfs type initramfs
into fitimage, it just fails without printing out any error or warning
messages, which is not right.
Introduce a FIT_SUPPORTED_INITRAMFS_FSTYPES variable to avoid
hard-coding the supported initramfs types, and it could be overridden
in config files. Also break the build when none of a supported
initramfs type is found.
(From OE-Core rev: b0f8d8a4c063936b50d3ec7c066b29157b3c3544)
Signed-off-by: Ming Liu <liu.ming50@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
