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.
(From OE-Core rev: ef0d0734866505c1c6e0528a0423e7248afb3ff8)
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Invoking meson with just paths is a deprecated use of the binary and
causes warnings, so call 'meson setup' explicitly.
Whilst doing this cleanup, we can also call 'meson compile' and 'meson
install' instead of ninja.
(From OE-Core rev: b248cb1ada04f57a5c4590c7e5efa04f68bc8843)
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>
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.
(From OE-Core rev: 8ed78ec262b2502dc3b673b24a868a3eec616a20)
Signed-off-by: Alejandro Enedino Hernandez Samaniego <alejandro@enedino.org>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
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!
(From OE-Core rev: 96ff9baa8ead57504f40f362ed3a4aaa776d1b58)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
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.
(From OE-Core rev: 7120b09a33af4c9a18063c0f2e51fb598697e39c)
Signed-off-by: Ovidiu Panait <ovidiu.panait@windriver.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
There is no need to patch native python so that it looks in the target
sysroot; the same can be achieved with just an environment variable.
(From OE-Core rev: c9617c03bceee54dc540318cada392799b137bd5)
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>
There is really no reason why we can't: it only needs to be available
in two versions (native/target), and the correct one can be picked
via PATH priority order.
This eliminates two un-upstreamable patches, one of which relies
on soon to be removed distutils.
(From OE-Core rev: d0145ead0f80ba4bde8e24617a8725f38eda8339)
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>
Currently TOOLCHAIN is strictly set to gcc in kernel-arch.bbclass.
And this prevents any TOOLCHAIN changes for any kernel recipe.
This change makes TOOLCHAIN configurable as usual.
(From OE-Core rev: be1634fc35dcc81f0301d942064a6eed584e0704)
Signed-off-by: Alexey Smirnov <pyih.soft@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The prepended code with the i and j variables clash with similar code in uboot.inc, so they need to be unset once we are done using them.
This commit fixes the issue that was introduced in d6858c9 "u-boot: Rework signing to remove interdependencies".
(From OE-Core rev: 6f668f85d7e5b0a9d36198db865cf1e1a012b2e1)
Signed-off-by: David Bagonyi <david.bagonyi@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Currently, the KERNEL_DEBUG_TIMESTAMPS is not working as expected
at rebuild. That is, even if we set it to "1", the kernel build time
is not changed. The problem could be reproduced by the following steps.
1. bitbake core-image-minimal; start image and check `uname -a` output.
2. set in local.conf: KERNEL_DEBUG_TIMESTAMPS = "1"
3. bitbake core-image-minimal; start image and check `uname -a` output.
It's expected that after enabling KERNEL_DEBUG_TIMESTAMPS, the kernel
build time will be set to current date. But it's not. This is because
the compile.h was not re-generated when do_compile task was re-executed.
In mkcompile_h, we have:
"""
# Only replace the real compile.h if the new one is different,
# in order to preserve the timestamp and avoid unnecessary
# recompilations.
# We don't consider the file changed if only the date/time changed,
# unless KBUILD_BUILD_TIMESTAMP was explicitly set (e.g. for
# reproducible builds with that value referring to a commit timestamp).
# A kernel config change will increase the generation number, thus
# causing compile.h to be updated (including date/time) due to the
# changed comment in the
# first line.
"""
It has made it very clear that it will not be re-generated unless
we have KBUILD_BUILD_TIMESTAMP set explicitly. So we set this variable
explicitly in do_compile to fix this issue.
(From OE-Core rev: 1b68c2d2d385013a1c535ef81172494302a36d74)
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>
This class should be in classes-recipe as, like testimage, it is specific
to image recipes.
This also solves a regression where simply IMAGE_CLASSES += "testexport"
doesn't work if testimage isn't already in IMAGE_CLASSES. The testexport
class adds testimage to IMAGE_CLASSES, but by this point it's too late.
Now that the class can only be inherited via IMAGE_CLASSES, we can
simply inherit testimage directly.
(From OE-Core rev: 87fcee9fed915ff3af528b976d7d44737f9f1a95)
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>
SafeConfigParser was renamed to ConfigParser in 3.2, and the
SafeConfigParser alias will be removed in 3.12.
(From OE-Core rev: 71b3e7f71727137b4b996cc4160c9cc1581824b8)
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>
Add 7-Zip support in conversion types and commands. Default arguments of
7-Zip are compression level: 9, method: BZip2, extension: 7z. All these are
overridable from local.con for any other use case in user side. First goal
was to make an alternative zip conversion which can operate faster in
multi-thread systems (on many CPU cores), normal zip is just a single
thread compression. 7-Zip can do a compression in multi-thread processing,
so it is fine to use and speed up it and 7-Zip is supported also in many
platform to use.
(From OE-Core rev: 5931b2898b5fa964768c2abf9b5aa556859152b3)
Signed-off-by: Benjamin Szőke <egyszeregy@freemail.hu>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When building with CONFIG_MODVERSIONS=y and CONFIG_RANDSTRUCT=y we need
to copy the build assets generated for the randstrutc seed to
STAGING_KERNEL_BUILDDIR, otherwise the out-of-tree modules build will
generate those assets which will result in a different
RANDSTRUCT_HASHED_SEED.
(From OE-Core rev: b36303158b2e0273ff415bdedefb379f680b30fc)
Signed-off-by: Diego Sueiro <diego.sueiro@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* RUSTC is normally defined in rust.bbclass, but rust-target-config doesn't
inherit rust.bbclass and RUSTC is then undefined when calling 'which'
in oe_cargo_build
* causes which help text ending in log.do_compile, e.g.:
| DEBUG: Executing shell function do_compile
| NOTE: Using rust targets from /OE/build/oe-core/tmp-glibc/work/core2-64-oe-linux/openssl-sys/0.9.77-r0/rust-targets/
| NOTE: cargo = /OE/build/oe-core/tmp-glibc/work/core2-64-oe-linux/openssl-sys/0.9.77-r0/recipe-sysroot-native/usr/bin/cargo
| Usage: which [options] [--] COMMAND [...]
| Write the full path of COMMAND(s) to standard output.
|
| --version, -[vV] Print version and exit successfully.
| --help, Print this help and exit successfully.
| --skip-dot Skip directories in PATH that start with a dot.
| --skip-tilde Skip directories in PATH that start with a tilde.
| --show-dot Don't expand a dot to current directory in output.
| --show-tilde Output a tilde for HOME directory for non-root.
| --tty-only Stop processing options on the right if not on tty.
| --all, -a Print all matches in PATH, not just the first
| --read-alias, -i Read list of aliases from stdin.
| --skip-alias Ignore option --read-alias; don't read stdin.
| --read-functions Read shell functions from stdin.
| --skip-functions Ignore option --read-functions; don't read stdin.
|
| Recommended use is to write the output of (alias; declare -f) to standard
| input, so that which can show aliases and shell functions. See which(1) for
| examples.
|
| If the options --read-alias and/or --read-functions are specified then the
| output can be a full alias or function definition, optionally followed by
| the full path of each command used inside of those.
|
| Report bugs to <which-bugs@gnu.org>.
| NOTE: rustc =
(From OE-Core rev: 0068e82f02354729b953e622889f96ca52468d6c)
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Add SDK_ZIP_OPTIONS to remove symbolic link creation in zip archive or add options, e.g. for encryption of the zip archive.
(From OE-Core rev: 04b62f9459b401c276255f166d0738b6f902a576)
Signed-off-by: Christoph Lauer <christoph.lauer@xtronic.de>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* use 4 spaces
* avoid trailing space
* add CARGO_LOCK_SRC_DIR to allow searching outside ${S}
* use BPN in output filename
* First I've used CARGO_LOCK_SRC_DIR as relative to ${S}, because that's what CARGO_SRC_DIR
in cargo.bbclass is using:
meta/classes-recipe/cargo.bbclass:CARGO_SRC_DIR ??= ""
meta/classes-recipe/cargo.bbclass:MANIFEST_PATH ??= "${S}/${CARGO_SRC_DIR}/Cargo.toml"
but change to absolute path (defaulting to ${S}) as requested by Alex:
11:44 < kanavin> JaMa, would prefer CARGO_LOCK_SRC_DIR ??= "${S}"
11:46 < kanavin> otherwise looks good
* I've resolved my usecase for CARGO_LOCK_SRC_DIR by changing
S back to ${WORKDIR}/git and using CARGO_SRC_DIR to select
the right subdirectory to be built, because the Cargo.toml
in this subdirectory was also referencing other subdirectories
with relative path:
88b147506d/token/cli/Cargo.toml (L30)
so including all Cargo.lock files in whole ${WORKDIR}/git
seems like reasonable approach
(From OE-Core rev: 7636a2b8080521ed2ad54b0edce47a8742a12d58)
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Rust targets have the form of 'arch-unknown-linux-gnu' while
oe's native targets are 'arch-linux-gnu', e.g. omit the vendor.
The effect this has on rust-native builds is that rust first builds
itself as stage0 for arch-unknown-linux-gnu, then builds itself
again for arch-unknown-linux-gnu, then finally uses the compiler
from second step to 'cross-compile' a compiler for 'arch-linux-gnu'.
This last step is really not necessary, and we could save 4 minutes
out of 12 if it is eliminated. Which is what this patch does
by setting the target directly to 'arch-unknown-linux-gnu'; rust's
build system then shortcuts the build process after the second step.
Given a working rust-native will be needed as early as possible in a
typical yocto build (e.g. when in a not too distant future making a
useful kernel will not be possible without rust), producing it faster
is important.
(From OE-Core rev: a918ea5645d8a67cedaf3ecf6c382520bbcad85b)
Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
For better or worse, more and more rust components are appearing that do
not include their dependencies in tarballs (or git trees), and rely on cargo
to fetch them. On the other hand, bitbake does not use cargo (and quite possible
won't ever be able to), and relies on having each item explicitly listed in SRC_URI
with a crate:// prefix. This however creates a problem of both making such lists in
the first place and updating them when a recipe is updated to a newer version.
So this class can be used to perform such updates by implementing a task that does it;
the next commit shows the outcome for python3-bcrypt (which has been tested to work
and produce a successful build).
Note: the python script relies on tomllib library, which appears in Python 3.11 and
does not exist in earlier versions - I've tested this by first updating python to 3.11-rc2
in oe-core.
(From OE-Core rev: 9eee3631124d64574b18a70a2fc42f446d58bfd2)
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>
The semaphore fix has landed and is available from 3.11 onwards:
1ee0f94d16
Drop 0001-Mitigate-the-race-condition-in-testSockName.patch
as it is merged upstream.
(From OE-Core rev: f10cdc155e47af5627ee999c57e1d083f9382a91)
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>
When building a FIT image with device trees, each device tree lands in a
FIT section and is referenced by a FIT configuration node.
FIT images however also allow referencing the same device tree from
multiple configurations. This can be useful to reduce FIT image size
while staying compatible with existing bootloaders. Allow
kernel-fitimage.bbclass users to take advantage of this by mapping
each symlink to a regular device tree included in the FIT to a
configuration that references a common device tree section.
(From OE-Core rev: 21e240da63239826f3ef50ceef40c9519e9030d8)
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This introduces no functional change, but will come in handy in a later
commit where a file lookup will have us using the device tree name. If
we keep it like it's now, we will lose the information whether an
underscore is an original underscore or a mangled slash.
(From OE-Core rev: 8bea426ca59d17715a3b32f7e3caf3e4b6db5ce9)
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The U-Boot signing code is a bit of a mess. The problem is that mkimage
determines the public keys to embed into a device tree based on an image
that it is signing. This results in all sorts of contortions: U-Boot has to
be available to the kernel recipe so that it can have the correct public
keys embedded. Then, the signed U-Boot has to be made available to U-Boot's
do_deploy. This same dance is then repeated for SPL. To complicate matters,
signing for U-Boot and U-Boot SPL is optional, so the whole process must be
seamlessly integrated with a non-signed build.
The complexity and interdependency of this process makes it difficult to
extend. For example, it is not possible to install a signed U-Boot binary
into the root filesystem. This is first because u-boot:do_install must run
before linux:do_assemble_fitimage, which must run before u-boot:do_deploy.
But aside from infrastructure issues, installing a signed U-Boot also can't
happen, because the kernel image might have an embedded initramfs
(containing the signed U-Boot).
However, all of this complexity is accidental. It is not necessary to embed
the public keys into U-Boot and sign the kernel in one fell swoop. Instead,
we can sign the kernel, stage it, and sign the staged kernel again to embed
the public keys into U-Boot [1]. This twice-signed kernel serves only to
provide the correct parameters to mkimage, and does not have to be
installed or deployed. By cutting the dependency of
linux:do_assemble_fitimage on u-boot:do_install, we can drastically
simplify the build process, making it much more extensible.
The process of doing this conversion is a bit involved, since the U-Boot
and Linux recipes are so intertwined at the moment. The most major change
is that uboot-sign is no longer inherited by kernel-fitimage. Similarly,
all U-Boot-related tasks have been removed from kernel-fitimage. We add a
new step to the install task to stage the kernel in /sysroot-only. The
logic to disable assemble_fitimage has been removed. We always assemble it,
even if the final fitImage will use a bundled initramfs, because U-Boot
will need it.
On the U-Boot side, much of the churn stems from multiple config support.
Previously, we took a fairly ad-hoc approach to UBOOT_CONFIG and
UBOOT_MACHINE, introducing for loops wherever we needed to deal with them.
However, I have chosen to use a much more structured approach. Each task
which needs to use the build directory uses the following pseudocode:
do_mytask() {
if ${UBOOT_CONFIG}; then
for config, type in zip(${UBOOT_CONFIG}, ${UBOOT_MACHINE}); do
cd ${config}
mytask_helper ${type}
done
else
cd ${B}
mytask_helper ""
fi
}
By explicitly placing the work in mytask_helper, we make it easier to
ensure that everything is covered, and we also allow bbappends files to
more easily extend the task (as otherwise they would need to reimplement
the loop themselves).
[1] It doesn't particularly matter what we sign. Any FIT will do, but I
chose the kernel's because we already went to the trouble of setting it up
with the correct hashes and signatures. In the future, we could create a
"dummy" image and sign that instead, but it would probably have to happen
in the kernel recipe anyway (so we have access to the appropriate
variables).
(From OE-Core rev: 5e12dc911d0c541f43aa6d0c046fb87e8b7c1f7e)
Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
uboot_fitimage_assemble is called from one place with the same
arguments. Instead of using shell variables as intermediaries, simplify
the function by using the bitbake variables directly. Removing a layer
of indirection makes it easier to determine what values are being
substituted in. Some variables can't be fully converted, but they will
be addressed in the a few commits.
(From OE-Core rev: 27f42c9b9a91742d3ee358aa8dc29627379b2539)
Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We have a specific variable for the path to the boot directory. Use it
instead of open-coding this path.
(From OE-Core rev: 725b75e83bc2b2111f2ab5103b7e7f60d6d3f34e)
Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Replacing sysroot_stage_all by a no-op recipe makes it difficult for
bbappends to stage files intentionally. Instead, just clear
SYSROOT_DIRS, allowing other bbappends to easily add new directories.
(From OE-Core rev: 849791e7086463a4c7c53c2c1ed9603a6c3a080d)
Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When generating our SPL-verifying certificate, we use FIT_KEY_REQ_ARGS,
which is intended for the U-Boot-verifying certificate. Instead, use
UBOOT_FIT_KEY_REQ_ARGS.
Fixes: 0e6b0fefa0 ("u-boot: Use a different Key for SPL signing")
(From OE-Core rev: a2d939ccb182a1ad29280d236b9f9e1d09527af1)
Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Improve the fatal error message of the yocto-kernel-tools symbol_why.py
and shows the command that generate the error as it can help understand
the root cause of the error.
(From OE-Core rev: 54ae08779071f2e97bff0ff6514ede3124312c3b)
Signed-off-by: Jose Quaresma <jose.quaresma@foundries.io>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When machine configuration defines a mount point, which is not used in
any recipe, allow to fall through and only report a note in the logs.
This can be expected behavior, when a mount point is defined for several
machines, but not used in all of them
(From OE-Core rev: a9c604b5e0d943b5b5f7c8bdd5be730c2abcf866)
Signed-off-by: Vyacheslav Yurkov <Vyacheslav.Yurkov@bruker.com>
Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The externalsrc class was moved to classes-recipe as part of oe-core
f5c1280, but it can be used in both recipe and global contexts so move
it back to classes/.
(From OE-Core rev: 7a2edcd4b7cb5a2d829289a11eff62663268fbf3)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Since "package_manager: Change complementary package handling to not
include soft dependencies"[1], complementary packages don't pull in
recommendations, just depends. However, ptest.bbclass has a recommends
on ptest-runner, so if ptest packages are in an image via the ptest-pkgs
IMAGE_FEATURE, ptest-runner doesn't get installed.
[ YOCTO #14928 ]
[1] oe-core b44b0b9294675f89aa51ff84f532664f4c479677
(From OE-Core rev: ecff0642be5781f7f6cca617158b04ac9a0e85f0)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Git has removed support for "git submodule--helper list".
31955475d1
This change provides an alternate method for gathering the submodules
information.
Tested:
Build recipes with and without submodules
(From OE-Core rev: 6d9364e5f3535954f65cbbc694ee7933ac1d664f)
Signed-off-by: Carson Labrado <clabrado@google.com>
Signed-off-by: John Edward Broadbent <jebr@google.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
GTKIC_CMD variable gets the wrong assignation leading into a post
install script error. Fix if-else condition in GTKIC_CMD variable
to assign gtk4-update-icon-cache when GTKIC_VERSION is 4 but
gtk-update-icon-cache when is 3.
Also, rename gtk-update-icon-cache-3.0.0 to gtk-update-icon-cache-3.0
to match the gtk-update-icon-cache binary name deployed in
meta/recipes-gnome/gtk+/gtk+3.inc.
(From OE-Core rev: 95fc819fca11115623d0f1f7f0c11e9c46195b21)
Signed-off-by: Daniel Gomez <daniel@qtec.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
github has recently changed how the releases page is structured:
the tarballs are no longer listed directly, but are included
via separate 'fragment' URIs. For now, we can change the check
to match against the release tags.
This also establishes a common base URI to use for both
fetching and checking the latest version.
(From OE-Core rev: afa57a02ecd12ad176302d9631f4181b26d94f5c)
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>
Instead of installing run-postinsts with it's postinst scripts causing
systemd restarts, use the new dnf-test-* packages instead.
Remove from the installroot tests entirely as they're exercised enough
using just busybox.
Rewrite the exclude test to be simplier now these packages are not going
to be part of an existing dependency chain.
[ YOCTO #14787 ]
(From OE-Core rev: fb1de2abc53bd742bc55cfecd384b78852c10d80)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Using a binary string as the login banner search expression is fraught with
risks. We've seen cases on the autobuilder where "login:" is clearly shown
but the code hasn't triggered. The most likely cause is hidden control characters
in the output causing the search to fail.
Take the opportunity to remove the horrible binary string search, at the expense of
decoding the bootlog multiple times.
Tweak the logging so we can know which log was printed (self.msg or bootlog)
just in case this isn't the issue and we need more information in future.
(From OE-Core rev: 91b9e30e08695e715ef14c3df7471e8c99f9deb5)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This makes sure that the postrm script it using the right kernel paths.
(From OE-Core rev: f7b191f80d4da740089a301062e7ac0b82d1d242)
Signed-off-by: Andrei Gherzan <andrei.gherzan@huawei.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This makes sure that the postinstall script it using the right kernel
paths.
(From OE-Core rev: 5105820df4800673c188366a76c1a3bd387a7148)
Signed-off-by: Andrei Gherzan <andrei.gherzan@huawei.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When 73fa855f6a was introduced, the "cleandirs" variable flag
started applying to functions exported using EXPORT_FUNCTIONS.
The externalsrc class is supposed to remove cleandirs in order to
prevent wiping out an external src folder (home directory?), but
doesn't take the previous point into account. The result is that
cleandirs is still in effect.
To fix this, apply the cleandirs manipulation to all variables, not
just predefined ones. This is expensive, but since it executes inside
an `if externalsrc` clause, and EXTERNALSRC is usually only set for
specific single recipes, it won't affect most recipes.
Richard Purdie <richard.purdie@linuxfoundation.org> is the original
author of this patch. I just submitted it under my name for blame
purposes, and also we have been testing it a lot in my company.
This is the original discussion:
https://lists.openembedded.org/g/openembedded-core/topic/91374926
(From OE-Core rev: 74ad497a55007960a4869905878e3ccbd11e4369)
Signed-off-by: Kristian Amlie <kristian.amlie@northern.tech>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Add the ability to expose the lower layer of /etc when mounting overlay.
This is the similar to what overlayroot script from initramfs-framework does.
By default, this option is turned off to keep an old behavior intact.
(From OE-Core rev: 791e8a8bacce5a7f31f4d7bcbfb17df2967fd258)
Signed-off-by: Vyacheslav Yurkov <v.yurkov@precitec.de>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When using IMAGE_FEATURE read-only-rootfs ssh host keys are moved to volatile
storage. If the feature overlayfs-etc is used in addition to read-only-rootfs
/etc is writable and the move is not wanted. But in the case also the IMAGE_FEATURE
stateless-rootfs is used the keys will be moved as storage of keys should not
be wanted in a stateless-rootfs.
This change only takes effect in the case IMAGE_FEATURE contains read-only-rootfs.
In adddition the following cases are handled:
IMAGE_FEATURES = "read-only-rootfs" --> ssh keys/config handled as ro root
IMAGE_FEATURES = "read-only-rootfs overlayfs-etc" --> ssh keys/config handled as rw root
IMAGE_FEATURES = "read-only-rootfs stateless-rootfs" --> ssh keys/config handled as ro root
IMAGE_FEATURES = "read-only-rootfs overlayfs-etc stateless-rootfs" --> ssh keys/config handled as ro root
(From OE-Core rev: d2ad7aa1f2153955adc044ea4eb11c48086a01d1)
Signed-off-by: Peter Bergin <peter@berginkonsult.se>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Openssh (main) package may be marked for installation via complementary
packages mechanism if sftp-server is installed
and this causes conflict with dropbear [Yocto #14858] [1].
Excluding openssh complementary packages if packagegroup-core-ssh-dropbear
is in PACKAGE_INSTALL fixes this issue.
To install openssh complementary packages in the images with
ssh-server-dropbear they may be added manually into the list because
they will be excluded from the installation even if corresonding class
(dev-pkg or dbg-pkgs) inherited.
[1]
Error:
Problem: problem with installed package dropbear-2020.81-r0.core2_64
- package dropbear-2020.81-r0.core2_64 conflicts with openssh provided by openssh-8.9p1-r0.core2_64
- package openssh-8.9p1-r0.core2_64 conflicts with dropbear provided by dropbear-2020.81-r0.core2_64
- package openssh-ptest-8.9p1-r0.core2_64 requires openssh, but none of the providers can be installed
- conflicting requests
(try to add '--allowerasing' to command line to replace conflicting packages or '--skip-broken' to skip uninstallable packages)
(From OE-Core rev: fa08030b32c2bf77889c23f964892f46e84994a3)
Signed-off-by: Pavel Zhukov <pavel@zhukoff.net>
Signed-off-by: Pavel Zhukov <pavel.zhukov@huawei.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We're currently only passing in a subset of the kernel make flags to
menuconfig. Fix this to be consistent with all the other kernel
operations since these are becomming increasingly reliant on host
compilers and flags and target toolchains as well.
(From OE-Core rev: 8c616bc090d1834a21073a33209323220c05d2e5)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
