Add the ability to specify cortexr52hf to get a version that enables the
hard-float calling convention.
(From OE-Core rev: be83bc937a76efaead4690237db25459a57673fc)
Signed-off-by: Mark Hatle <mark.hatle@amd.com>
Signed-off-by: Mark Hatle <mark.hatle@kernel.crashing.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Add the ability to specify cortexr5hf to get a version that enables the
hard-float calling convention.
(From OE-Core rev: 992149023a896ecb38d07c54fa9ae9df82249a07)
Signed-off-by: Mark Hatle <mark.hatle@amd.com>
Signed-off-by: Mark Hatle <mark.hatle@kernel.crashing.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Currently hard-float is dependency upon the special TUNE_CCARGS_MFPU which
contains custom -mfpu= values. However, newer 32-bit architectures like
cortex-r52 use 'simd' instead. There is no 'simd' entry for -mfpu=
according to the GCC manual, it's more or less automatic based on the
cpu settings.
Add this as an exception to the TUNE_CCARGS_FLOAT setting of hard or softfp.
(From OE-Core rev: f122eb8b19528e1192fd0b68b0639d84aa288155)
Signed-off-by: Mark Hatle <mark.hatle@amd.com>
Signed-off-by: Mark Hatle <mark.hatle@kernel.crashing.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Current include logic goes into generic arm v8/v9 architecture tunes and
adds corresponding -march option after synthesizing it from various tune
fragments, this is fine for a machine which is using armv8/armv9 based
tunes but cortex tunes are intentionally using -mcpu option based on
selected tune value. So when cortex based default tune is selected for a
machine, it will add both -mcpu and -march to the compiler commandline
which can result in invalid combinations for this pair in gcc's own
logic. This can then result in compiler warnings/errors reporting this
e.g.
aarch64-yoe-linux-gcc -mcpu=cortex-a72.cortex-a53 -march=armv8-a+crc+crypto -mbranch-protection=standard
...
cc1: error: switch '-mcpu=cortex-a72.cortex-a53' conflicts with '-march=armv8-a+crc+crypto' switch and resulted in options '+crc+crypto' being added [-Werror]
This is seen in lot of configure test results in glibc 2.39 and the
warning is promoted to errors by gcc in some of these checks especially
with gcc-14, the logs also show it as warning in other places in
configure checks.
mcpu option will compute relevant march implicitly as it specifies a cpu
implementation and this will be the right value to use, therefore do not
specify -march when -mcpu is already describing the cpu.
(From OE-Core rev: e64f0c1b6ac5d598a79a21de5f3060f83cb9523e)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Cc: Ross Burton <ross.burton@arm.com>
Cc: Jon Mason <jdmason@kudzu.us>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
There are two types of cases: executables and POSIX shell scripts.
All test cases PASS.
Add xz-ptest to PTESTS_FAST because test duration less than 30s
on qemux86-64.
root@qemux86-64:~# ptest-runner xz
START: ptest-runner
2024-01-26T03:32
BEGIN: /usr/lib/xz/ptest
=== test_bcj_exact_size.c ===
PASS: test_exact_size
PASS: test_empty_block
(From OE-Core rev: 2704983f972e4fe1d4e0bee8491a07eb4f629346)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
1. Patch changes:
0004-Move-sysusers.d-sysctl.d-binfmt.d-modules-load.d-to-.patch
is removed because it has no real effect now. The /lib is now
/usr/lib because 'usrmerge' is a required distro feature for systemd.
0002-binfmt-Don-t-install-dependency-links-at-install-tim.patch is
refreshed for the new version to avoid patch-fuzz issue.
2. root user's home directory now defaults to "/root":
The sysuers.d/basic.conf is still modified to respect the ROOT_HOME
value, so if users set ROOT_HOME to "/home/root", the behavior is the
same as before. However, this is only for backward compatibility. With
this patch, The ROOT_HOME value is set to "/root" in init-manager-systemd.inc.
This is because systemd's source codes are hardcoding "/root", and other
values are not officially supported. See the list below.
$ grep -rl '"/root"' src/ | grep -v 'src/test'
src/core/namespace.c
src/basic/user-util.c
src/nss-systemd/nss-systemd.c
src/nspawn/nspawn.c
src/firstboot/firstboot.c
src/shared/userdb.c
src/shared/user-record.c
$ grep -rl /root network/ factory/ sysctl.d/ sysusers.d/ rules.d/ tmpfiles.d/ units/ xorg/ tools/
sysusers.d/basic.conf.in
tmpfiles.d/provision.conf
units/emergency.service.in
units/rescue.service.in
tools/list-discoverable-partitions.py
Previously, the recipe was just substituting sysusers.d/basic.conf.in,
which is not enough to be treated as 'fully support'. I deliberately put
a warning message in do_install to warn users about non "/root" ROOT_HOME
value. Don't remove it until all above places are handled.
3. cgroupv2 is now the default.
cgroupv2 is the default for systemd for many years and it's the default
for distros such as ubuntu and fedora. Let's also use it as the default.
(From OE-Core rev: ebafe463799b39025a0b24a0a14a2f02b6de9bac)
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>
Add recipe for yamllint. There is an upcoming change in u-boot where
the binman tool is now configured to call yamllint to verify the configs
during compile time.
There was a previous patch a year ago from Trevor Woerner that never
made it into oe-core. This patch is a reworking of his patch but
pointing to a newer version.
(From OE-Core rev: 128cfc5222752a6337a9cbb9bc9023c13ee19f2f)
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We can drop the mips workaround patch since there were fixes in 8.2.0.
The build system changed and we should drop cross.patch and replace it with
explicit settings for cross-prefix, and host-cc. To make that work we need
to indicate we don't use a cross pkg-config.
PIE isn't availale for mingw so tweak PACKAGECONFIG accordingly.
(From OE-Core rev: 8917fa10b8afb1413b34a6134beea129e416f8c6)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
As agreed upon in the bug triage meeting, disable mdadm ptest for the time being
Related: [Yocto #15181] [Yocto #15159] [Yocto #15308] [Yocto #15309]
(From OE-Core rev: 22ae573aa1a1244d4dea498d4fa4fcdf195bedf8)
Signed-off-by: Pavel Zhukov <pavel@zhukoff.net>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This reverts commit d2b445384d.
Revert this commit since:
- some systems using oe-core master may still be using kernels from
before 5.6 pulled in the rng-tools algorithm, and
- some hardware platforms may not have a hardware random number generator
and could therefore need to run rngd to avoid long boot-time initialization
due to a depleted entropy pool.
(From OE-Core rev: 828afafb3bff54079fcba9bdab2ec87ac13e4ce6)
Signed-off-by: Randy MacLeod <Randy.MacLeod@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Nothing in oe-core depends on rng-tools anymore:
e7e1bc43ca rng-tools: splitting the rng-tools systemd/sysvinit serivce as a package
so move it to meta-oe for people who still want to run rngd
as a service for some reason or for those who want to run rng-test.
(From OE-Core rev: 9a651e5bc2492cf864261e8f17e4cbe34d6d48ce)
Signed-off-by: Randy MacLeod <Randy.MacLeod@windriver.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
shadow 4.14.x adds a number of libraries it dynamically links with
(md, bsd, attr). This causes troubles in setscene tasks where
shadow executables are used (such as useradd), as pulling in
the needed dynamic libraries needs unpleasant special-casing.
(From OE-Core rev: 495ff95eae14a91c94187f78a0b30c7957c9b168)
Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We need bitbake 2.7.1 due to the find_siginfo changes.
(From OE-Core rev: b8b6c3ebf6821b28dd1a350ccd74e47dc4bbd593)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Replace musl fixes with backports from upstream.
Add sysconfdir to config.toml to fix:
| thread 'main' panicked at install.rs:92:9:
| User doesn't have write access on `install.sysconfdir` path in `config.toml`.
https://blog.rust-lang.org/2023/11/16/Rust-1.74.0.html
(From OE-Core rev: 84f46dd2503bb0ef238fef0097c66fda88f6cbda)
Signed-off-by: Alex Kiernan <alex.kiernan@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Drop getrandom-open64.patch (merged upstream).
Revert c4f414f449
("Map source absolute paths to OUT_DIR as relative. (#684)") which
causes hashed versions of the build path to be used as part of the
filename of generated objects and hence reproducibility issues.
https://blog.rust-lang.org/2023/08/24/Rust-1.72.0.html
(From OE-Core rev: 30637cdeb31fae02544fdc643a455d0ebb126ee6)
Signed-off-by: Alex Kiernan <alex.kiernan@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* The tarfile module doesn't support xz options or environment varible
XZ_DEFAULTS, this makes do_ar_patched incrediblely slow when the file is
large, for example, chromium-x11 is about 3GB:
- "bitbake chromium-x11 -car_patched" hasn't been done after 3 hours on my
host, I checked the partial tar.xz file is only 1.5GB, so maybe more than 6
hours is required to complete the task.
- Now only less than 4 minutes is needed on the same host.
* Need add xz to HOSTTOOLS when archiver.bbclass is enabled and compression is xz.
(From OE-Core rev: 6548354f049b173e8d443bc547d35c9d9fc05259)
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
There is actually a use for nativesdk MACHINE_FEATURES; for example
qemu-usermode being supported, as this is needed to build profile-guided
optimised code.
We shouldn't use the target MACHINE_FEATURES for this because the target
and the SDK can be entirely different, so instead set the
MACHINE_FEATURES in nativesdk.bbclass to SDK_MACHINE_FEATURES (which
defaults to "") and let the conf/machine-sdk/*.conf files set that as
appropriate.
(From OE-Core rev: 14571764b7e046507f81bbe589a9f42c5b16665a)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Link: https://lore.kernel.org/r/20231221170159.1995650-2-ross.burton@arm.com
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Converts IPK package generation to use zstd instead of xz. zstd has a
much larger compression/speed tradeoff range allowing users to choose
what suits them best, and fast decompression speeds. It also continues
to support parallel compression as xz did.
A new variable called ZSTD_DEFAULTS is provided to set the defaults for
places that want to use zstd for compression; the zst image conversion
command is also modified to use this.
Finally, in order for this to function properly, opkg must include zstd
support, so it is enabled all the time with no PACKAGECONFIG to turn it
off.
(From OE-Core rev: 1bc3e9bbaa670b6128c74c76b4b5264e60ce3463)
Signed-off-by: Joshua Watt <JPEWhacker@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
As per the Arm ARM section A2.3.2, FEAT_CRC32 is mandatory in Armv8.1
onwards, so there's no need to have an explicit optional tune for it.
SVE (which was a choice here) and SVE2 (which was not) are both optional
with v9.0, but GCC's -march=armv9-a defaults to enabling both SVE and
SVE2 on the grounds that there are no implementations of Armv9 that
don't support SVE2. This means there's no point having an explicit sve
feature as it's enabled out of the box.
(From OE-Core rev: 28fabebf812cdf8e76eb508115a4b76944f1e729)
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 a minimal initramfs image recipe that just contains enough initramfs
to find the real root filesystem and pivot to it.
BSPs can use MACHINE_EXTRA_RRECOMMENDS to pull in any kernel module
packages that are required to bring up the hardware so that the rootfs
can be found.
(From OE-Core rev: 3e862cddf28e58957b7c6d9c983fdb40eb5442fb)
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>
Adds the path to the runtime state directory (/run). In particular,
systemd units need to have the correct path to the runtime directory
because RequiresMountsFor doesn't follow symbolic links. This means that
if a unit calls out a directory in /var/run (a symbolic link to /run),
it may actually start before /run is mounted. The fix is to have the
unit specify a directory in /run instead.
(From OE-Core rev: 90bc18108230f6d41a50cebc8348444e119e95bf)
Signed-off-by: Joshua Watt <JPEWhacker@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Build and publish crates with pyo3, rust-cpython, cffi and uniffi bindings
as well as rust binaries as python packages.
This project is meant as a zero configuration replacement for setuptools-rust
and milksnake. It supports building wheels for python 3.7+ on windows, linux,
mac and freebsd, can upload them to pypi and has basic pypy and graalpy support.
https://github.com/pyo3/maturin
* Referesh -crates.inc for 1.4.0
(From OE-Core rev: b92d923991e57e1e3a8e4a3bf1b7f7ec138f7723)
Signed-off-by: Tim Orling <tim.orling@konsulko.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Require the presence of a recent change in bitbake 2.6.1 for runqueue
setscene dependency handling improvements.
(From OE-Core rev: a63bd8d5aa2cdcbfa899b1cfb8acad06dee44752)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
UEFI bootloader specification says the primary place for UKI images
is the /EFI/Linux/ directory under $boot. Defining a varaible for
the same to reuse in all UEFI-related packages.
(From OE-Core rev: f32a3f55a6602bb71effe28c931ac387f71af640)
Signed-off-by: Viswanath Kraleti <quic_vkraleti@quicinc.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
systemtap-uprobes package was not used for a long time - since kernel
itself provided uprobes support. Now source code of old uprobes kernel
module was removed from systemtap git repo by "PR30434 continuation:
Removed old uprobes, uprobes2 implementation, uprobes-inc.h & any
mentions of CONFIG_UTRACE." it is good time for us to gid rid of it
too.
(From OE-Core rev: 42fd5abedb835b6f87721674001c52304e43cfc5)
Signed-off-by: Victor Kamensky <victor.kamensky7@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The current description is only pertinent to the kernel, even though
do_menuconfig task is used by other projects, such as Busybox and
U-Boot.
Replace "for the kernel" by an agnostic alternative (i.e., "in the
compilation directory").
(From OE-Core rev: 52e053bce5e359995ebdaa21d6899f04ad2306a0)
Signed-off-by: Joao Marcos Costa <joaomarcos.costa@bootlin.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The gcp fetcher uses gsutil to check if the file exists before
fetching. This change ensures the tool is included as a nonfatal
hosttool so that it's included in the build environment when
available.
(From OE-Core rev: d5e481bbd96210e608fa0aa5f7b54c1398c7d273)
Signed-off-by: Charlie Johnston <charlie.johnston@loftorbital.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* Allow glslang to find external spirv-tools to build.
* Vulkan-validation-layers has been split. So a new recipe
Vulkan-Utility-Libraries has been added and vulkan-validation-layers now
depends on it.
(From OE-Core rev: 63bdeb74b133840f9d6f55093bf16f5a8a5d0ab7)
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
By setting this, event data from during the build is saved out. This can
be used for analytic including toaster to allow builds to be examined
later. This will use some disk space but it is worth it for the ability
to look into what happened during the build. Users can set the value to
something else in local.conf to override if they wish.
(From OE-Core rev: 4719b68bb9e3cccb85aed55350e53a439360abcb)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* ${EXTENDPE} and ${PR} was already dropped from ${WORKDIR} in:
https://lists.openembedded.org/g/openembedded-core/message/185818
adjust -fmacro-prefix-map and -fdebug-prefix-map similarly.
As we don't depend on PR in WORKDIR maybe we don't need it in
/usr/src/debug paths as well anymore.
* add TARGET_DBGSRC_DIR variable with "/usr/src/debug/${PN}/${PV}"
because there is quite a few places where this location is being
used in various recipes
* The interesting side-effect of this is that for DISTROs which still
use PR (e.g. webOS OSE), the difference in -fdebug-prefix-map causes
all binaries to be slightly different when rebuilt with different PR
(due to this path, even when otherwise they are binary reproducible)
and when the number of digits in PR changes (e.g. from r9 to r10) it
also results in slightly bigger binaries (and more unnecessary changes
in buildhistory diffs).
For example this foo binary, the difference between "new" and "old" is
extra patch applied in SRC_URI (which doesn't affect the foo binary,
but was the reason for PR bump).
And the binary is bigger with r10, identical with r6 and slightly
different due to debug-prefix-map between r6 and r7:
$ ls -la 1.0.0-175-*/image/usr/sbin/foo
-rwxr-xr-x 1 bitbake bitbake 30182460 Nov 8 08:29 1.0.0-175-new-r10/image/usr/sbin/foo
-rwxr-xr-x 1 bitbake bitbake 30182224 Nov 8 08:02 1.0.0-175-new-r6/image/usr/sbin/foo
-rwxr-xr-x 1 bitbake bitbake 30182224 Nov 8 07:49 1.0.0-175-new-r7/image/usr/sbin/foo
-rwxr-xr-x 1 bitbake bitbake 30182224 Nov 8 07:39 1.0.0-175-old-r6/image/usr/sbin/foo
$ md5sum 1.0.0-175-*/image/usr/sbin/foo
8e3ab67d596f8cc2f9c6320dab10af01 1.0.0-175-new-r10/image/usr/sbin/foo
f6ff1e64fe6affb9ba0d8f278cf7ed79 1.0.0-175-new-r6/image/usr/sbin/foo
6469a0b10aac233911e63c5ea97b03c0 1.0.0-175-new-r7/image/usr/sbin/foo
f6ff1e64fe6affb9ba0d8f278cf7ed79 1.0.0-175-old-r6/image/usr/sbin/foo
(From OE-Core rev: c84c8ba295eeaa0e0b7592364b0978f17a44c70b)
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>
Autoconf defines this as:
"The directory for installing C header files for non-GCC compilers."
Whilst this is something autoconf does allow changing, I find it hard
to believe it has much use in the wild now and that headers don't get
split like this in reality, it would probably only be useful on really
old unixes.. The values are the same in our configuration anyway.
Drop the value and just use includedir everywhere.
(From OE-Core rev: 506c91cbc6a604a84e37e53ccff430436369802e)
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>
Since ccache version 4.0, according to
https://github.com/ccache/ccache/blob/master/doc/NEWS.adoc#ccache-40
* An appropriate cache directory level structure is now chosen automatically. The cache_dir_levels (CCACHE_NLEVELS) configuration option has therefore been removed.
Therefore remove the option which has not been supported by ccache
recipe version since Yocto Hardknott.
(From OE-Core rev: 1d7c1b71e2639b8d069403920368292bd214347d)
Signed-off-by: Niko Mauno <niko.mauno@vaisala.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
machine features like vfat are needed for ptests to pass ( e..g. parted)
This brings it closer to what x86 qemu config looks like as well.
(From OE-Core rev: a6d97ecb25567ec7e56c061be9ab281b54b4d524)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
