Now that xilinx-embeddedsw is compatible with the toolchain vendor buildroot,
this patch modifies the toolchain-bare-metal-buildroot help example to use
the microblazeel-buildroot-elf tuple.
Signed-off-by: Neal Frager <neal.frager@amd.com>
Signed-off-by: Julien Olivain <ju.o@free.fr>
This patch renames the bare-metal toolchain vendor used by the
xilinx-embeddedsw package from Xilinx to Buildroot to be consistent with all
other toolchains built by Buildroot.
To build the Microblaze applications available with the xilinx-embeddedsw
package, the following config is now needed:
BR2_TOOLCHAIN_BARE_METAL_BUILDROOT_ARCH="microblazeel-buildroot-elf"
This change keeps backwards compatibility for users already using the
following architecture tuple:
BR2_TOOLCHAIN_BARE_METAL_BUILDROOT_ARCH="microblazeel-xilinx-elf"
Either vendor name is now valid, but there is a warning message to use the
Buildroot vendor name, and the support for the "microblazeel-xilinx-elf"
tuple will be removed once all zynqmp and versal defconfigs have migrated.
Signed-off-by: Neal Frager <neal.frager@amd.com>
Signed-off-by: Julien Olivain <ju.o@free.fr>
With the Segmented Configuration, the Versal PDI files will have new names:
<design>_boot.pdi - Minimal DDR and PS config for booting
<design>_pld.pdi - Remainder of FPGA design to be loaded at run-time
Since this will be the new default, this patch aligns the Buildroot file
naming by replacing the obscure vpl_gen_fixed.pdi filename with a simpler
boot.pdi filename.
Signed-off-by: Neal Frager <neal.frager@amd.com>
Reviewed-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Signed-off-by: Julien Olivain <ju.o@free.fr>
AMD has created a new segmented configuration for Versal products. It splits
the Vivado hardware design into two PDI files, one containing the minimal
required configuration for the DDR and booting the processors, and a second PDI
file which contains the rest of the FPGA design and can be loaded at run-time
via U-boot or Linux.
The file names generated when using the Vivado Segmented Configuration are as
follows:
<design>_boot.pdi - Minimal DDR and PS config for booting
<design>_pld.pdi - Remainder of FPGA design to be loaded at run-time
Since two PDI files will be included in the XSA file when using Segmented
Configuration, the xilinx-prebuilt package needs to be updated to support
this feature.
For Buildroot purposes, the <design>_boot.pdi is the file that needs to be
included in the boot.bin for booting the processors, so this patch checks
for a file named *boot*.pdi which will indicate that Segmented Configuration
is being used and will make sure to use the <design>_boot.pdi file and not
the <design>_pld.pdi file when generating the boot.bin image.
If no pdi files contain the word "boot" in the filename, it can be assumed
that Segmented Configuration is not being used, so the single file *.pdi
method is the appropriate one for this case, and this patch is thus
backwards compatible with prior designs not using Segmented Configuration.
Also, Segmented Configuration is going to become the default mode for Versal
products, so the xilinx-prebuilt github location will also soon have two PDI
files for each board. For this reason, this patch is also handling Segmented
Configuration for files downloaded from the xilinx-prebuilt repo.
For further information about the AMD Segmented Configuration, please see the
github tutorial below.
https://github.com/Xilinx/Vivado-Design-Tutorials/tree/2024.2/Versal/Boot_and_Config/Segmented_Configuration
Signed-off-by: Neal Frager <neal.frager@amd.com>
Reviewed-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Signed-off-by: Julien Olivain <ju.o@free.fr>
The versal vpl_gen_fixed.pdi file is not a target executable, so the file
permissions should be 644 instead of 755.
Signed-off-by: Neal Frager <neal.frager@amd.com>
Reviewed-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Signed-off-by: Julien Olivain <ju.o@free.fr>
CoolPi 4B is a rk3588s based SBC[0].
Specification:
- Rockchip RK3588S
- LPDDR4 2/4/8/16 GB
- TF scard slot
- eMMC 8/32/64/128 GB module
- Gigabit ethernet drived by PCIE with RTL8111HS
- HDMI Type D out
- Mini DP out
- USB 2.0 Host x 2
- USB 3.0 OTG x 1
- USB 3.0 Host x 1
- WIFI/BT module AIC8800
- 40 pin header
[0] https://www.cool-pi.com/product/cp4b/
Signed-off-by: Andy Yan <andyshrk@163.com>
[Julien:
- add missing BR2_GLOBAL_PATCH_DIR and hash files
- add missing BR2_TARGET_UBOOT_NEEDS_GNUTLS
]
Signed-off-by: Julien Olivain <ju.o@free.fr>
This test verifies that clinfo correctly detects the rustiCL platform
and llvmpipe as device.
We check:
- exit code,
- Platform name matches rusticl,
- Device name matches llvmpipe.
Signed-off-by: El Mehdi YOUNES <elmehdi.younes@smile.fr>
[Julien:
- rename file to test_mesa3d.py
- add Medhi in DEVELOPERS for this test file
]
Signed-off-by: Julien Olivain <ju.o@free.fr>
RustiCL is an OpenCL implementation written entirely in
RUST and integrated into Mesa3D. It enables the use of OpenCL
programs on top of existing Gallium drivers. RustiCL is a
modern and maintainable replacement for the old C-implemention
Clover.
To build Rusticl in Mesa3D, the following components are required:
- host-rustc (Rust compiler)
- host-rust-bindgen (for generating bindings)
- spirv-tools
- spirv-llvm-translator (providing libLLVMSPIRVLib.so matching the LLVM version)
- LLVM built with libclc and -DLLVM_ENABLE_DUMP=ON
- Mesa headers via mesa-clc-bundle-headers=enabled
- There is a known meson bug and in order to fix it we set -Drust_std=2021, see:
https://docs.mesa3d.org/rusticl.html#known-issueshttps://github.com/mesonbuild/meson/issues/10664
More details in the Mesa documentation:
https://docs.mesa3d.org/rusticl.html
Rusticl is the provider of OpenCL in Buildroot.
However, Mesa does not provide the `libOpenCL.so` library itself.
We therefore depend on the `opencl-icd-loader` package to
install this loader in the toolchain sysroot, so that other
OpenCL-dependent packages can build and link against it.
Without the ICD loader, only `libRustiOpenCL.so` is available and
tools like `clinfo` or `clpeak` fail.
- 'Dmesa-clc-bundle-headers=enabled' ensures that the OpenCL header
files are bundled and available in the target. Without this,
runtime compilation of OpenCL kernels via clpeak or clinfo fails
with:
fatal error: 'opencl-c-base.h' file not found
see 3619611b31
We add a Rust compiler override to Meson to use Buildroot’s
host Rust toolchain using MESA3D_MESON_EXTRA_BINARIES.
Rusticl requires Meson to invoke rustc with the correct target and linker:
* 'rust=['$(HOST_DIR)/bin/rustc','--target=$(RUSTC_TARGET_NAME)']'
* 'rust_ld='$(TARGET_CROSS)gcc''
Another solution was proposed to set rust and rust_ld globally in
cross-compilation.conf, we prefer to define them locally within mesa
for now to ensure precise control over the rust toolchain behavior
during the Mesa build.
see https://lore.kernel.org/buildroot/ZcOeEMfwX4XlRXO7@landeda/
Currently, we should always use RUSTICL_ENABLE= driver to use rusticl.
More information:
https://docs.mesa3d.org/envvars.html#envvar-RUSTICL_ENABLE
Signed-off-by: El Mehdi YOUNES <elmehdi.younes@smile.fr>
[Julien: add references to meson bug]
Signed-off-by: Julien Olivain <ju.o@free.fr>
This adds the OpenCL installable client driver (ICD) loader
which provides the 'libOpenCL.so' shared library needed to
run OpenCL programs.
Mesa3D can offer opencl implementations (such as rusticl),
but it does not provide 'libOpenCL.so'.
Without it, programs like 'clinfo' fail to build not
finding 'libOpenCL.so'.
see:
https://github.com/KhronosGroup/OpenCL-ICD-Loader
Signed-off-by: El Mehdi YOUNES <elmehdi.younes@smile.fr>
[Julien: add Medhi in DEVELOPERS for this package]
Signed-off-by: Julien Olivain <ju.o@free.fr>
MessagePack ships both in a C++ and a C variant. Add the latter to
Buildroot. This will be used by fluent-bit.
Signed-off-by: Thomas Devoogdt <thomas@devoogdt.com>
[Julien:
- switch homepage URL to use https
- add a note in package/msgpack{,-c}/Config.in to clarify C/C++
variants
]
Signed-off-by: Julien Olivain <ju.o@free.fr>
For change log, see:
https://gitlab.com/gpsd/gpsd/-/blob/release-3.26.1/NEWS
Support for RTCM104v2 and v3, as well as for ublox, are now
unconditionally built, so drop our corresponding options. No
needfor legacy options: users that had those enabled will
still get them.
Support for Oceanserver has been dropped upstream, so drop our
option and add legacy handling to infrom users they should
complain upstream if they needed it (very unlikely).
Also, a bug introduced in gpsd 3.26 makes the Buildroot gpsd runtime
test failing. See [1]. As a workaround, this commit also changes
this runtime test to emulate a gps over TCP, instead of UDP.
[1] https://gitlab.com/gpsd/gpsd/-/issues/337
Signed-off-by: Yann E. MORIN <yann.morin@orange.com>
[Julien:
- rebased on branch next
- add gpsd version removing OceanServer option in Config.in.legacy
- add link to NEWS in commit log
- switch test_gpsd to use tcp instead of udp
]
Signed-off-by: Julien Olivain <ju.o@free.fr>
The biggest change for 3.14 is the maximum inittab entry size has increased
from 127 charachters to 253 characters long AND it logs a warning when a line
longer than this is found AND it refuses to run a line longer than 253
characters. It doesn't truncate too-long lines anymore, it just drops them.
Changelog: https://github.com/slicer69/sysvinit/releases/tag/3.14
Signed-off-by: Adam Duskett <adam.duskett@amarulasolutions.com>
Signed-off-by: Julien Olivain <ju.o@free.fr>
Noteable changes:
- Add -Dipe=false and -Dsysupdated=disabled as defaults to CONF_OPTS and
HOST_CONF_OPTS. If another user has a use for either, they can submit a
patch at a later date.
- systemd-networkd now requires kernel >= 5.4 due to the inclusion of the
sockios.h header, which uses the SIOCGSTAMP_OLD macro. This macro was first
introduced in kernel 5.2, but the official README recommends 5.4 as a
baseline, so stick with that. Due to this change:
- systemd-networkd now depends on BR2_TOOLCHAIN_HEADERS_AT_LEAST_5_4
- default y only if BR2_TOOLCHAIN_HEADERS_AT_LEAST_5_4
- Add a section in migrating.adoc about the new requirement
- The sha256sum of README.md changed due to the addition of some CentOS CI
lines.
- Add the new license file LICENSES/alg-sha1-public-domain.txt.
For change log, see:
https://github.com/systemd/systemd/blob/v257.5/NEWS
Tested with `./utils/test-pkg -p systemd -a`
40 builds, 28 skipped, 0 build failed, 0 legal-info failed, 0 show-info failed
Tested with `./support/testing/run-tests tests.init.test_systemd`
All tests passed
Signed-off-by: Adam Duskett <adam.duskett@amarulasolutions.com>
Reviewed-by: Marcus Hoffmann <buildroot@bubu1.eu>
[Julien: add link to change log in commit log]
Signed-off-by: Julien Olivain <ju.o@free.fr>
Both packages where added as host-only dependencies to build
package/python-aiomqtt, and are not used by any other package. With
aiomqtt switching to hatchling they are not needed any more.
Signed-off-by: Fiona Klute (WIWA) <fiona.klute@gmx.de>
Signed-off-by: Julien Olivain <ju.o@free.fr>
Aiomqtt switched its build system to hatchling with 2.3.1, change
PYTHON_AIOMQTT_SETUP_TYPE and dependencies accordingly.
Signed-off-by: Fiona Klute (WIWA) <fiona.klute@gmx.de>
Signed-off-by: Julien Olivain <ju.o@free.fr>
Drop the -Djs_engine=duktape option as mozjs support has been removed entirely
and as such, the option no longer exists.
Tested with `./support/testing/run-tests tests.package.test_polkit`
Both tests pass.
Also tested with `./utils/test-pkg -p polkit -a`
40 builds, 8 skipped, 0 build failed, 0 legal-info failed, 0 show-info failed
Signed-off-by: Adam Duskett <adam.duskett@amarulasolutions.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
This fixes the following CVE:
- CVE-2025-47287: When Tornado's ``multipart/form-data`` parser encounters
certain errors, it logs a warning but continues trying to
parse the remainder of the data. This allows remote
attackers to generate an extremely high volume of logs,
constituting a DoS attack.
Fixes:
https://www.cve.org/CVERecord?id=CVE-2025-47287
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
[Julien: reword commit log]
Signed-off-by: Julien Olivain <ju.o@free.fr>
GOI now has a build option to disable tests [1], so use that and drop
the "disable tests" patch.
Meson now checks LD_LIBRARY_PATH during configure and sets the result
when calling g-ir-scanner [2]. This means overriding it in _NINJA_ENV
does not work any more, set LD_LIBRARY_PATH in _CONF_ENV instead.
The hash for giscanner/scannerlexer.l changed due to small code
additions.
Upstream changelog:
https://gitlab.gnome.org/GNOME/gobject-introspection/-/blob/1.84.0/NEWS?ref_type=tags
[1] 3e41addbd0
[2] d3f684559f
Signed-off-by: Fiona Klute (WIWA) <fiona.klute@gmx.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Since Python 3.13 BaseSelectorEventLoop.add_reader() and
BaseSelectorEventLoop.add_writer() use the mapping returned by
selector.get_map() to detect if a file object is already
registered. This fails with the implementation in gi.events._Selector
if some calls use a file object, and others the raw file descriptor.
Full upstream bug report:
https://gitlab.gnome.org/GNOME/pygobject/-/issues/689
This bug breaks package/python-aiomqtt, because its client object uses
file objects in some places for the connection socket, and the file
descriptor in others. The result is that the connection attempt times
out because source registration fails, and the Future that marks
successful connection never resolves.
This commit adds the fix as backported to PyGObject 3.50 [1] so it can
be cherry-picked to Buildroot stable versions using that version.
[1] https://gitlab.gnome.org/GNOME/pygobject/-/merge_requests/423
Signed-off-by: Fiona Klute (WIWA) <fiona.klute@gmx.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Bump bootgen to xilinx_v2025.1 release.
With version 2025.1, bootgen executable is now in the build/bin dir, so
the install step is changed accordingly.
The LICENSE file hash has changed as it was converted from DOS end of
lines to Unix end of lines.
Signed-off-by: Neal Frager <neal.frager@amd.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
