Since pmem kernel drivers were enabled as modules, the initrd
size limit is hit. On genericarm64 all kernel modules and some
firmware files get installed to initrd by default which make
the thing large. I'm working on patches to reduce the kernel drivers
installed to initrd and to make the kernel in general more modular
(btrfs 5 Mb etc built into kernel by default). For now just increase
the size limit to unblock genericarm64 builds and testing.
Fixes: https://bugzilla.yoctoproject.org/show_bug.cgi?id=15833
(From meta-yocto rev: 09bef6a491c087f0f0371874e759aae08a720c4e)
Signed-off-by: Mikko Rapeli <mikko.rapeli@linaro.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The AMD KV260 has a serial console on ttyPS1, so until we can dynamically
detect the console under sysvinit we can add it to SERIAL_CONSOLES so
this platform has a working console when not running systemd.
(From meta-yocto rev: c1824bbee5d7febe4154d683e895470c08b681a9)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Assignment in machine config can not be overwritten in local.conf
because former is evaluated later than latter. Use ?= conditional
assignment for u-boot and qemuboot variables so that they can be
overwritten in local.conf if a different test setup is wanted
for kernel, initrd and rootfs built for genericarm64 machine
config.
(From meta-yocto rev: fbb2b2a2dcfd4dd990071d4d73b149d4099a36ab)
Signed-off-by: Mikko Rapeli <mikko.rapeli@linaro.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
There have been no specific patches on beaglebone branche for many years,
so I have decided to switch to using the base branch for this machine.
This approach eliminates the need to adjust the corresponding
SRCREV_machine and LINUX_VERSION with each stable kernel version bump.
(From meta-yocto rev: 1570badbf93511f370614c21ccd967af28549b2c)
Signed-off-by: Kevin Hao <kexin.hao@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Some boards - such as the BeaglePlay - have the serial console on ttyS2.
Add a few ttyS? consoles to SERIAL_CONSOLES to cater for these boards.
(From meta-yocto rev: 6732332ef6ff640189e51ef6839598be693fd9e1)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The 6.10 build has been tested on both a BeaglePlay and qemu, so we can
remove the pinning now.
(From meta-yocto rev: d19d6046e96b1c94ee50a6140b1338a50a2cef20)
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 want to introduce the 6.10 linux-yocto recipe, but genericarm64
hasn't been valided for it yet. As such, let's temporarily set the
preferred version to 6.6 so the builds don't break.
(From meta-yocto rev: 6ba3b6f15bf2460432da3a3aee782976e778381c)
Signed-off-by: Bruce Ashfield <bruce.ashfield@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
genericarm64 supports runqemu, so testimage should work. However, out of
the box it does not:
runqemu - ERROR - BIOS .../tmp/deploy/images/genericarm64/u-boot.bin not found
That is because we make the user explicitly build the qemu-targetting u-boot
to avoid the confusing situation where they have real hardware but the
deploy directory contains a u-boot for qemu.
In automated testing situations, however, we can be a bit more helpful
and make testimage depend on u-boot. This will lead to u-boot binaries
being in deploy, but at this point the user is already running the images
inside a qemu.
Reported-by: James McGregor <James.McGregor2@arm.com>
(From meta-yocto rev: 90b45c62d34396a20078b55d7d36f66b4e2177f7)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
testimage on core-image-sato fails in parselogs and xorg tests. We can
add the virtio gpu to address these issues (like what is being done in
qemuarm, qemuarm64, and qemuarmv5).
(From meta-yocto rev: ddf075e2abfba6479efc2bc5fe1c41ba6079e355)
Signed-off-by: Jon Mason <jdmason@kudzu.us>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
A basic SystemReady IR system can be provided by qemu-system-aarch64
and u-boot, so tell u-boot to build the qemu_arm64 machine and configure
qemuboot to start that u-boot and search the virtio-attached wic image
for the EFI boot partition.
Currently this machine support emulated (Cortex-A76) and virtualised (KVM)
execution, and virtio storage/network/console. Display support will be
added shortly.
Note that this machine still doesn't build U-Boot by default, as a u-boot
binary for qemu in deploy would potentially confuse users who want to
boot on real hardware and think this u-boot is needed. If you wish to
use genericarm64 with runqemu, you'll need to manually bitbake u-boot.
(From meta-yocto rev: cd40f405844c1ccfabcbe64ad6be5d98d0221f72)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Don't install _all_ of the firmware, as that's huge (almost 1GB). Instead
install a few pieces of firmware for common hardware.
Also use the same list of packages to populate the initramfs, so there's
no need to manually sync the package lists (as initramfs doesn't install
the MACHINE_EXTRA_RRECOMMENDS automatically).
(From meta-yocto rev: a5aa914990f36cc5175577983dd1ad1aa0bb81f2)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This is a new 64-bit "generic" Arm machine, that expects the hardware to
be SystemReady IR compatible.
(From meta-yocto rev: 68de209f58917d8e7108caacfefc55bbe0e0c5a2)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Prepare for the scarthgap release.
(From meta-yocto rev: 5be4f867a3fba39103a569eb4ecf10c39eca3742)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Create the bbappend for v6.6 and bump the default kernel to v6.6 for
all the machines.
In the v6.6 kernel the dtb files for beaglebone have been moved to a
vendor specific directory, so we also need to adjust the corresponding
variables for its DTB generation.
(From meta-yocto rev: 58ac51b7a44fbb165311c7226ae8b15cd004114d)
Signed-off-by: Kevin Hao <kexin.hao@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Now we only support the zImage kernel type for this machine, so remove
the obsolete variables for uImage kernel type.
[YOCTO #15232]
Reported-by: Robert Berger <pokylinux@reliableembeddedsystems.com>
(From meta-yocto rev: a8a5abe05ec4e96287acc994883f4b977290cde1)
Signed-off-by: Kevin Hao <kexin.hao@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The kernel-devicetree is set in both MACHINE_ESSENTIAL_EXTRA_RDEPENDS
and MACHINE_EXTRA_RRECOMMENDS. Since it is an essential package for
boot, drop the one in MACHINE_EXTRA_RRECOMMENDS.
[YOCTO #15229]
(From meta-yocto rev: 5935d4fd751153f0f1c930924a5e2771f6dd5e72)
Signed-off-by: Kevin Hao <kexin.hao@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
GLIBC_ADDONS was removed from the glibc recipe in 2018 (oe-core 9dc9983).
matchbox-panel-2 uses the ACPI battery by default if the machine has the
'acpi' feature, which genericx86 has.
(From meta-yocto rev: 0e50dbf558fda7595f5cb13667cbe0a01481a9f5)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(From meta-yocto rev: 5746c1b425859f5b8ba3a434c92b1d8798bedd71)
Signed-off-by: Lee Chee Yang <chee.yang.lee@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The default XSERVER value is good enough for this BSP, so we don't need
to set it explicitly.
(From meta-yocto rev: a0077d9132d1e86f97fa3d4a4607b008622aa17e)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Drop the gma500-gfx-check script. This hardware is ancient and times have moved
on, stop carrying around this old compatibility script.
(From meta-yocto rev: fa57b62988010a398b0defa89c9e831d75618d14)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
As per ongoing discussions, there is a general desire to try and update
our hardware reference platforms when an appropriate (newer) substitute
is available. However, in the case of MIPS, there isn't an obvious one
to choose. But continuing to maintain support for an old board that
hasn't been available for purchase for years doesn't make sense either.
So we are dropping support for the Edgerouter for future releases. The
MIPS architecture will still get coverage via QEMU build/boot testing.
(From meta-yocto rev: 5ea231c6c6387804a85adf15d3f03bdae9b2be25)
Signed-off-by: Paul Gortmaker <paulg@kernel.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Create the corresponding 6.1 bbappend and also update the default
kernel to 6.1 for all the boards.
(From meta-yocto rev: 7494f08fa1c85611a1dce3444a2737ef06fc8fdf)
Signed-off-by: Kevin Hao <kexin.hao@windriver.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Appending IMAGE_INSTALL with the kernel and device tree is a bad idea
because it will bring these packages to every images, like a container
image or sysroot image to pivot to, where the kernel artifacts are
not wanted.
Instead use the MACHINE_ESSENTIAL_EXTRA_RDEPENDS which results
in having the kernel artifacts where they should be, part of
packagegroup-core-boot.
(From meta-yocto rev: d4cbe0e43393ec70e98caf006372b5311597d377)
Signed-off-by: Vivien Didelot <vdidelot@pbsc.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
If BBLAYERS contains a layer providing a package for
virtual/bootloader, building for the beaglebone-yocto machine will
fail. Setting u-boot as the preferred provider for virtual/bootloader
ensures u-boot is choosen and the build succeeds.
(From meta-yocto rev: 2ffbaf54f4d88ba251d282c90755ea964d224ba6)
Signed-off-by: Vivien Didelot <vdidelot@pbsc.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
5.14 has been removed from core, so we change our default version
to the 5.15 LTS kernel.
(From meta-yocto rev: a0241305e3785930f806529fd1b1d1a3831203d0)
Signed-off-by: Bruce Ashfield <bruce.ashfield@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Introduce the bbappend for v5.14 and also set the default kernel to
v5.14 for all the boards.
(From meta-yocto rev: bed70a64b5b772fd71a7ff88a9f3719082e9398c)
Signed-off-by: Kevin Hao <kexin.hao@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Build and boot test for all the boards.
(From meta-yocto rev: b296bf6b0cadcff7c9000a3f4a3ca162785b9a33)
Signed-off-by: Kevin Hao <kexin.hao@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This marks the layers as compatible with honister now they use the new override
syntax.
(From meta-yocto rev: ea637ebd5a79e5d023bbf347d633568bcb7506a2)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This converts the two layers to the new override syntax. This was done using
<oe-core>/scripts/contrib/convert-overrides.py <this-layer>
with no manual fixes.
(From meta-yocto rev: 387eab70f343966267c2a68746aea1bbb8525191)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The beaglebone-yocto machine currently hardcodes "u-boot" as a
dependency for the image and the wic format, and this prevents one
from choosing a different bootloader via:
PREFERRED_PROVIDER_virtual/bootloader = "non-u-boot-bootloader"
Depending on "virtual/bootloader" instead of "u-boot" fixes this.
Cc: Denys Dmytriyenko <denis@denix.org>
(From meta-yocto rev: 4c4c8327645457a5d68762ff64b64d4b5845cd20)
Signed-off-by: Vivien Didelot <vdidelot@pbsc.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Build and boot for all these machines and also explicitly set the
preferred kernel version for the x86 machines so that we can override
the default setting in oe-core.
(From meta-yocto rev: a93344439e01e8aba4ba15d3123ce31b4500add1)
Signed-off-by: Kevin Hao <kexin.hao@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This reverts commit da27ca50eebf80463cb8d7b85f3b705254823413.
We usually select the preferred kernel version in the machine conf. This
is also what recommend in the BSP Developer's Guide. The beaglebone-yocto
and edgerouter have already done this in their machine conf files.
Resetting it in the layer conf is redundant. So revert this patch first,
we will also explicitly select the kernel version for the x86 machines
in the following patch.
(From meta-yocto rev: 825645ff2b663ae5c5845e8ae4679f6e346eac8c)
Signed-off-by: Kevin Hao <kexin.hao@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We already have the following variables defined above IMAGE_BOOT_FILES:
SPL_BINARY = MLO
KERNEL_IMAGETYPE = zImage
KERNEL_DEVICETREE = am335x-bone.dtb am335x-boneblack.dtb am335x-bonegreen.dtb
So use them instead of repeating their values.
(From meta-yocto rev: be07cc4f4e729905065ac6fbf1d46ebfdd79cf8f)
Signed-off-by: Vivien Didelot <vdidelot@pbsc.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When we are bumping the default kernel version for the qemu*
reference boards, but before the h/w reference boards have been
updated, we need to explicitly set the default kernel version
for those boards. Otherwise, we'll end up with warnings about
preferred versions of kernel modules not being available
(From meta-yocto rev: da27ca50eebf80463cb8d7b85f3b705254823413)
Signed-off-by: Bruce Ashfield <bruce.ashfield@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Boot test for all the supported boards. There is a kernel config
check warning for beaglebone-yocto, the fix [1] has been sent to
kernel-meta for merging.
[1] https://lists.yoctoproject.org/g/linux-yocto/message/8992
(From meta-yocto rev: ccc825335f257ca2259440c561eea02538edd762)
Signed-off-by: Kevin Hao <kexin.hao@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Build and boot test for the beaglebone and edgerouter.
(From meta-yocto rev: db4247b8c95ec997fa90226a49df36d7366c469f)
Signed-off-by: Kevin Hao <kexin.hao@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This hardware is old/obsolete and unobtainable. Its proving hard to support
with nobody fixing bugs or helping keep the platform running/up to date.
Whilst there is value in real hardware testing, this platform ist just too
old and obsolete to support. This does leave a gap for the power architecture
but at this point there is nobody willing to step up to cover it.
The TSC did discuss and agree support for this platform should be removed.
It calls into question the support/testing of the architecture by Yocto
Project which is being discussed by the TSC and governing board.
(From meta-yocto rev: aa691d49ceb6700eb1881b789fe34f8369bcb0ef)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
replace .wks with .wks.in and refer loader to EFI_PROVIDER
to enable different bootloader.
soft assign default EFI_PROVIDER "grub-efi" for genericx86.
(From meta-yocto rev: 11ecd2febed95e923c8cc68e2c438cd846a88d02)
Signed-off-by: Chee Yang Lee <chee.yang.lee@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
