Cleanup the J* APM patch a little. Apparently I could have stood to read
the docs a little more. Found some better runtime pm functions and also
discovered something new about the power island definition.
Notes from upstream:
- Add the CCB coherency HACK from DDK 1.15 to improve stability
on J* devices
- Fix power island support on J* devices
- Use better APM calls so user modification to power domains
don't cause the device to enter a bad state
Signed-off-by: Randolph Sapp <rs@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
Update the ti-rpmsg-char library to 0.6.0 version that enables the
support for 6.1 kernel.
Signed-off-by: Praneeth Bajjuri <praneeth@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
Set the variables needed to put the DTBs into the same vendored
directory structure as they are in the kernel.
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
Change over to using the kernel_devicetree_vendor_dtb.bbclass to have
the DTBs installed into the same vendored subdirectory that is in the
kernel. Setting the variables comes in a follow on patch.
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
There is a series of patches [1] [2] [3] that were submitted to oe-core
master that have gone through some iterations. It has partially stalled
in the upstreaming process but we need the effective changes to continue
our LTS plans for u-boot 2023.
This backports the effective changes of that patch against the kirkstone
version of the two classes, but renames them to make it obvious that we
are overriding them. This should be a temporary change as we expect the
changes to be backported to oe-core kirkstone once the stall is gone.
This patch will be removed once it makes it into oe-core kirkstone.
[1] https://lists.openembedded.org/g/openembedded-core/message/180753
[2] https://lists.openembedded.org/g/openembedded-core/message/181190
[2] https://lists.openembedded.org/g/openembedded-core/message/181191
Signed-off-by: Randolph Sapp <rs@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
Rename the .wks to .wks.in so the EFI_PROVIDER variable is expanded
properly. We only ever set a soft recommendation for GRUB. If anyone
attempted to change that previously it was ignored (for more than just
the reason here, but that's out of my control).
Add a timeout for the bootloader menu. Currently setting it to 3
seconds.
Remove the read only flag on the rootfs. Not sure why that was there.
Signed-off-by: Randolph Sapp <rs@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
Now that are past the inital stages of our LTS migration, we can remove
the DEFAULT_PREFERENCE settings that prevented these recipes from being
the default.
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
As part of the initial LTS migration, move to the latest commits for the
various repos.
- ti-linux-fw
- u-boot-ti-staging_2021.01
- u-boot-ti-staging_2023.04
- k3conf_git
- linux-ti-staging-rt_5.10
- linux-ti-staging-rt_6.1
- linux-ti-staging_5.10
- linux-ti-staging_6.1
Going forward all updates will come from the CICD flow.
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
Bump the src revision to enable the following features on Jacinto:
- Active power managment
- Power island control via firmware
Signed-off-by: Randolph Sapp <rs@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
Stop looking for the rpmsg_rpc.h in the kernel. While we are doing the
LTS migration for kirkstone, this file is not going to be in the kernel
for a bit. This patch works in conjunction with a recipe change that
downloads the header and adds a -I to the CFLAGS to pick it up. This
will be removed once the 6.1 kernel has been updated.
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
uboot will now use a merged defconfig build so keep the defconfig same
for hs here
Signed-off-by: Manorit Chawdhry <m-chawdhry@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
u-boot 2023.04 merged def configs for HS and GP EVM into one.
This patch updates config for J721S2 HS EVM for both A72 and R5.
Signed-off-by: Udit Kumar <u-kumar1@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
u-boot 2023.04 merged def configs for HS and GP EVM into one.
This patch updates config for J7200 HS EVM for both A72 and R5.
Signed-off-by: Udit Kumar <u-kumar1@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
When creating the boot partition we copy all of the images for the
different board types (gp, hs-fs, hs-se), but one of those we consider
to be a default. That default is not copied with the longer name, but
rather with the tiboot3.bin name.
The issue comes when someone wants to boot with a different bin file.
They need to copy the approriate longer named .bin file to tiboot3.bin.
But then the original name of what tiboot3.bin is lost unless the use
backs it up.
This patch just additionally copies that longer named bin file so that
anyone who wants to look at the boot partiion and decide to boot for a
different board can know which file to copy to tiboot3.bin and does not
need to worry about making a backup.
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
Avoid inheriting ti-secdev class and adding unconditional dependency
on TI_SECURE_DEV_PKG and other variables, when meta-ti-bsp is in the
bblayers.conf stack, but not building for TI platforms. This solves
yocto-check-layer signature test for Yocto Project compliance.
Signed-off-by: Denys Dmytriyenko <denys@konsulko.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
Avoid inheriting ti-secdev class and adding unconditional dependency
on TI_SECURE_DEV_PKG and other variables, when meta-ti-bsp is in the
bblayers.conf stack, but not building for TI platforms. This solves
yocto-check-layer signature test for Yocto Project compliance.
Signed-off-by: Denys Dmytriyenko <denys@konsulko.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
There are some unobvious issues with adding PVR support to Mesa
via a bbappend:
1. We need to mark mesa package as machine-specific, due to
differences in builds between SGX, Rogue and software-rendering
2. We also need to then mark mesa package as providing safe
ABIs (EGL/GLES/GBM) in order for all generic dependent packages
to not be treated as machine-specific, allowing their re-use
across different machines of the same architecture
But doing the above alters the upstream mesa package and changes
its signatures even when not building for TI platforms, which is
a Yocto Project compliance violation.
In order to resolve this issue, convert Mesa bbappend, that adds
PVR support, into its own standalone alternative provider, called
mesa-pvr and allow selecting it with PREFERRED_PROVIDER settings.
Signed-off-by: Denys Dmytriyenko <denys@konsulko.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
Some of the settings for KERNEL_DEVICETREE_PREFIX were missed in the
first pass. Additionally, some new dts files that were not present in
upstream are now there and can be updated in KERNEL_DEVICETREE.
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
The DT name k3-am62-lp-sk.dts was adopted instead of k3-am62x-lp-sk.dts
because the community prefers avoiding the wildcard letter 'x'.[1]
Drop the letter 'x' from the strings in KERNEL_DEVICETREE_PREFIX and
KERNEL_DEVICETREE to accurately reflect the DT names.
Signed-off-by: Aradhya Bhatia <a-bhatia1@ti.com>
[1]:
https://lore.kernel.org/all/06cbcd7d-bc83-bfeb-0821-72c7caf9a5e7@linaro.org/
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
Add and enable Jailhouse support for TI's k3-am625-sk similarly to how other
TI platforms were added and enabled in the past by adding platform-specific
configuration variables for the root cell DTB overlay, and the Linux demo inmate
DTB, configuration, and Kernel command line settings.
Migrate the jailhouse recipe to kirkstone from dunfell.
Switch to latest upstream source tree for jailhouse. The TI Jailhouse repo has
gone stale and now upstream is really the happening place where new development
happens including for new devices such as AM62x. Hence go ahead and switch to
using the upstream repo.
Update the platform-specific config files to such that the root cell DTB overlay
gets included in the file system images.
Enable additional console on ttyS3 (usually mapped to main_uart1) and ttyS2 for
the Linux demo inmate to use.
Add patches for the Jailhouse source tree for am62xx.
Signed-off-by: Andreas Dannenberg <dannenberg@ti.com>
Signed-off-by: Ricardo Salveti <ricardo@foundries.io>
Signed-off-by: Paresh Bhagat <p-bhagat@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
Only include the ti-img-rogue-umlibs-vulkan package if the
DISTRO_FEATURES x11 is enabled. Right now, this recipe has a failed
build dependency on images without x11.
Signed-off-by: Daniel Schultz <d.schultz@phytec.de>
Tested-by: Denys Dmytriyenko <denys@konsulko.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
Select the correct WKS file according to the MACHINE_FEATURE efi.
Right now, the sdimage-2part-efi.wks is set by default. Switch to the
sdimage-2part.wks file, if efi was removed from MACHINE_FEATURES to allow
non-efi boot methods.
Signed-off-by: Daniel Schultz <d.schultz@phytec.de>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
* Add support for additional config fragments passed via semi-standard
KERNEL_CONFIG_FRAGMENTS variable
* Also don't build all DTBs, only those specified in the machine config
Signed-off-by: Denys Dmytriyenko <denys@konsulko.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
Split the firmware into it's own package so we can make the insane
checks a little more paranoid again.
Signed-off-by: Randolph Sapp <rs@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
Cairo can either be built for use with OpenGL or OpenGLES. By default it
assumes it can use OpenGL if X11 is provided as a distro feature. This
assumption is not true when using the proprietary SGX / Rogue drivers.
Vulkan libraries need some X libs at runtime now, so we need to manually
override the PACKAGECONFIG for cairo to use OpenGLES.
Signed-off-by: Randolph Sapp <rs@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
Add the required vulkan configuration options and the new PVR WSI file to
the corresponding package.
Also bump the PV to match the version of Mesa actually being shipped.
Signed-off-by: Randolph Sapp <rs@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
Update the source revision for the following upstream changes:
- Drop the included ICD loader. Use the Mesa/Khronos one
instead.
- Enable all vulkan window system extensions
Separate the vulkan libs into their own package so people can choose to
ignore them if they want, considering the new runtime deps.
Signed-off-by: Randolph Sapp <rs@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
Currently we move this firmware around and put it in a directories named
pdk-ipc/ and ethfw/. This does not match what is in linux-firmware which
is what projects should expect /lib/firmware to match. This causes issues
for projects that want to work both on meta-ti and also on any other distro.
Switch to shipping the same directory structure as linux-firmware here.
Signed-off-by: Andrew Davis <afd@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
The version information should come from the ti-linux-fw.inc and this
version is old anyway.
Signed-off-by: Andrew Davis <afd@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
For TI platforms with SGX and Rogue graphics we patch upstream Mesa with
PVR support and conditionally enable features based on virtual/gpudriver
preference. This makes Mesa package platform-specific and it contains the
corresponding PACKAGE_ARCH = "${MACHINE_ARCH}" line.
In order to get greater re-use from generic non-machine-speciofic packages
down the dependency chain, such as Wayland, Weston, Qt5 modules, etc. we
need to mark mesa recipe as still providing stable OpenGL ABIs, such as
EGL, GLES, GBM, etc. - add it to SIGGEN_EXCLUDERECIPES_ABISAFE list.
Signed-off-by: Denys Dmytriyenko <denys@konsulko.com>
Reviewed-by: Andrew Davis <afd@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
Neither of recipes nor their ABI is all that stable. OpenGL might be
slightly more stable, but that is not what these provide anymore.
The variable helps with re-use of generic packages down the dependency
tree, when they depend on a machine-specific package. SGX UM used to
provide different standard OpenGL ABIs (EGL, GLES, GBM) and hence was
listed here. Now those ABIs are provided by Mesa, so SGX is removed.
Signed-off-by: Andrew Davis <afd@ti.com>
Signed-off-by: Denys Dmytriyenko <denys@konsulko.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
Use these to select build type and window system support to match how
the Rouge driver does the same. No functional change.
Signed-off-by: Andrew Davis <afd@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
Of the K3 family only the AM65x device is currently supported by this
driver. COMPATIBLE_MACHINE should match.
Signed-off-by: Andrew Davis <afd@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
The "clean" target has been fixed along with some other issues that
caused us to need these set. Remove them now.
Signed-off-by: Andrew Davis <afd@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
Small changes, the "upstream" doesn't look to be maintained anymore,
so if we get any more breaks it might be time to remove this recipe.
Signed-off-by: Andrew Davis <afd@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
The traditional order of serial consoles for K3 has ttyS2 being the
primary UART attached console. ttyS0 is also a common choice for
distros and the TI kernel may switch at some point. To prepare
we check both UARTs on all K3 devices. Move this common check
to a K3 common location.
Signed-off-by: Andrew Davis <afd@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
Update the rogue graphics recipes to point at the new 23.1.6404501
release. This adds support for:
- kernel 6.1
- vulkan
- opencl (more testing needed)
- zink (more testing needed)
Signed-off-by: Randolph Sapp <rs@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
Update the pvr-mesa include to use the 22.3.5 patches from
oe-core/master and the branch for DDK version 23.1 on freedesktop.
Also remove the PACKAGECONFIG options elf-tls and xvmc as those
functions were dropped by the oe-core/master version of this package and
mesa package respectively.
Signed-off-by: Randolph Sapp <rs@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
Locally overlay verbatim copies of mesa 22.3.5 patches from
oe-core/master, so the same codebase works in kirkstone and master.
Also drop the old patches from 22.0 that were from the tip of
oe-core/kirkstone previously.
Signed-off-by: Randolph Sapp <rs@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
Was used to build kernel DTBs prior to KERNEL_DEVICETREE getting populated
with the list of preset of upstream DTBs. Not needed anymore so remove.
Signed-off-by: Andrew Davis <afd@ti.com>
Reviewed-by: Denys Dmytriyenko <denys@konsulko.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
Randolph Sapp
Praneeth Bajjuri
Ryan Eatmon
Jai Luthra
LCPD Automation Script
Manorit Chawdhry
Udit Kumar
Denys Dmytriyenko
Aradhya Bhatia
Paresh Bhagat
Daniel Schultz
Andrew Davis