Age | Commit message (Collapse) | Author |
|
Avoid PREFERRED_PROVIDER_u-boot and PREFERRED_PROVIDER_virtual/bootloader
drift apart as good as it gets without creating circular dependencies.
Read: there in one primary definition (PREFERRED_PROVIDER_u-boot) and the
second specification is relying on the primary.
Signed-off-by: Jens Rehsack <sno@netbsd.org>
|
|
u-boot-fw-utils have been replace by libubootenv which has been finalized with
commit d15639a0a27393aa455adae113f4692c27367e89
Author: Otavio Salvador <otavio@ossystems.com.br>
Date: Thu Apr 2 11:19:34 2020 -0300
u-boot-qoriq-fw-utils: Drop recipe as libubootenv works as replacement
Now we are using libubootenv and dropped the u-boot-qoriq-fw-utils, we
need to move the configuration file to the main recipe.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Remove choosing a preferred provider, too.
Signed-off-by: Jens Rehsack <sno@netbsd.org>
|
|
Instead of maintaing 98% copy of OE-core recipe without being up-to-date
as good as possible, inject patches into OE-core recipe of openssl when
compiling for the right platform.
Signed-off-by: Jens Rehsack <sno@netbsd.org>
|
|
This is an adaptation of the original recipe
https://bitbucket.org/retotech/meta-jailhouse
- add recipe for Jailhouse version 0.11
- enable for mx8mq, mx8mm, mx8mn
- add pyjailhouse/python tools like jailhouse-cell-linux
- update to python3 classes
Signed-off-by: Cristinel Panfir <cristinel.panfir@nxp.com>
|
|
|
|
Machines or distros can define which BSP it should use by default. We
are intending to default for NXP BSP by default and specific
machines or DISTROs might change it if need.
Two values are considered valid: mainline, nxp
For example:
QORIQ_DEFAULT_BSP = "nxp"
Signed-off-by: Ting Liu <ting.liu@nxp.com>
|
|
We should avoid leaking NXP settings when using mainline BSP so we need
to filter it out.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
The library should not be overrided for SoCs without official BSP
releases from NXP as it is untested.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
Current assignment of the IMX_DEFAULT_BSP variables is too strong. For
example, any IMX_DEFAULT_BSP value set in local.conf currently gets
overridden by the assingment in imx-base.inc. Soften the assignment
to allow for alternative values for these variables.
Signed-off-by: Carlos Rafael Giani <crg7475@mailbox.org>
|
|
Bump revisions of optee, wayland-protocols and weston since recipes were
upgraded.
Signed-off-by: Andrey Zhizhikin <andrey.zhizhikin@leica-geosystems.com>
|
|
The i.MX 8QXP has Cortex-A35 cores plus Crypto
extensions, so enable them by default.
Signed-off-by: Tom Hochstein <tom.hochstein@nxp.com>
|
|
We cannot change the default BSP based on SoC as we trim out those when
moving to the mainline BSP.
For this reason, we moved the i.MX8 forced setting to our respective
machines as they lack mainline BSP support.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
The imx-alsa-plugins use i.MX kernel specific APIs so we cannot use it
with mainline.
Now, the imx-alsa-plugins are added only when NXP BSP is in use.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
We need to postpone the expanion so we avoid cycles.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
Like other i.MX 8M SOCs, the 8M Quad has Cortex-A53 cores
and supports ARM Crypto extensions. Enable them by default.
Signed-off-by: Tom Hochstein <tom.hochstein@nxp.com>
|
|
We are forcing the mainline BSP for 'mxs' and 'mx5' so it will always
use '-imx' suffix. This allows us to drop the unused values.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
The mainline BSP is the only well supported option for i.MX23, i.MX28,
i.MX51 and i.MX53 so force its use.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
The `use-mainline-bsp` must have higher priority over SoC one so it can
override settings when need.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
Machines or distros can define which BSP it should use by default. We
are intending to default for mainline BSP by default and specific
machines or DISTROs might change it if need.
Two values are considered valid: mainline, nxp
For example:
IMX_DEFAULT_BSP ??= "mainline"
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
The community forks are not being actively maintained by community and
vendors and as consequence are not being updated as need. Aiming to
reduce the maintenance burden we are dropping them as we will move to
NXP's GA kernels when using NXP BSP.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
The providers were incorrectly set to the -native recipe instead of the
nativesdk- recipe
Signed-off-by: Joshua Watt <Joshua.Watt@garmin.com>
|
|
Splits out the IMX u-boot tools build to it's own recipe. This is
required because these are native tools and thus need to be in a -native
recipe to be properly relocated (instead of being built as part of the
target u-boot-imx recipe). If the tools aren't built in a -native
recipe, then the mkimage tool will contain a hard coded path to the
uninative program loader from the sandbox where it was first built and
this will not get corrected when restored from sstate. The result is
that the mkimage_uboot executable can't run because it's looking for a
hard coded program loader that does not exist, resulting in strange
build errors like:
./mkimage_uboot: Command not found
even though the executable is obviously in the correct location.
Signed-off-by: Joshua Watt <Joshua.Watt@garmin.com>
|
|
See similar commit fb753d427add595d9c425fab9d4720a041db1b6d:
imx-base: Enable wic.bmap image generation as default
Bmaptool is a generic tool for creating the block map (bmap) for a
file and copying files using the block map. Is faster than use dd
to flash images to SD Cards. More info about bmap here [1].
Signed-off-by: Nicola Lunghi <nick83ola@gmail.com>
|
|
Use the cortexa72-cortexa53-crypto tune for IMX.8 QuadMax devices
instead of the more generic aarch64. Update the QuadMax MEK machine to
require the file that provides this tune.
Signed-off-by: Joshua Watt <Joshua.Watt@garmin.com>
|
|
On mx8,
`WKS_FILE_DEPENDS_mx8 += "imx-boot"`
overrides the content of `WKS_FILE_DEPENDS` instead
of appending, causing `do_image_wic` to complain about
missing dependencies for `wic-tools`.
|
|
Add firmware package definitions for Broadcom wifi modules, which are
used on NXP EVKs (for i.MX8M series).
Signed-off-by: Andrey Zhizhikin <andrey.zhizhikin@leica-geosystems.com>
|
|
Add necessary machine defines and overrides required to support i.MX8M
Nano SoC and is similar to i.MX8M Mini SoC.
New SoC lacks VPU support, therefore it is excluded from
MACHINEOVERRIDES_EXTENDER, leaving only GPU-related features.
Signed-off-by: Andrey Zhizhikin <andrey.zhizhikin@leica-geosystems.com>
|
|
Signed-off-by: Andreas Müller <schnitzeltony@gmail.com>
|
|
Refresh recipe base from poky, drops patches no longer used.
Signed-off-by: Mihai Lindner <mihai.lindner@nxp.com>
|
|
Signed-off-by: Mihai Lindner <mihai.lindner@nxp.com>
|
|
Default u-boot load address is set to 0x40480000 but fitImage
support requires the UBOOT_ENTRYPOINT to be defined as well, so
just extend imx-base.inc to also include a default uboot entrypoint
for mx8mm.
Signed-off-by: Ricardo Salveti <ricardo@foundries.io>
|
|
Looks like the variable was accidentally re-introduced.
https://github.com/Freescale/meta-freescale/commit/2cbca9a255674649730b3f440a6142cf16030267
Signed-off-by: Tom Hochstein <tom.hochstein@nxp.com>
|
|
optee imx fork like most other imx forked components is
more fully featured for imx devices and supports things
like CAAM accleration.
Signed-off-by: Peter Griffin <peter.griffin@linaro.org>
|
|
BOOT_SPACE and SDCARD_ROOTFS are no longer used with the move to wic.
Signed-off-by: Tom Hochstein <tom.hochstein@nxp.com>
|
|
The UBOOT_ENTRYPOINT setting is used when generating fitImage data.
Specifically where to extract the kernel binary into memory.
When fitImage is enabled for i.MX 7ULP targets, we are incorrectly
setting the extraction location to 0x80008000 (the mx7 default).
This results in a hung boot.
The default configuration for i.MX 7ULP is to load the kernel binary
(zImage) and initramfs into memory directly from the vfat boot
partition so this issue does not present itself until fitImage is used.
To fix the extraction point for fitImage kernel binaries, let's add
a specific setting for mx7ulp so that mx7 setting isn't used.
Signed-off-by: Michael Scott <mike@foundries.io>
|
|
Signed-off-by: Peter Griffin <peter.griffin@linaro.org>
|
|
Signed-off-by: Tom Hochstein <tom.hochstein@nxp.com>
|
|
Current implementation of i.MX8M Mini is utilizing generic aarch64 tune
(arch-arm64) provided from OE meta layer.
According to IMX8MMRM [1] this CPU contains Cortex-A53 cores, so this commit
sets a default tune to cortexa53-crypto.
Note, that according to [1] ARM Crypto extensions are supported, therefore it
is beneficial to set the tune to -crypto, as this also enables NEON and VFP
support in GCC (via -march and -mcpu feature modifiers, see [2]).
[1]: https://www.nxp.com/webapp/Download?colCode=IMX8MMRM
[2]: https://gcc.gnu.org/onlinedocs/gcc/AArch64-Options.html#aarch64-feature-modifiers
Signed-off-by: Andrey Zhizhikin <andrey.z@gmail.com>
|
|
It make sense for the default config to use the same provider for
for both u-boot and u-boot-fw-utils. We now assume that there is a
matching -fw-utils package, if not the build will fail instead of
silently using a mismatching u-boot version.
Signed-off-by: Alban Bedel <alban.bedel@avionic-design.de>
|
|
This is needed for GStreamer 1.16 Wayland support
Signed-off-by: Carlos Rafael Giani <crg7475@mailbox.org>
|
|
Function make_dtb_boot_files now always returns and raises an error if a
KERNEL_DEVICETREE entry filename extension is not .dtb or .dtbo.
Signed-off-by: Jacopo Dall'Aglio <jacopo.dallaglio@kynetics.com>
|
|
Image container generation of imx-boot requires sha384sum also
when building for i.MX 8 targets (e.g. i.MX 8QM). Make sure it
is available for all i.MX 8 families.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
Split alldtbs by any number of whitespace instead of just one to fix machines
with dtbs listed on multiple lines of KERNEL_DEVICETREE variable.
Signed-off-by: Diego Rondini <diego.rondini@kynetics.com>
|
|
Move gstreamer1.0 to the generic i.MX override, all i.MX use the 1.14.imx
override, not only the once with a GPU.
Completes '0d5f4e27 imx-base: make gstreamer1.0 packages socarch specific
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
|
|
UBOOT_SUFFIX is already set in machine configuration files and
we can use this variable to set u-boot suffix for UBOOT_BINARY.
|
|
Some boards expects a specific binary name.
Example : WaRP7 required the u-boot-dtb.imx file (instead of u-boot.imx).
Signed-off-by: Pierre-Jean Texier <pjtexier@koncepto.io>
Signed-off-by: Joris Offouga <offougajoris@gmail.com>
|
|
The RDEPENDS from kernel-module-qca* is not working, so that firmware-qca*
and qca-tools are missing from the image. Use a packagegroup instead
as a workaround.
Signed-off-by: Tom Hochstein <tom.hochstein@nxp.com>
|
|
Signed-off-by: Tom Hochstein <tom.hochstein@nxp.com>
|
|
gstreamer and its plugins have been forked for i.MX. So all of the generated
packages should have their PACKAGE_ARCH set to SOCARCH.
Prevents, when building gstreamer for an i.MX machine followed by a non i.MX
machine:
| ERROR: gstreamer1.0-1.14.4-r0 do_packagedata_setscene: QA Issue: Package version for package gstreamer1.0 went backwards which would break package feeds from (0:1.14.imx-r0 to 0:1.14.4-r0) [version-going-backwards]
| and many more for each generated package
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
|