The result of the modprobe operation doesn't guarantee the driver loaded
correctly. The module may succeed and be loaded, but failures during the
probe may not propagate and result in an error of modprobe.
The valid condition is to check the existence of the interface in
/sys/class/net/wlan0.
Signed-off-by: Hector Palacios <hector.palacios@digi.com>
https://onedigi.atlassian.net/browse/DEL-8821
Commit d3f3cfdb62 removed the inclusion of STM's
linux-stm32mp.inc from meta-st-stm32mp in our linux-dey recipe, but this
inadvertently removed the logic in do_configure() necessary to use our custom
ccmp1_defconfig. Since this commit, the kernel was being built with the default
ARM defconfig, which is very different from our custom defconfig and doesn't
even boot on MP1 platforms.
Rework the logic used to copy our platform's defconfigs to prevent this.
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
An anonymous function in linux-stm32mp.inc produces a bbfatal error when
KERNEL_DEVICETREE variable contains more than one device tree. This is our
case since we build the main DT plus a number of DT overlays.
This commit removes the dependency to this include file since we have our
own recipe to build the kernel and it is not needed at all.
It also removes the build of a uImage and the need to provide a
LOADADDR.
Signed-off-by: Hector Palacios <hector.palacios@digi.com>
This commit adds signed FIT image support for the CCMP1
platforms when using Trustfence.
https://onedigi.atlassian.net/browse/DEL-8591
Signed-off-by: Mike Engel <Mike.Engel@digi.com>
Some platforms do not support signing external artifacts (kernel, dtb,
etc.) yet, so we need to decouple the signing of the bootloader from the
signing of the external artifacts.
This commit generalizes the code, so instead of having platform exceptions
scattered along the recipes, we create a new variable used conditionally
to sign or not the external artifacts.
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
This branch contains the latest BSP changes from STM's v5.15-stm32mp-r2.1
release.
https://onedigi.atlassian.net/browse/DEL-8659
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
The btnxpuart driver is used for the bluetooth chip. We want to control
when to load and unload it, and when power/unpower the chip.
Therefore, blacklist it, so we can manage it in our scripts.
https://onedigi.atlassian.net/browse/DEL-8632
Signed-off-by: Isaac Hermida <isaac.hermida@digi.com>
The only user of 'linux-dey-src.inc' was the linux recipe itself, so
instead rename that file to a more generic 'linux-dey.inc' and include
more common code in that renamed file.
This is in preparation for the new linux 6.1.1 recipe for the ccimx93.
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
We still need to use this downgraded version of dtc to be able to build
U-Boot v2017.03, so backport this patch to be able to build dtc with the latest
version of gcc
https://onedigi.atlassian.net/browse/DEL-8540
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
For the moment, do not sign aditional artifacts, such as the ramdisk,
the kernel or the boot scripts for STM platforms.
In the specific case of the ramdisk, simply copy it over with the
expected filename extension.
Signed-off-by: Hector Palacios <hector.palacios@digi.com>
Certain platforms share a processor family but need to be differentiated
between them. DEY was using the variable DIGI_FAMILY as the SOM name
rather than the family. It becomes useful to have both (DIGI_SOM as the
more specific, and DIGI_FAMILY as the more generic).
This is the case, for example, of:
- ccmp1 (family)
- ccmp15 (SOM)
- ccmp13 (SOM)
- ccimx8m (family)
- ccimx8mm (SOM)
- ccimx8mn (SOM)
Both variables are used on the machine overrides.
Where DIGI_FAMILY was used, use now DIGI_SOM.
Signed-off-by: Hector Palacios <hector.palacios@digi.com>
The name of the variable was not very intuitive of what
it contains. This variable expands to the SoC vendor
(NXP or STM).
Signed-off-by: Hector Palacios <hector.palacios@digi.com>
rename interface sta_name to be "wlan0" instead of "mlan0", so it keeps
compatibility with other platforms.
Signed-off-by: Isaac Hermida <isaac.hermida@digi.com>
1)
CcxTagControlData struct in CORE/HDD/inc/CcxTagDefs.h had some fields missing compared
with 'cctags api’ which was producing CCX_TAG_DRVR_RSP_ERR_CTRL_SIZE.
2)
Then CCX_TAG_DRVR_RSP_ERR_PACKET_SIZE aroused because following paddings:
CCX_TAG_MPDU_HEADER_PAD_SIZE = 2
CCX_TAG_CONTENT_HEADER_PAD_SIZE = 3
where producing wrong alignements and sizes between the 'cctags api’ and the driver.
With changes in this commit, master branch of ‘CCxTagTest’ repository compiled
out of source code gets on well with 'cctags api’ repository and with qca driver in
both dey-2.4 and dey4.0.
Packets in the air match 100% in content in both dey-2.4 and dey4.0.
https://onedigi.atlassian.net/browse/DEL-8446
Signed-off-by: Hector Bujanda <hector.bujanda@digi.com>
1)
CcxTagControlData struct in CORE/HDD/inc/CcxTagDefs.h had some fields missing compared
with 'cctags api’ which was producing CCX_TAG_DRVR_RSP_ERR_CTRL_SIZE.
2)
Then CCX_TAG_DRVR_RSP_ERR_PACKET_SIZE aroused because following paddings:
CCX_TAG_MPDU_HEADER_PAD_SIZE = 2
CCX_TAG_CONTENT_HEADER_PAD_SIZE = 3
where producing wrong alignements and sizes between the 'cctags api’ and the driver.
With changes in this commit, master branch of ‘CCxTagTest’ repository compiled
out of source code gets on well with 'cctags api’ repository and with qca driver in
both dey-2.4 and dey4.0.
Packets in the air match 100% in content in both dey-2.4 and dey4.0.
https://onedigi.atlassian.net/browse/DEL-8446
Signed-off-by: Hector Bujanda <hector.bujanda@digi.com>
add a verbatim copy of those recipes as they are provided by nxp in langdale-6.1.1-1.0.0
https://onedigi.atlassian.net/browse/DEL-8346
Signed-off-by: Isaac Hermida <isaac.hermida@digi.com>
In rare cases, the wlan module may fail to load. In this scenario, forcing the
MMC driver to rebind the interface resolves the issue, making the WLAN module
load more reliably.
This only affects to modules where the qca chip is connected through the
MMC interface.
https://onedigi.atlassian.net/browse/DEL-8361
Signed-off-by: Isaac Hermida <isaac.hermida@digi.com>
Symbols are needed for DTB overlays. In the CC6UL we are not using
overlays, so disable symbols generation.
https://onedigi.atlassian.net/browse/DEL-8397
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
The SW encryption is much faster than decicated HW in our platforms.
The HW encryption is limited to the use of CAAM which is used through
blob calls, therefore the cryptodev module is not required.
https://onedigi.atlassian.net/browse/DEL-8371
Signed-off-by: Isaac Hermida <isaac.hermida@digi.com>
Also enable recipe for ccimx93, and pass the correct DTC flags to create
overlays capable DTBs.
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
Co-authored-by: Javier Viguera <javier.viguera@digi.com>
Signed-off-by: Isaac Hermida <isaac.hermida@digi.com>
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
The qualcomm.sh script has been losing functionality over time, to the
point that now is just loading the kernel module using 'modprobe'.
Instead, run modprobe directly from the udev rule and get rid of the
qualcomm script.
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
Verbatim copy of the recipes in DEY-3.2, using command:
git restore -s dey-3.2/master -- meta-digi-arm/recipes-kernel/kernel-module-qualcomm
Signed-off-by: Isaac Hermida <isaac.hermida@digi.com>
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
For development this would not be needed as it points to AUTOREV, but
for releases we need to specify the SHA1 revision, and this recipe
builds two different branches (for NXP and STM platforms), so we need a
place to define two different revisions.
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
The qualcomm.sh script has been losing functionality over time, to the
point that now is just loading the kernel module using 'modprobe'.
Instead, run modprobe directly from the udev rule and get rid of the
qualcomm script.
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
The SRCREV may change depending on the version of the kernel, so it
cannot be a common variable for all kernel versions.
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
Use the canonical Stash repository to avoid synchronization problems
between the repo and the mirror.
Moreover the LOG mirror is hosted in an old VM machine with limited
specs, and although the transfer of the objects is faster, the counting
of the objects to transfer (which is done in the server) takes ages to
complete, so at the end, there is no time gain using the mirror.
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
Commit 8cba3aaefb in 'meta-freescale' changed the kernel_localversion
function we are using in our kernel recipe. This leads to a not-properly
configured kernel build, where CONFIG_MODULES is disabled. That
config option is needed to build external kernel modules in other
recipes, so those recipes (kernel-module-qualcomm, cryptodev-module,
etc) fail to build.
The commit makes sure the kernel is properly configured for building by
extending the do_configure function in a "prepend".
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
It's just a waste of space as we already have the kernel image in the
'linux' partition and that gets mounted under /mnt/linux.
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
The enable p2p patch add support for enabling p2p as module parameter setting
it by default to false.
The cfg80211 netdevice fixes the registration of virtual interfaces for kernels
v5.12+.
https://onedigi.atlassian.net/browse/DEL-7986
Signed-off-by: Isaac Hermida <isaac.hermida@digi.com>
Also, fix the sequence number of the GPIO_SET_DEBOUNCE_IOCTL ioctl. Same
fix has been done in Linux.
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
Remove deprecated versions of recipes updated in other general layers
(poky, meta-openembedded). Also remove duplicated IMX specific recipes that
are available in other BSP layers (meta-freescale, meta-fsl-demos, etc).
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
Yocto is becoming more strict on not using tags in the SRCREV, as that
may lead to problems in the different tasks of a bitbake recipe.
As of Yocto 4.0, bitbake errors when such condition happens, with:
"Recipe uses a floating tag/branch without a fixed SRCREV"
So convert the SRCREV of all tags to the proper SHA1 revision.
Also, extend the 'fix-build-issues' to delete '-Werror' from the
compilation flags. Newer versions of GCC (as the one used in Yocto 4.0)
throw more (new) warnings, and the '-Werror' flag was making the build
to fail.
cc1: all warnings being treated as errors
lim_api.c:1057:17: error: this 'if' clause does not guard... [-Werror=misleading-indentation]
lim_admit_control.c:169:17: error: this 'if' clause does not guard... [-Werror=misleading-indentation]
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
This commit adds a patch for the qcacld-3.0 wireless driver to read the
wireless MAC addresses from device tree. This patch is based on the original
commit from qcacld-2.0 driver improving and simplifying the code to work with
the new qcacld-3.0 driver.
https://onedigi.atlassian.net/browse/DEL-7916
Signed-off-by: Arturo Buzarra <arturo.buzarra@digi.com>
This commit updates the Linux wireless driver for Qualcomm's chipset to the
new qcacld-3.0 driver.
This update is based on the release "r10016.1 - Post-CS3 1.0.016.1" from
"qca6574au-le-2-2-1_qca_oem" product branch, which uses for a Linux kernel v5.4
the tag "CHSS.LNX_FSL.5.0-01200-QCA6574AUARMSDIOHZ".
https://onedigi.atlassian.net/browse/DEL-7916
Signed-off-by: Arturo Buzarra <arturo.buzarra@digi.com>
This commit avoids creating a subfolder for the kernel artifacts, to homogenize
it with the NXP platforms and avoid using custom recipes for the result
artifacts.
Signed-off-by: Arturo Buzarra <arturo.buzarra@digi.com>
The sign mode needed for each platform is invariable, and since the platform
is already a mandatory parameter for the script, we can store this information
implicitly. Reflect this change in every recipe where the script is used, but
keep the variable at the Yocto level since it's still needed in several places.
https://onedigi.atlassian.net/browse/DEL-7862
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
The license checksum and the source branch should be on the properly
versioned bitbake recipes. Otherwise, when we include this common file
in a different version linux recipe (for example, the upcoming 5.10) it
will inherit incorrect values.
While on it, fix the incorrect license checksum values for the linux 5.4
recipes.
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
This makes it so bitbake obtains the SDMA firmware from the firmware-imx recipe
instead.
This also avoids including all of the linux-firmware packages in i.MX6 SDK
builds, preventing errors when building the ccimx6sbc SDK (linux-firmware
provides the community versions of the ath6k and ar3k firmwares, which clash
with our versions).
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
This package is used by the eIQ recipes, so make sure we have NXP's .bbappend
so it works as expected
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
Digi Embedded Yocto 3.2-r2.2
Manually changed recipes to use the master branches instead of the fixed SHA1
from the last release.
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
Create this symbolics links in build time instead of doing it
in run-time. This simplifies the system and allows it to work
with read only filesystems.
Signed-off-by: Francisco Gil <francisco.gilmartinez@digi.com>
Before this patch the MAC addresses were read from the device
tree in the qualcomm.sh script executed at init, and a creates
a file where the MAC where located.
In 0a92f5d9e8df644de74e74f152ed6ff3dd8d3369 in the qca driver
the support to read these MAC addresses from the filesystem
was added.
This commits creates the simbolic link between the device tree
and the expected path where the MAC addresses where located, and
gets rid of the file with the MAC addresses.
Signed-off-by: Francisco Gil <francisco.gilmartinez@digi.com>
This change was accidentally reverted when merging to dey-3.2/maint, causing
the qualcomm module to be missing in the rootfs.
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
This allows the packages to be included in the ccimx6sbc images. While at it,
include the Qualcomm bluez patches in ccimx6 builds. These patches aren't
destructive, they simply add functionality required by the Qualcomm chip, so
they shouldn't have any secondary effects when using the Atheros chip.
https://onedigi.atlassian.net/browse/DEL-7661https://onedigi.atlassian.net/browse/DEL-7666
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
- Add the '-m' parameter. Without it, the script doesn't
generate the output merged configuration file, and the
fragments are not applied.
- Fix the if condition. The '*.cfg' expands to a list of
files, but the '-f' only admits one parameter. Use instead
the Python find_cfgs() function.
Signed-off-by: Hector Palacios <hector.palacios@digi.com>
This allows users to apply small changes to our kernel configuration without
having to create a completely new defconfig. Use a simplified version of the
kernel-yocto.bbclass implementation.
https://onedigi.atlassian.net/browse/DEL-6706
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
Enabling DIGI_INTERNAL_GIT defaults to LOG server. The build from local
MTK Digi server was broken.
Fix uboot and linux recipes declaring different repo paths depending on
whether the local remote is LOG or MTK.
Signed-off-by: Hector Palacios <hector.palacios@digi.com>
This commit backports a patch to fix the following build issue with
the GCC v10:
dtc-parser.tab.c:1069: multiple definition of `yylloc'
Signed-off-by: Arturo Buzarra <arturo.buzarra@digi.com>
Commit ffeb0556("classes: fsl-kernel-localversion: Fix task dependency") in
meta-freescale layer fixes a task dependency issue, so do_preconfigure() was
renamed to do_kernel_localversion().
https://jira.digi.com/browse/DEL-7508
Signed-off-by: Arturo Buzarra <arturo.buzarra@digi.com>
Remove all v4.9 recipes and build Linux using the same branch on all platforms.
The .inc files were only needed because of the existence of multiple Linux
versions, but that's not the case anymore, so remove them.
https://jira.digi.com/browse/DEL-7221
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
Arturo Buzarra
Hector Palacios
Isaac Hermida
Gabriel Valcazar
Mike Engel
Javier Viguera
Francisco Gil
Hector Bujanda
Gonzalo Ruiz