The 'resize2fs' command we use to resize the filesystem on first boot
has been moved from standard 'e2fsprogs' package to its own
'e2fsprogs-resize2fs', so this is the one we need to include now.
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
Also remove kernel configuration fragments for kernel 3.10. This is not
needed for 3.10 as all the configuration is done via device-trees.
https://jira.digi.com/browse/DEL-1695
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
(cherry picked from commit b0031915f1c7f3eeba680161827e78bbac9eebb4)
This is a regression introduced in a74363c. The MOD_VARIANT variable was
not correctly set because in kernel 2.6.35 there is no device-tree
support.
The fix is to get the module variant from the sysfs as a fallback for
old kernels.
https://jira.digi.com/browse/DEL-1694
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
(cherry picked from commit 6357d71152c38c4713fc534b78b6c2993b201aa8)
The following section configures the Audio Data Path to Voice Over UART.
//Audio - Audio main config
[H:S]0041
[H:S]0004
[H:A]01 00 00 00
The syntax is explained in the AR3002 2.2 Tag definition 1.0 document from
Qualcomm/Atheros.
Other audio sections are removed as they are not being used.
Signed-off-by: Alex Gonzalez <alex.gonzalez@digi.com>
Adapt u-boot bootscript after latest changes in 'mmcdev' variable
semantics. Now the boot media device is stored in 'mmcbootdev' variable.
Also set the device tree filename depending on the hardware variant.
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
Support new memory chips for 'ccardimx28js'.
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
(cherry picked from commit b7d108c35f6a26a78a4fbab399fd2c0878976a67)
Some CC6 variants have wireless but not bluetooth. Those variants need a
different calibration file.
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
(cherry picked from commit a74363cf4a9b239edf9d628f84fe7e88002c0f5c)
Add post-installation script that runs on first boot to reconfigure the
u-boot environment depending on the media (EMMC,SD) the target is
booting from.
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
(cherry picked from commit 8486d120355d80ea8ad5864596c8486a1d3c0db2)
This is needed in the context of CC6 variants 0x7 and 0x9 (EMMC-less).
Changes:
* Do not use 'meta-fsl-arm' image generation class: removed include and
override SDCARD generation function
* Use same VFAT boot image for EMMC and SD card. The u-boot bootscript
has been adapted to be able to boot Linux from both SD and EMMC.
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
(cherry picked from commit 8b8456b8fd90facd92dfc87632effa582ca60475)
This will allow in following commit to use the same VFAT boot image for
EMMC and SD. Until now we were creating different boot images for EMMC
and SD card images.
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
(cherry picked from commit 13e9ec7608d42430c2a01a728d030752a71f7948)
Use Yocto's UBOOT_CONFIG new functionality to select the u-boot
configuration to build.
https://jira.digi.com/browse/DEL-1584
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
Remove the current DEY variants support and framework for CC6 because in
following commits we will add support for all *hardware* variants in
just one set of DEY images for CC6.
https://jira.digi.com/browse/DEL-1569
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
The previous class 1 settings for the atheros bluetooth resulted in
output power higher than what is allowable in all regulatory
domains where class 1 devices are allowed on the CC6. The updated class 1
regulatory settings here reduce the maximum output power to levels
which pass when tested at an external lab.
This change resolves CCIMX6-389.
Signed-off-by: Paul Osborne <Paul.Osborne@digi.com>
The IMAGE_FSTYPES variable for ccimx51js and ccimx53js was not correct,
it was taking the default one from Poky (IMAGE_FSTYPES="tar.gz")
instead of the one in 'imx-digi-base.inc' file. This happened because we
were using the weak '??=' assignement while the default in Poky was
using '?='.
Also rework a bit how this variable is set, so it is easily overriden on
project config files or initramfs recipes.
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
Those packages are not machine specific, so there is no need to set
PACKAGE_ARCH to ${MACHINE_ARCH}.
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
athtestcmd and wmiconfig are proprietary tools from Atheros. When
installed from GitHub we use prebuilt binaries because we cannot provide
the source code to the customers.
The prebuilt binaries need to be platform dependent, because the
'ccardimx28' uses soft-float binaries while the 'ccimx6' uses hard-float
binaries. Otherwise there is a failure:
root@ccimx6sbc:~# /usr/sbin/athtestcmd
-sh: /usr/sbin/athtestcmd: not found
While on it, removed the checksums for the binaries, as Yocto does
not verify checksums for local files by default, only for downloaded
files.
https://jira.digi.com/browse/DEL-1586
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
(cherry picked from commit 8a3a124c75a885a619b41bf1e17118f1b3dec13c)
The SD card images support added a dependence to mkimage native tool.
Otherwise the build may fail with:
mkimage: not found
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
Also remove support for device tree variants as it's not needed. We will
only have different device tree for the different CPUs, but not for the
different variants of the same CPU.
https://jira.digi.com/browse/DEL-1569
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
With yocto 1.8 we need to pass CC and LD from yocto environment to
imx-bootlets makefile. Otherwise it fails:
arm-dey-linux-gnueabi-ld: cannot find -lgcc
make[1]: *** [output-target/linux_prep] Error 1
https://jira.digi.com/browse/DEL-1558
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
Yocto 1.8 is more strict in the syntax when you append more flags to the
CFLAGS variable, because with previous yocto versions these packages
used to build fine and now fail with:
run.sstate_create_package: Syntax error: "(" unexpected
Use a different EXTRA_CFLAGS variable for our custom flags instead of
the standard CFLAGS.
https://jira.digi.com/browse/DEL-1558
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
We need to append our config to EXTRA_OEMAKE variable instead of
overriding it, because otherwise we lose some config from the
environment (e.g. KERNEL_SRC) and the build fails with:
make: *** M=...kernel-module-.../git: No such file or directory
Also remove redundant PACKAGE_ARCH setting, as it's inherited from
module-base class.
https://jira.digi.com/browse/DEL-1558
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
The way the kernel is built in yocto 1.8 has changed. Now the
build is done under ${WORKDIR}/build while the source code is at
STAGING_KERNEL_DIR, so we need to modify the '.config' in the build
directory instead of in the source directory.
This fixes:
kernel-source is not clean, please run 'make mrproper'
make[2]: *** [prepare3] Error 1
make[1]: *** [sub-make] Error 2
https://jira.digi.com/browse/DEL-1558
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
This synchronizes the machine configuration files with the latest
changes in meta-fsl-arm for yocto 1.8
https://jira.digi.com/browse/DEL-1558
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
Variant 0x8 does not have wireless/bluetooth, but it uses the same
device tree than the other variants.
https://jira.digi.com/browse/DEL-1524
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
(cherry picked from commit 3ba5bff51ebbe6be8325128ae3beb950993607d2)
Variant 0x5 (Furuno) does not have bluetooth, but the device tree in the
kernel is the same than the variant with bluetooth.
https://jira.digi.com/browse/DEL-1524
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
(cherry picked from commit 7de11f08e95fecb8d4895556292a488a91af6817)
It has never been a mirror of upstream packages.
It's the URL of packages needed for the BSP. These packages are mostly
Digi specific and not provided in any other location.
https://jira.digi.com/browse/DEL-1493
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
(cherry picked from commit d4cf9a88c9db39c6ecfd0099437d82af81c96e85)
This completes the support for ccimx6sbc variants:
* support device tree variants
* split variants '0x01,0x02,0x04' group in two different groups (with or
without kinetis)
https://jira.digi.com/browse/DEL-1359
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
(cherry picked from commit 3f20015411f4e270b2e2c8faf292678c75ff2aff)
https://jira.digi.com/browse/DEL-1359
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
(cherry picked from commit 015d029255eb407cca4c986562a304aef7de3940)
Kernel 3.10 does not fail to build perf features (as kernel 2.6.35
does), so there is no need to disable them for platforms using that
kernel version.
https://jira.digi.com/browse/DEL-1418
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
Revert to the old way of setting the console for ccardimx28js. Configure
the console on the generic ttyS0 port name and let the device managers
create a symbolic link to the correct device at runtime.
Trying to set it statically at compile time lead to some side effects
with some packages not being recompiled after a change of the preferred
kernel version in the project's local.conf (e.g. udev-extraconf and
sysvinit-inittab).
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
A post-down script was unloading the Atheros wireless driver
when the network interfaces were brought down.
On rare occasions the 'rmmod' command was not able to complete
and may leave the system in an unexpected state.
This patch removes the post-down script for the ccardimx28 platform
and leaves it only for the ccimx6 platform, where the unloading
is wanted to reset the driver after suspend.
Signed-off-by: Hector Palacios <hector.palacios@digi.com>
https://jira.digi.com/browse/DEL-1314
SERIAL_CONSOLES_CHECK is new functionality provided by Yocto, but it
needs '/proc/consoles' supported by the kernel. Our kernel 2.6.35 does
not support '/proc/consoles' so rework the serial console configuration
to work with both kernels (2.6.35 and 3.10).
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
Notes:
* This version of u-boot does not support comments in scripts
* Enabled EXT4 in kernel config (for the SDCARD rootfs)
https://jira.digi.com/browse/DEL-197
Signed-off-by: Javier Viguera <javier.viguera@digi.com>