Age | Commit message (Collapse) | Author |
|
These machines use an old Nvidia provided downstream kernel 3.1.10. They no
longer much the way more recent sysroot components, e.g. libc, xserver et. al.
The maintance burden to keep the userspace components in their old version
becomes simply to high. Keep using the rocko based 2.8 BSP for these machines.
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
|
|
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
|
|
Replace a bunch of += assignments with = .
Put all RDEPENDS into one statement.
Cleanup indent to the prefered OE 4 spaces per level.
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
|
|
The update is mostly bugfixing, notably:
This fixes a build error with glibc 2.28.
<sys/sysmacros.h> is no more included by <sys/types.h>
additionally:
Add support for UAC2 function
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
|
|
Update to latest git for imx6, imx7 and tegra support and integrate
automatic board configuration support.
Signed-off-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
Acked-by: Max Krummenacher <max.krummenacher@toradex.com>
|
|
Libsoc is currently built with no directly enabled board config
(we will wait until all configs are available). All available
configs are nonetheless included in the rootfs and can be
symlinked to /etc/libsoc.conf to enable them. The library can be
otherwise used with the Linux GPIO numbers.
Signed-off-by: Gerard Salvatella <gerard.salvatella@toradex.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
|
|
Using "10#" to treat the product ID as a decimal value does not
work with the BusyBox shell. Use eval productid + 0 which works
well with bash and BusyBox to convert to a decimal value.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
|
|
Since the OS Descriptor feature has been reviewed and upstreamed
the USB Gadget library configuration format changed slightly.
Update the configuration files to make sure they work with the
current upstream library version.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
|
|
go now has its own compile class. Use it and make it work.
snapd uses both autoconf and go, but the autoconf and go class compete
for the configure/compile class, so manually add what is additonally needed
for go.
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
|
|
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
|
|
When compiling for thumb one gets errors like this:
| ../recipe-sysroot-native/usr/lib/arm-angstrom-linux-gnueabi/go/pkg/tool/linux_amd64/link \
| -o $WORK/github.com/snapcore/snapd/cmd/snap/_obj/exe/a.out -L $WORK \
| -L .../recipe-sysroot/usr/lib/arm-angstrom-linux-gnueabi/go/pkg/linux_arm \
| -extld=arm-angstrom-linux-gnueabi-gcc "-extldflags=-march=armv7-a -mthumb \
| -mfpu=neon -mfloat-abi=hard --sysroot=.../recipe-sysroot" -buildmode=exe \
| -buildid=e86c00f573bc1df95256a4a05d36066839bfed34 $WORK/github.com/snapcore/snapd/cmd/snap.a
| # github.com/snapcore/snapd/cmd/snap
| /tmp/go-build238845398/os/user.a(_all.o): unknown relocation type 30; compiled without -fpic?
| invalid rela size 1
| /tmp/go-build238845398/net.a(_all.o): unknown relocation type 30; compiled without -fpic?
| invalid rela size 1
| ...
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
|
|
As OE generally hides the build output from the console it does not
clutter the screen, however if something goes wrong the log file
contains important debug info.
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
|
|
meta-snappy now uses go from oe-core.
This commit backports the relevant changes to the recipe.
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
|
|
The OS descriptor changes have been merged upstream, hence we can
use upstream libusbgx again.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
|
|
In the Toradex Easy Installer we do not use the example programs
or configuration. Create a separate package for the library and
the example programs and configurations.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
|
|
So far a generic USB vendor and product ID has been used in the
Toradex specific schema files. However, the startup service wrote a
Toradex specific USB product ID. This lead to the awkward combination
of a generic USB vendor ID and Toradex specific product ID.
Use the Toradex vendor ID by default and do not specify the product ID
since it will get filled in by the usbg service.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
|
|
This allows to use the same schema for Colibri iMX7 with raw
NAND and eMMC flash.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
|
|
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
|
|
Signed-off-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
Acked-by: Dominik Sliwa <dominik.sliwa@toradex.com>
|
|
Signed-off-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
Acked-by: Dominik Sliwa <dominik.sliwa@toradex.com>
|
|
Signed-off-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
Acked-by: Dominik Sliwa <dominik.sliwa@toradex.com>
|
|
https://github.com/morphis/meta-snappy/commit/f6df6a63756d9976bcccb5fefdbc64e5cc794c99
Signed-off-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
Acked-by: Dominik Sliwa <dominik.sliwa@toradex.com>
|
|
The Toradex product ID from the config block is decimal coded
ASCII with leading zeros. In C and bash, leading zeros means octal
number... This leads to wrong calculation of the USB product ID
and to errors if the product ID contains the number 8 or 9 (since
those are not valid under the octal numeric system).
Fix both issues by explicitly force decimal system for the Toradex
product ID.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
|
|
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
|
|
Prevents a warning when not building for one of the machines which have a g1.schema
| WARNING: ...libusbgx/libusbgx_git.bb: Unable to get checksum for libusbgx SRC_URI entry g1.schema: file could not be found
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
|
|
Update libusbgx to a version with OS Descriptor support for better
interoperability with Windows operating systems. Add configuration
using RNDIS with Toradex vendor id. Read serial number and product
id from the device tree to configure the USB gadget device.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
Acked-by: Max Krummenacher <max.krummenacher@toradex.com>
|
|
libusbgx (libusbg-neXt) is a fork of libusbg which is more actively
maintained.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
Acked-by: Max Krummenacher <max.krummenacher@toradex.com>
|
|
Move the appends on recipe's which are not in an essential required layer
into a target layer specific directory and add them only if said layer
is active in build/conf/bblayers.conf.
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
Acked-by: Stefan Agner <stefan.agner@toradex.com>
|
|
Despite the reordering of when to load the USB Gadget configuration
with commit fadb16f153aa ("libusbg: enable USB Gadgets after Connman
to avoid interference") the network device ended up to be down after
startup. It seems that there is a race condition in systemd-networkd
which triggers especially in combination with USB Gadget, a issue on
the systemd issue tracker has been created:
https://github.com/systemd/systemd/issues/1784
There have been various changes related to link handling between
systemd 219 and the latest version 227. It is possible that the issue
is already solved.
Until this has been resolved, explicitly load the USB Gadget before
starting systemd-networkd seems to get around the issue.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
|
|
It seems that Connman disables ALL interfaces on startup, even if
they are listed in NetworkInterfaceBlacklist. This leads to the
following situation:
networkd starts...
usb0 gets created...
networkd sets the device up (event triggered)
connman starts... and sets the device down!
networkd does not restart the device in this situation. To avoid
the situation, create the USB Gadget functions after connman. It
seems that if connman is up and running, it does not touch the
device anymore.
|
|
Add updated version of USB gadget library. Also add a systemd start-
script which loads the default schema located at /etc/usbg/g1.schema.
|