Also known as glinet spitz. The mips processor is big endian, so the
current release driver 1.14.1 will not work. I have however tested this
with the big endian patchset and it works.
The board has only 16MB flash, so I have restricted the packages.
The usb port is disabled by a gpio in the upstream dts. I have edited
that to enable the port at boot.
Approved-by: James Haggerty
Approved-by: Lyall Beveridge
If you don’t have this a clean build now prompts you to select
an option for RAVE_SP (causing a failure unless you run with V=sc
and answer the option).
I'm a bit confused about why this is suddenly necessary AND why the
default make kernel_*config doesn't populate it, but it's echoing a
change that went in with the matter PR for artini.
I’m also concerned that this means we probably should be regenerating
other kernel configs (since people might take this SDK and expect it
to work for other archiectures), but the good news there is that most
of them already have the RAVE thingy.
Approved-by: Sophronia Koilpillai
Approved-by: Evan Benn
SW-12275 add support for releasing matter with openwrt
* Update bluetooth make components (other.mk) with correct config option
* Add matter defconfig for all ekh boards
Approved-by: Cedric Van den Bergh
Approved-by: James Haggerty
Approved-by: Julia Page
Plan is to rename ekh01s to include the chip id.
For now, let's just mess with the image name; we can
consider whether we want to modify the model name
(and the internal identifier) once MM8108 settles.
Approved-by: Sophronia Koilpillai
Approved-by: Arien Judge
It's more interesting to see a HaLow activity LED than an
SD card LED, and this lines up with our priorities on the EKH03
(where we now show HaLow link/activity on the single RGB LED).
Approved-by: Milad Mohtashamirad
Approved-by: Lyall Beveridge
This removes the extra functionality based aliases from the ekh03
(which are no longer used) and adds colour based aliases, which
make the diag_override.sh script clearer when we're using
multiple LEDs to construct a colour.
Note that we leave the old aliases in so that the board
will work without our custom LED package (morse-led-rgb).
Approved-by: Sophronia Koilpillai
Approved-by: Evan Benn
Note that this is tricky because we want the LED to be green
when the link is off and purple when the link is on.
This requires a patch to the netdev driver to support invert.
Approved-by: Sophronia Koilpillai
Approved-by: Evan Benn
Some targets may not support USB host mode. So move the kmod packages,
and disable the hardware in artini's dts.
Approved-by: James Haggerty
Approved-by: Sophronia Koilpillai
Add configuration to artini dts and code to the LED drivers to support
bringing up the status led in the desired pattern and color. Without
this patch the multicolor-led driver briefly turns the led off (uboot
turned it on) before diag.sh runs to turn it back on.
Approved-by: Sophronia Koilpillai
Approved-by: James Haggerty
The status LED is now driven by the PWM driver. When the driver is built as
a kmod module, it is loaded after the preinit state is set. This means
the diag.sh script cannot set the color and flashing pattern. Making the
driver built in means the LED can be controlled at preinit stage.
Approved-by: James Haggerty
Approved-by: Sophronia Koilpillai
Add OpenWrt target definitions to produce images for the EKH03 and EKH04
versions of Morse Micro evaluation kits. Included in this commit are:
* kernel config changes to enable pstore for ramoops, USB serial devices,
and other miscellaneous generated changes.
* a preinit script to mutate the board_name propagated through OpenWrt
* inittab changes to remap the default serial interface
The mainline mmc driver offers improvement in CPU performance by
allowing more data utilise DMA. This change resulted in a 3 Mbps
increase in Morse Micro HaLow throughput.
Add the driver by default to the mt76x8 kernel config and the mt7628
device tree.
Previously, we'd just been modifying the distroconfig.txt for
all RPIs. This has the advantage that we'd pick up any upstream
changes.
However, for SDK users this is potentially annoying, as they
may want their own set of customisations. It's also a bit
strange that a 'normal' RPI build would have our HAT overlay.
This changes it so we have a custom distroconfig.txt for
ekh01 devices only which completely replaces the base
distroconfig.txt.
Approved-by: Arien Judge
Approved-by: Evan Benn
We want to use the normal rpi4 device tree, but have OpenWrt
report different model names. To easily use multi-profile builds,
this has to happen somewhere outside the rootfs.
Approved-by: Arien Judge
Approved-by: Evan Benn
Add OpenWrt target definitions to produce images for the EKH01 versions
of Morse Micro evaluation kits. Included in this commit are:
* kernel changes to enable by default SPI support, pstore for ramoops,
USB serial device and the default CPU governor.
* a preinit script to mutate the board_name propagated through OpenWrt
* kernel cmdline changes to support a USB serial adapter for console
access
config.txt is the 'user' config which is retained on upgrade,
whereas distroconfig is the distro defaults that might change
on upgrade.
This also means we pick up the OpenWrt default of disabling bluetooth
and the miniuart.
Approved-by: Arien Judge
Approved-by: Evan Benn
keep.d: (APP-1962) Remove config.txt and cmdline from keep.d
* keep.d: (APP-1962) Remove config.txt and cmdline from keep.d
All filenames in keep.d are retained if the user chooses 'retain config'
in an update. There is no way for the update to override this at update
time. We may want to modify the command line or bootloader config in
a future update, and as we are supporting 'retain configs' we must
remove these files from keep.d. This will only affect updates FROM the
image that includes this patch.
Approved-by: Arien Judge
Approved-by: Takura Mapani
Otherwise, on upgrade we retain the overlay if the rootfs size is the
same (causing confusing behaviour). This was taken from
https://github.com/openwrt/openwrt/issues/9113#issuecomment-1572846457
However, ensure we only clear the bit at the end of the partition to
ensure we don't generate unnecessarily large images for the EKH01.
See https://github.com/openwrt/openwrt/pull/13793
Approved-by: Milad Mohtashamirad
Approved-by: Arien Judge
Approved-by: Evan Benn
For the Raspberry Pi based platforms, we add device tree overlays to
instantiate the appropriate hardware configuration for MM6108.
SPI and SDIO options are available, as well as relevant power save
configuration.
Also add a ramoops overlay.
MM6108 does not support 1.8V but offers voltage negotiation. Add
required quirk to kernel.
Additionally, issues in kernel 5.x versions of the BCM2835 SPI driver
have been reported, particular regarding CS polarity. The patch rolls
the driver back to 4.19.
When you have an RGB LED with every colour mapped individually,
it's useful to have a different colour for link on vs link off,
particularly in the case where the LED has multiple meanings
(i.e. it's a generic status/power LED as well, so shouldn't just be
turned off in the absence of a link).
This invert code is largely copied from similar code
in other triggers which already support invert.
Approved-by: Sophronia Koilpillai
Approved-by: Evan Benn
There's a bug in drivers/dma-buf/heaps/Makefile or /driver/dma-buf/Makefile
that doesn't allow heaps source files to get compiled.
we raised an issue in openwrt:
https://github.com/openwrt/openwrt/issues/13887
Approved-by: Arien Judge
Approved-by: James Haggerty
libcamera requires a number of v4l2 modules without packages
in OpenWrt.
Additionally, add support for the raspberry pi camera variants
APP-2145, APP-2097, APP-2003, APP-2197
Approved-by: James Haggerty
Approved-by: Matthew Forgie
Approved-by: Evan Benn
Sadly, KConfig mechanism for enabling CONFIG_DEV_COREDUMP is for
an (in-tree) driver to do CONFIG_WANT_DEV_COREDUMP.
But because the morse driver is built separately, it has no
way to demand this, and there's no way to manually specify
CONFIG_DEV_COREDUMP (or CONFIG_WANT_DEV_COREDUMP for that
matter). Therefore we hack the kconfig files.
Approved-by: Sophronia Koilpillai
Approved-by: Evan Benn
The MAC address of the GMAC is contained inside the CWMP-Account
number on the label.
The label MAC address alias was defined previously, but it has been
removed with the switch to IPQESS / DSA.
Restore the label MAC address alias.
Fixes: 27b441cbaf ("ipq40xx: drop ESSEDMA + AR40xx DTS nodes")
Signed-off-by: Fabian Bläse <fabian@blaese.de>
Reviewed-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit b22d382ae4eaa1af42930115d91855f402314cac)
Signed-off-by: Tom Herbers <mail@tomherbers.de>
Link: https://github.com/openwrt/openwrt/pull/16459
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
With commit a22d359fa56fe0 VLAN handling was fixed for kernel 6.6.
This restored network connectivity of the devices. For easy testing
backport the fix for 5.15 too.
Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
Link: https://github.com/openwrt/openwrt/pull/16391
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 35e13244aab36958831195a1d2e152aa3ad0ab6c)
Signed-off-by: Goetz Goerisch <ggoerisch@gmail.com>
Change I2C pin flags to GPIO_ACTIVE_HIGH and change SDA to GPIO1.
On my late production GL-AR750 (purchased 2024):
SCL = GPIO16
SDA = GPIO1
Bug report:
I2C bus doesn't work in GL-AR750
https://github.com/openwrt/openwrt/issues/16319
Signed-off-by: Ryan Salsbury <ryanrs@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16406
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 2e626ae2d2488f2d3708d3aed57eea7a1402f00e)
Signed-off-by: Ryan Salsbury <ryanrs@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16446
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
The sysupgrade-tar image build is not defined for this target, do not
add a build instruction for it. The build system will use the definition
from the dna_valokuitu-plus-ex400 board and the build will fail.
This fixes the build of the ramips target.
Fixes: 665c2154ef ("ramips: add basic support for tp-link er605-v2")
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 746e894877e5130a7d5227cd90f90dfe18cc274b)
Add the GPIO pin of the PoE passthrough switch on the Aruba AP-303H.
Power is activated when the pin is low. It enables a PSE chip, so power
is only supplied to downstream devices when they are 802.3af/at
compliant devices.
Ensure you use a sufficient power supply when chaining a consuming
device after the AP.
Signed-off-by: David Bauer <mail@david-bauer.net>
(cherry picked from commit b40cc46cc882ff97f2752e7837173ffc9f884eb9)
This fixes a well-known "LZMA ERROR 1" error on Sercomm NA502s, reported
on the OpneWrt forum [0].
[0] https://forum.openwrt.org/t/206640
Signed-off-by: Andreas Böhler <dev@aboehler.at>
(cherry picked from commit 6d89aa29878a93c744e8c752a68bc05d41252cfc)
Link: https://github.com/openwrt/openwrt/pull/16412
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
configure 2.5G PHY LEDs to:
2500/1000: green with blink on TX/RX
100/10: green+yellow with blink on TX/RX
which is similar to other 1G PHY LEDs, which are:
1000: green with blink on TX/RX
100/10: green+yellow with blink on TX/RX
Fixes: 6cc14bf66aa ("filogic: support Telenor branded ZyXEL EX5700")
Signed-off-by: Yan Cangang <nalanzeyu@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16082
(cherry picked from commit 3de653a0af35c2d6fbc0fd14909f362c6b977e4a)
Link: https://github.com/openwrt/openwrt/pull/16390
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Ubiquiti has a set of UniFi 802.11ax (Wi-Fi 6) AP devices. All models
include "U6" in their names and also have code names with no special
characters (including spaces).
Examples:
1. U6 Lite (codename U6-Lite)
2. U6 Long-Range (codename U6-LR)
3. U6+ (codename U6-PLUS)
4. U6 Pro (codename U6-Pro)
5. U6 Mesh (codename U6-Mesh)
6. U6 Mesh Pro (codename U6-Mesh-Pro)
7. U6 Enterprise (codename U6-Enterprise)
Use proper full names for those devices. Names in OpenWrt/DTS code may
need updating too but it can be handled later.
Cc: Elbert Mai <code@elbertmai.com>
Cc: Daniel Golle <daniel@makrotopia.org>
Cc: Henrik Riomar <henrik.riomar@gmail.com>
Cc: David Bauer <mail@david-bauer.net>
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
(cherry picked from commit 74879140a1aa0b8a8c237e0b67e94c3508e2e249)
Aruba boards now ship with multiple DTS and image-configurations per
image. Newer apboot revs expect a configuration for their hardware to be
present.
Signed-off-by: David Bauer <mail@david-bauer.net>
(cherry picked from commit c7ba5574f52f25334add66c15190b578ac873739)
Some devices lock up on PCIe initialization:
[ 64.309697] PCIe: Port 0 in endpoint mode, skipping.
[ 64.320496] PCIe: Initializing port 1
[ 64.325257] PCIe: BIST FAILED for port 1 (0xffffffffffffffff)
(system hangs here)
Given the ER contains no PCIe peripherals, has no way to attach any
and the stock kernel doesn't have PCIe support either, just disable it.
(cherry picked from commit 308630aea6da1b5618d169d6ebf208aaa249ef72)
Signed-off-by: Jakob Haufe <sur5r@sur5r.net>
Link: https://github.com/openwrt/openwrt/pull/16223
Signed-off-by: Robert Marko <robimarko@gmail.com>
Sometimes the mmc deivce may come up later than kernel attempts to
mount rootfs, resulting kernel panic. Enable rootwait to fix it.
Reported-by: Yangyu Chen <cyy@cyyself.name>
Signed-off-by: Tianling Shen <cnsztl@immortalwrt.org>
Link: https://github.com/openwrt/openwrt/pull/15077
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
According to RTL8221B's datasheet, the PHY requires at least 10ms
for assert and 68ms (recommended) for de-assert. So increase the
assert/de-assert time to 15ms and 68ms respectively.
Fixes: c0c3234e1720 ("mediatek: add support for JDCloud RE-CP-03")
Signed-off-by: Tianling Shen <cnsztl@immortalwrt.org>
Link: https://github.com/openwrt/openwrt/pull/16106
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit d1954aa5353eca59bb76d860b09a68a01e7ec05a)
