Backport upstream patch fixing support for 6.18 due to kernel upstream
commit b74947b4f6ff ("wifi: cfg80211/mac80211: Add support to get radio
index")
Link: https://github.com/openwrt/openwrt/pull/20964
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Migrate wifi compatible from ralink,rt2880-wmac to ralink,rt2880-wifi to
align to new 6.18 backports version.
Link: https://github.com/openwrt/openwrt/pull/20964
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Drop all upstreamed patch and refresh all affected patch.
Minor changes were needed to the
350-mac80211-allow-scanning-while-on-radar-channel due to changes in
6.18.
Also some changes were needed to downstream patch due to upstream commit
b74947b4f6ff ("wifi: cfg80211/mac80211: Add support to get radio index")
Link: https://github.com/openwrt/openwrt/pull/20964
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Add opending patch fixing resource_size(9 if resource start and end is
set to zero.
This is to indirectly fix a helper rework in ath11k qmi code that cause
firmware load panic.
Link: https://github.com/openwrt/openwrt/pull/20964
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Move reserved_mem backports patch from airoha to generic dir as it will
be also used by backports package based on 6.18.
Link: https://github.com/openwrt/openwrt/pull/20964
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Backport support for faux base driver needed for backports package to
compile with kernel 6.18.
Link: https://github.com/openwrt/openwrt/pull/20964
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Add missing address-cells and size-cells properties under the switch's
ports node to fix a multitude of below build warnings:
Warning (reg_format): /soc@0/mdio@90000/ethernet-switch@11/ports/port@1:reg: property has invalid length (4 bytes) (#address-cells == 2, #size-cells == 1)
Warning (avoid_default_addr_size): /soc@0/mdio@90000/ethernet-switch@11/ports/port@1: Relying on default #address-cells value
Signed-off-by: George Moussalem <george.moussalem@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/20982
Signed-off-by: Robert Marko <robimarko@gmail.com>
Although the bootloader tries to 'fixup' the qpic nand node, it actually
can't find it as the node was renamed to spi based on the new driver
architecture. The added benefit is that it also silences build warning:
Warning (spi_bus_reg): /soc@0/spi@79b0000/partitions: missing or empty reg property
Signed-off-by: George Moussalem <george.moussalem@outlook.com>
---
Node rename:
old node name: qpic-nand@79b0000
new node name: spi@79b0000
Bootloader logs:
Booting using the fdt blob at 0x44d3fc78
Uncompressing Kernel Image ... OK
Loading Device Tree to 4a3f6000, end 4a3ff805 ... OK
fdt_fixup_qpic: QPIC: unable to find node '/soc/qpic-nand@79b0000'
parse_fdt_fixup: unable to find node '/soc/qpic-nand@79b0000/'
Link: https://github.com/openwrt/openwrt/pull/20982
Signed-off-by: Robert Marko <robimarko@gmail.com>
Avoids having to create a custom LED for wifi.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/21059
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
libunwind fails to compile because the include for the WORDSIZE
definition was missing when compiling with musl libc.
This lead to unw_word_t being defined as 64 bit long instead
of the correct 32 bit.
Signed-off-by: David Bauer <mail@david-bauer.net>
This partially reverts 78d259e7d2.
That commit broke mounting the root fs on Sophos RED 15w due to
partition size mismatch, and changed the partition sequence, causing
the OpenWrt partition numbers to change from mtd3:kernel/mtd4:ubi to
mtd4:kernel/mtd5:ubi.
Fixes: 78d259e7d2 ("mpc85xx: fix some dtc warnings")
Signed-off-by: Shine <4c.fce2@proton.me>
Link: https://github.com/openwrt/openwrt/pull/20745
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
The stability issue doesn't seem to be found on newer rockchip SoCs
like RK3528, RK3576 and RK3588, so drop the sdr50 limitation.
In addtion, the HS200 limitation for NanoPC T6 is also removed as
upstream fixed the real issue in mmc driver.
Signed-off-by: Tianling Shen <cnsztl@immortalwrt.org>
Link: https://github.com/openwrt/openwrt/pull/20432
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
f19dd37fb467 odhcp6c: accept EUI64 and random for interface ID
e955729c781f formal: fix workflow permissions
9557d3e245aa vscode: set tabs to 8 whitespaces by default
98dfd156f399 src: fix shadowed local variables
f98b6ec823...f19dd37fb4
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
Every attempt to update a device with NOR flash
and RouterBOOT v6 resulted in the error:
“RouterBOOT 6 and earlier requires ELF-in-YAFFS image.”
The cause was that xargs did not fully remove whitespace,
so the condition "$bootentry" != "kernel" always evaluated to true.
Signed-off-by: Marcin Maj <marcinmajsc@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/21034
Signed-off-by: Robert Marko <robimarko@gmail.com>
Commit 8f1914053b ("airoha: replace I2S patch with upstream pending
version") introduced a kernel panic with I2S driver.
Add the fixed patch to fix the kernel panic.
Fixes: 8f1914053b ("airoha: replace I2S patch with upstream pending version")
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Enable I2S sound driver and add nodes for eMMC RFB board since it's the
one that always have the I2C codec and can use the sound card.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Replace I2S sound patch with upstream pending version. This adds ETDM
support and fix various bug in the downstream implementation.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
This is about as simple as it gets, it's able to start up and put
the onboard switch into dumb switch mode and then send untagged
frames which become available on every switch port.
It is out-of-tree to allow for rapid development and it is being
proposed now because even in this state it brings a lot of value
to the EcoNet platform and it is a fairly complex ethernet system
so it will take some time before the driver is in a state that
may be considered for upstreaming.
Signed-off-by: Caleb James DeLisle <cjd@cjdns.fr>
Link: https://github.com/openwrt/openwrt/pull/20685
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
In preparation for the introduction of ethernet, this patch normalizes the
partitions in the devices in DT so that the partition to upgrade will
always be called tclinux, no matter what it is called from the factory.
Also fix a math error in Nokia G240G-E partition table and remove leading
zeros from paritition offsets.
Finally, add NVMEM mac-address entries where they are left by the
factory. In preparation for the introduction of Ethernet and Wifi
support.
Signed-off-by: Caleb James DeLisle <cjd@cjdns.fr>
Link: https://github.com/openwrt/openwrt/pull/20685
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Added missing platform data struct for RK3576 to fix boot issues,
fixed general stability issues and also improved random write
performance.
Signed-off-by: Tianling Shen <cnsztl@immortalwrt.org>
Link: https://github.com/openwrt/openwrt/pull/21028
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Fixes:
```
LzmaDecode.c: In function 'RangeDecoderReverseBitTreeDecode':
LzmaDecode.c:199:5: error: 'symbol' undeclared (first use in this function)
199 | symbol |= (bit << i);
| ^~~~~~
LzmaDecode.c:199:5: note: each undeclared identifier is reported only once for each function it appears in
```
Fixes: 4ffd5aa239 ("treewide: fix coccinelle checks")
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Fixes this compile problem:
```
drivers/net/ethernet/ralink/esw_rt3050.c: In function 'rt3050_esw_init':
drivers/net/ethernet/ralink/esw_rt3050.c:1467:12: error: 'pdev' is used uninitialized [-Werror=uninitialized]
1467 | if (!pdev)
| ^
drivers/net/ethernet/ralink/esw_rt3050.c:1461:33: note: 'pdev' was declared here
1461 | struct platform_device *pdev;
| ^~~~
cc1: all warnings being treated as errors
```
Fixes: 4ffd5aa239 ("treewide: fix coccinelle checks")
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Instead of using tag name, directly use git commit for the base feed
string if we detect we are generating an SDK from a buildbot (using
CONFIG_BUILDBOT)
This improve security and generates consistent feed.conf.default files
following the pattern used by feeds repository.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
There is currently a logic error with generating the base feed string.
In the scenario where the commit of a Tag match the latest commit of a
stable branch, the stable branch have priority over the tag name.
This results in generating SDK that actually target the branch instead
of the specific tag.
To fix this invert the logic and give priority to the tag name.
Fixes: #18583
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Branch named main was never filter out on parsing the GIT_BRANCH define.
Add main to the name of branch that should be filtered.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Makefile foreach works only on parsing the Makefile and in this specific
case only works if the package is already extracted and file actually
exist.
On scenario where the package still has to be built, foreach doesn't
find any file causing Host/Install to not install any .m4 file.
To handle this, use a shell for loop that scan files in the
Host/install.
Fixes: c820f097e0 ("tools: gnulib: install .m4 file with gl_ prefix")
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Fix wrong output using '%d' format when byte count parameter is not
given.
Signed-off-by: Shiji Yang <yangshiji66@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/21013
Signed-off-by: Robert Marko <robimarko@gmail.com>
cdedfd891bdf nss-dp: fix hardware STP state sync with the kernel
19c51af0c5be nss-dp: switchdev: improve FDB cleanup for roaming
Signed-off-by: Robert Marko <robimarko@gmail.com>
a284e7751fa7 file: bump sz_size to 64 bits
ffb9961c1f8b file: provide user name and group name lookups for stat listings
Signed-off-by: Robert Marko <robimarko@gmail.com>
Fixes compile errors caused by static const arrays;
The lantiq firmware is const now.
ATTRIBUTE_GROUPS() does not take const.
Fixes: 5b263f3360 ("treewide: add const to static arrays")
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
GL-MV1000 external SD controller was broken by a recent kernel update
due to a mistake in Device Tree: "regulator-gpio" was defined without
required "gpios" property.
Removing this regulator definition from the device tree fixes this issue
without any side effects so far.
"main" branch uses upstream devicetree for this device, so add a kernel
patch for the "mvebu" kernel family.
Relevant discussions:
https://github.com/openwrt/openwrt/issues/20309https://github.com/openwrt/openwrt/pull/20378
Signed-off-by: Mikhail Zadorozhnyi <zador.blood.stained+git@protonmail.com>
Link: https://github.com/openwrt/openwrt/pull/20390
Signed-off-by: Robert Marko <robimarko@gmail.com>
Backport upstream patch for C++23 support on ax_cxx_compile_stdcxx.m4.
This is needed by Asterisk package to correctly configure.
It's currently not possible to bump to a more recent autoconf-archive
release 2024.10.16 as it's currently affected by multiple BUGs ([1])
[1] e1acc7a67b
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
It was found that there is currently a conflict for the cond.m4 that
is also shipped by automake, making the gnulib one having priority causing
problem with finding AM_CONDITIONAL macro.
To handle this, install gnulib .m4 file with a gl_ prefix to the
filename.
This make sure gnulib .m4 file won't have name conflict with automake
.m4 default files permitting correct autoreconf run of any affected
package by this.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Microchip EV23X71A is a LAN9696 based EVB.
Specifications:
* CPU: Microchip LAN9696 switch SoC
* DRAM: 1GB DDR4
* Storage:
* 2MB QSPI NOR
* 4GB eMMC
* Networking:
* 24 x 10/100/1000 RJ45 via LAN8814 Quad PHY-s over QSGMII
* 4 x 100/1000/2500/5000/10000 SFP+ ports
* 1 x 10/100/1000 management RJ45 via LAN8840 PHY over RGMII (U-Boot too)
* USB: 1 x USB2.0 Type-A
* Management via USB-C (MCP2200):
* UART @ 115200 baud
* GPIO-s for bootstrap, reset and clock selection
* DIP switch for boostrap configuration
* LED-s:
* 2 per networking port (Green and Yellow)
* Green status LED
* Yellow reset LED
* Hard reset button
* Power:
* 12V DC barrel jack
* 48/56V DC screw terminal
* Selectable via toggle switch
* PTP support:
* Sync-E DPLL ZL30732B to generate the board required clocks
* Two SMAs for PTP and two for Station clock inputs and outputs
* Two ITU-T G.8275-compliant RS-422 interfaces for PTP applications
* External PoE:
* Option for PoE add-on, like EV14Y36A (IEEE 802.3af/at/bt Type 4
standard com-pliant)
* Option for external CPU control via SPI and PCIe
Installation instructions:
1. Connect to UART via the USB-C port
2. Connect the management port
3. Boot and interrupt U-Boot
4. TFTP the OpenWrt initramfs image and boot it
5. SCP the OpenWrt eMMC GPT image to a running OpenWrt initramfs to /tmp
openwrt-microchipsw-lan969x-microchip_ev23x71a-squashfs-emmc-gpt.img.gz
And decompress it via:
gzip -d /tmp/openwrt-microchipsw-lan969x-microchip_ev23x71a-squashfs-emmc-gpt.img.gz
6. Wipe eMMC with:
dd if=/dev/zero of=/dev/mmcblk0 bs=1M
7. Flash OpenWrt eMMC image with:
dd if=/tmp/openwrt-microchipsw-lan969x-microchip_ev23x71a-squashfs-emmc-gpt.img
of=/dev/mmcblk0
After a restart OpenWrt will boot, and then regular sysupgrade can be used
for upgrades.
Signed-off-by: Robert Marko <robert.marko@sartura.hr>
Add a new microchipsw target aimed add supporting Microchip switch
SoC-s.
Start by supporting LAN969x SoC-s as the first subtarget.
Signed-off-by: Robert Marko <robert.marko@sartura.hr>
Userspace handling is deprecated.
Also use led-sources for simplicity.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/20508
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
This AX3000 flying saucer-shaped device is sold under a few names:
- Widelantech WAP430X
http://www.widelantech.com/?m=home&c=View&a=index&aid=131
- Felicomm
- UeeVii UAP200 (sold on amazon.com)
True name, as marked on the PCB and as appears in the DTS, is WAP430X.
Hardware:
- SoC: Mediatek MT7981B
- RAM: 256MiB DDR3
- ROM: 16MiB NOR flash (Winbond 25Q128JVSQ)
- Wired: one GbE RJ45 port
- Wireless: Mediatek MT7976CN DBDC
- Antennas: Internal (two 2.4 GHz 4.5 dBi, three 5 GHz 6 dBi)
- GPIO: two blue LEDs (status, WLAN), one button (reset)
- Power: DC5521 (12V) or 802.3af PoE (48V 0.3A)
- UART: J500 on PCB [3v3] (Rx) (Tx) (Gnd), 115200n
Original firmware is GECOOS 8.0_2024081000, a fork of OpenWrt 21.02
created by Widelantech. All ports are closed except 443 (web management
interface) and 80 (redirects to 443). UART is locked down, but U-Boot
bootloader is friendly.
Installation is done via UART (see note below).
Method 1: press any key during boot to enter MTK U-Boot bootloader,
then choose option 2 (Upgrade firmware) and upload sysupgrade binary
via TFTP (or via Xmodem).
Method 2: put sysupgrade binary on a webserver (http, not https); after
GECOOS loads, choose option 4 (固件在线升级 = online firmware upgrade)
in the menu and enter the URL to the sysupgrade binary.
Note: GECOOS has a slick web management, and firmware can be uploaded
at https://6.6.6.6/#/system/sysupgrade but it's still not clear how to
convert the sysupgrade binary to a factory binary accepted by the web
interface (WIP).
Signed-off-by: Rani Hod <rani.hod@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/20402
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Userspace handling is deprecated.
Also switch LED to led-sources for simplicity.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/20529
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Unneeded semicolon
WARNING comparing pointer to 0
WARNING: NULL check before some freeing functions is not needed.
WARNING: casting value returned by memory allocation function to (u32 *)
ERROR: allocation function on line 378 returns NULL not ERR_PTR on failure
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/19932
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Userspace handling is deprecated.
This is a weird one. It should be using u-boot,env, but the variable
name is unknown. Keep it as is.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/20520
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Tests with ingress shaping and enabled flow control showed really high
packet loss. It seems like the MAC pause frames are not created correctly
when both burst high off is set to the same value as burst high on.
By default, RTL930x has set the burst high values to:
* on: 64K
* off: 32K
Using the same 1:2 ratio seems to solve the high packet loss rate during
UDP tests.
Fixes: 2e74eb6d93 ("realtek: dsa: rtl93xx: Support per port throttling")
Signed-off-by: Sven Eckelmann <se@simonwunderlich.de>
Link: https://github.com/openwrt/openwrt/pull/21011
Signed-off-by: Robert Marko <robimarko@gmail.com>
Fix the cipher implementation to avoid treating empty input as finalizer.
This issue is fixed in the openssl 3.6 branch, but the fix approach from
that branch is not suitable for 3.5, since the code is completely different.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
