Conversion to DSA broke 802.2+LLC+SNAP packet processing. Frames
received by napi_complete_done with GRO and DSA have transport_header
set two bytes short, or pointing 2 bytes before network_header &
skb->data. As snap_rcv expects transport_header to point to SNAP
header (OID:PID) after LLC processing advances offset over LLC header
(llc_rcv & llc_fixup_skb), code doesn't find a match and packet is
dropped.
Image built at this commit operates properly:
86dadeba48 - generic: add patch for GPON-ONU-34-20BI quirk
Image built at following commit exhibits the issue:
337e36e0ef - ipq806x: convert each device to DSA implementation
As issue is LLC specific, to avoid impacting non-LLC traffic, and to
follow up on original assumption made on kernel commit fda55eca5a33
("net: introduce skb_transport_header_was_set()") stating "network
stacks usually reset the transport header anyway", llc_fixup_skb to
reset and advance the offset. llc_fixup_skb already assumes the LLC
header is at skb->data, and by definition SNAP header immediately
follows.
Signed-off-by: Antonio Pastor <antonio.pastor@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/17220
Signed-off-by: Robert Marko <robimarko@gmail.com>
Adds latest 6.6 patches from the Raspberry Pi repository.
These patches were generated from:
https://github.com/raspberrypi/linux/commits/rpi-6.6.y/
With the following command:
git format-patch -N v6.6.67..HEAD
(HEAD -> 811ff707533bcd67cdcd368bbd46223082009b12)
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
Simplify patch with gpiochip_add_data, struct reduction, new GPIO API,
and header cleanup.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16635
Signed-off-by: Robert Marko <robimarko@gmail.com>
These only work with and are useful with mach files. Now that those are
gone, this can go too.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16635
Signed-off-by: Robert Marko <robimarko@gmail.com>
This is no longer used by mac80211 local patches.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16635
Signed-off-by: Robert Marko <robimarko@gmail.com>
This goes away in kernel 6.12 upstream.
Get ahead of the curve by fixing all the local patches as well.
Get rid of CONFIG_OF. It's needed now.
Refreshed all patches.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16635
Signed-off-by: Robert Marko <robimarko@gmail.com>
Commit 56d97fff55 backported leds-ktd202x from upstream but didn't add the
generic config symbol.
Fixes: 56d97fff55 ("generic: backport support for KTD2026/7 rgb(w) led controller")
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/17396
Signed-off-by: Robert Marko <robimarko@gmail.com>
Fix bogus reference to kmod-nf-conntrack-timeout, fixing the warning
`WARNING: Makefile 'package/kernel/linux/Makefile' has a dependency on
'kmod-nf-conntrack-timeout', which does not exist`.
Fixes: 0e2dcfc4f4 ("netfilter: add kmod-nfnetlink-ct{helper,timeout}")
Signed-off-by: Joel Low <joel@joelsplace.sg>
Link: https://github.com/openwrt/openwrt/pull/17388
Signed-off-by: Robert Marko <robimarko@gmail.com>
Kernel 6.6 requires LED node names to be prefixed via "led-", otherwise
probing the LED will fail, so update our downstream patch adding the LED.
Signed-off-by: Richard Schneidt <ricsc@users.noreply.github.com>
Link: https://github.com/openwrt/openwrt/pull/17330
Signed-off-by: Robert Marko <robimarko@gmail.com>
The former is deprecated in favor of nvmem-layout. In preparation for
eventual removal from the kernel, do so here.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16097
Signed-off-by: Robert Marko <robimarko@gmail.com>
The former is deprecated in favor of nvmem-layout. In preparation for
eventual removal from the kernel, do so here.
Some of these are leftovers from nvmem-layout conversion.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16097
Signed-off-by: Robert Marko <robimarko@gmail.com>
Add kmods for the following RP1 options that not all users
will necessarily need or want compiled in:
* Composite video
* Display video
* LED control
* PWM control
* Serial video
Build system: x86/64
Build-tested: bcm2712/RPi5B
Run-tested: bcm2712/RPi5B
Signed-off-by: John Audia <therealgraysky@proton.me>
Link: https://github.com/openwrt/openwrt/pull/17233
Signed-off-by: Robert Marko <robimarko@gmail.com>
Build in several options RP1-specific features rather than
generating additional kmods for them since bcm2712 is unique to
RPi5B only.
Signed-off-by: John Audia <therealgraysky@proton.me>
Link: https://github.com/openwrt/openwrt/pull/17233
Signed-off-by: Robert Marko <robimarko@gmail.com>
Cherry-pick patches to support building RP1 modules.
Signed-off-by: John Audia <therealgraysky@proton.me>
Link: https://github.com/openwrt/openwrt/pull/17233
Signed-off-by: Robert Marko <robimarko@gmail.com>
This script was rendered obsolete after xrx200 target switched to DSA and
breaking config migration from old releases to 22.03.
Signed-off-by: Mustafa Can Elmacı <mustafacan@elmaci.net>
Link: https://github.com/openwrt/openwrt/pull/17383
Signed-off-by: Robert Marko <robimarko@gmail.com>
This platform has a bogus PCI ID for the chip. Correct it in DTS.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16555
Signed-off-by: Robert Marko <robimarko@gmail.com>
These two devices use different wifi chips and as a result, different
calibration sizes. Move the differences out of dtsi.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16555
Signed-off-by: Robert Marko <robimarko@gmail.com>
- drop patches accepted upstream
- include build fixes in the tarball
- based on https://github.com/nbd168/backports commit 410656ef04d2
Signed-off-by: Felix Fietkau <nbd@nbd.name>
The upstream mtk-sd driver did not perform specific timing
optimization for MT762x series SoC, hence the SDHC peripheral
of some boards cannot run at too high frequency. Reduce the
maximum clock frequency to fix the mmc read/write error.
Closes: https://github.com/openwrt/openwrt/issues/17364
Signed-off-by: Shiji Yang <yangshiji66@qq.com>
Link: https://github.com/openwrt/openwrt/pull/17375
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Short specification:
* 650/600/216 MHz (CPU/DDR/AHB)
* 2x 10/100 Mbps Ethernet, passive PoE support
* 64 MB of RAM (DDR2)
* 16 MB of FLASH
* 2T2R 2.4 GHz with external PA, up to 30 dBm (1000mW)
* 2x internal 14 dBi antennas
* 8x LED, 1x button
* No UART on PCB on some versions
* Display panel with 2x buttons (F/N) not supported (and not relevant in OpenWrt)-
Flash instructions
* Connect PC with 192.168.0.141 to WAN port
* Install a TFTP server on your PC ('atftp' is doing the job for instance)
* Copy your firmware in the TFTP folder as upgrade.bin
* Power up device pushing the 'reset' button
* The device shall upload upgrade.bin, install it and reboot
* Device shall be booting on 192.168.1.1 as default
Signed-off-by: Joan Moreau <jom@grosjo.net>
Link: https://github.com/openwrt/openwrt/pull/17279
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
The kernel logs the error "bcm6368_nand 10000200.nand: there is not valid
maps for state default" on boot and all nand pins show as UNCLAIMED in
sysfs pinmux-pins.
bcm6362.dtsi, bcm6368.dtsi and bcm63268.dtsi use the undocumented property
group which the driver doesn't understand. This has been documented upstream
in commit caf963efd4b0b9ff42ca12e52b8efe277264d35b.
Replacing group with pins allows the nand pins to be properly configured.
Signed-off-by: Kyle Hendry <kylehendrydev@gmail.com>
[add bcm636/bcm6368 and fix commit title]
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
The GS1900 images have been updated to have a larger firmware partition,
bumping the compatibility version to 2.0. However, since this version is
generated on first boot and the default was used, these images still
advertised 1.0 after a fresh install.
Add a new uci-defaults script that will generate the correct version for
all affected Zyxel GS1900 devices.
Fixes: 35acdbe909 ("realtek: merge Zyxel GS1900 firmware partitions")
Signed-off-by: Sander Vanheule <sander@svanheule.net>
Not needed. Everything can be replaced with devm.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16869
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Shorter and more consistent with other drivers.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16869
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
I have no idea why this is even here. Simplifies the code with direct
returns.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16869
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Not needed as all the older kernels are gone.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16869
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
This file does not use a coding style consistent with the kernel. Run
clang-format with the kernel's .clang-format file.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16869
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Two use AR9287 and one AR9381. Both have different calibration sizes.
Move differences out of wifi node to make it clearer what's what.
qca,no-eeprom needs to stay for 7362sl as there's no nvmem equivalent
for caldata_extract_reverse
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/17278
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
These units use AR9287, which has a calibration size of 3d8.
Also fixed compatible string to the proper one to indicate AR9287.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/17278
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Makes it clear that the calibration size is correct as most ar72xx
devices use older wifi chips.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/17278
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
These devices use AR9287, which uses 3d8 as the calibration size, not
440 like newer chips do. Add a compatible line to make it clear that
this is the case.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/17278
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
These three devices use AR9287 chips, which have a calibration size of 3d8.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/17278
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Specification:
- MT7986 CPU using 2.4GHz and 5GHz WiFi (both AX)
- MT7531 switch
- 512MB RAM
- 128MB NAND flash (MX35LF1GE4AB-Z4I) with two UBI partitions with identical size
- 1 multi color LED (red, green, blue, white) connected via GCA230718 (Same as D-Link M30 A1)
- 3 buttons (WPS, reset, LED on/off)
- 1x 2.5 Gbit WAN port with Maxlinear GPY211C
- 4x 1 Gbit LAN ports
Disassembly:
- There are five screws at the bottom: 2 under the rubber feet, 3 under the label.
- After removing the screws, the white plastic part can be shifted out of the blue part.
- Be careful because the antennas are mounted on the side and the top of the white part.
Serial Interface
- The serial interface can be connected to the 4 pin holes next to/under the antenna cables.
- Note that there is another set of 4 pin holes on the side of the board, it's not used.
- Pins (from front to rear):
- 3.3V (do not connect)
- TX
- RX
- GND
- Settings: 115200, 8N1
MAC addresses:
- MAC address is stored in partition "Odm" at offset 0x81 (for example XX:XX:XX:XX:XX:52)
- MAC address on the device label is ODM + 1 (for example XX:XX:XX:XX:XX:53)
- WAN MAC is the one from the ODM partition (for example XX:XX:XX:XX:XX:52)
- LAN MAC is the one from the ODM partition + 1 (for example XX:XX:XX:XX:XX:53)
- WLAN MAC (2.4 GHz) is the one from the ODM partition + 2 (for example (XX:XX:XX:XX:XX:54)
- WLAN MAC (5 GHz) is the one from the ODM partition + 5 (for example (XX:XX:XX:XX:XX:57)
Flashing via OEM web interface:
- Currently not supported because image crypto is not known
Flashing via recovery web interface:
- This is only working if the first partition is active because recovery images are always flashed to the active partition and OpenWrt can only be executed from the first partition
- Use a Chromium based browser, otherwise firmware upgrade might not work
- Recovery web interface is accessible via 192.168.200.1 after keeping the reset button pressed during start of the device until the LED blinks red
- Upload the recovery image, this will take some time. LED will continue flashing red during the update process
- The after flashing, the recovery web interface redirects to http://192.168.0.1. This can be ignored. OpenWrt is accessible via 192.168.1.1 after flashing
- If the first partition isn't the active partition, OpenWrt will hang during the boot process. In this case:
- Download the recovery image from https://github.com/RolandoMagico/openwrt/releases/tag/M60-Recovery-UBI-Switch (UBI switch image)
- Enable recovery web interface again and load the UBI switch image. This image works on the second partition of the M60
- OpenWrt should boot now as expected. After booting, flash the normal OpenWrt sysupgrade image (for example in the OpenWrt web interface)
- Flashing a sysupgrade image from the UBI switch image will make the first partition the active partition and from now on, default OpenWrt images can be used
Flashing via Initramfs:
- Before switching to OpenWrt, ensure that both partitions contain OEM firmware.
- This can be achieved by re-flashing the same OEM firmware version again via the OEM web interface.
- Flashing via OEM web interface will automatically flash the currently not active partition.
- Open router, connect serial interface
- Start a TFTP server at 192.168.200.2 and provide the initramfs image there
- When starting the router, select "7. Load Image" in U-Boot
- Settings for load address, load method can be kept as they are
- Specify host and router IP address if you use different ones than the default (Router 192.168.200.1, TFTP server 192.168.200.2)
- Enter the file name of the initramfs image
- Confirm "Run loaded data now?" question after loading the image with "Y"
- OpenWrt initramfs will start now
- Before flashing OpenWrt, create a backup of the "ubi" partition. It is required when reverting back to OEM
- Flash sysupgrade image to flash, during flashing the U-Boot variable sw_tryactive will be set to 0
- During next boot, U-Boot tries to boot from the ubi partition. If it fails, it will switch to the ubi1 partition
Reverting back to OEM:
- Boot the initramfs image as described in "Flashing via Initramfs" above
- Copy the backed up ubi partition to /tmp (e.g. by using SCP)
- Write the backup to the UBI partition: mtd write /tmp/OpenWrt.mtd4.ubi.bin /dev/mtd4
- Reboot the device, OEM firmware will start now
Signed-off-by: Roland Reinl <reinlroland+github@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/17296
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Referencing commit a1837135e0
Hardware
--------
SoC: Qualcomm Atheros QCA9558
RAM: 128M DDR2 (Nanya NT5TU64M16HG-AC)
FLASH: 128M SPI-NAND (Spansion S34ML01G100TFI00)
WLAN: QCA9558 3T3R 802.11 bgn
ETH: Qualcomm Atheros QCA8337
UART: 115200 8n1
BUTTON: Reset - WPS - "Router" switch
LED: 2x system-LED, 2x wlan-LED, 1x internet-LED,
2x routing-LED
LEDs besides the ethernet ports are controlled
by the ethernet switch
MAC Address:
use address(sample 1) source
label cc:e1:d5:xx:xx:ed art@macaddr_wan
lan cc:e1:d5:xx:xx:ec art@macaddr_lan
wan cc:e1:d5:xx:xx:ed $label
WiFi4_2G cc:e1:d5:xx:xx:ec art@cal_ath9k
Installation from Serial Console
------------
1. Connect to the serial console. Power up the device and interrupt
autoboot when prompted
2. Connect a TFTP server reachable at 192.168.11.10/24
to the ethernet port. Serve the OpenWrt initramfs image as
"openwrt.bin"
3. Boot the initramfs image using U-Boot
ath> tftpboot 0x84000000 openwrt.bin
ath> bootm 0x84000000
4. Copy the OpenWrt sysupgrade image to the device using scp and
install it like a normal upgrade (with no need to keeping config
since no config from "previous OpenWRT installation" could be kept
at all)
# sysupgrade -n /path/to/openwrt/sysupgrade.bin
Installation from Web Interface
------------
To flash just do a firmware upgrade from the stock firmware (Buffalo
branded dd-wrt) with squashfs-factory.bin
Signed-off-by: Edward Chow <equu@openmail.cc>
Link: https://github.com/openwrt/openwrt/pull/17227
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
The ARIA block cipher is pretty uncommon in TLS, deactivate it for now.
This saves some space and reduces the possible variations and attack
vectors of mbedtls.
ARIA support was deactivated in OpenWrt 23.05 by default.
Link: https://github.com/openwrt/openwrt/pull/17342
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Fixes the following error by backporting upstream update:
```
scripts/dtc/pylibfdt/libfdt_wrap.c: In function ‘_wrap_fdt_next_node’:
scripts/dtc/pylibfdt/libfdt_wrap.c:5581:17: error: too few arguments to function ‘SWIG_Python_AppendOutput’
5581 | resultobj = SWIG_Python_AppendOutput(resultobj, val);
| ^~~~~~~~~~~~~~~~~~~~~~~~
```
Fixes: https://github.com/openwrt/openwrt/issues/17345
Signed-off-by: Tianling Shen <cnsztl@immortalwrt.org>
Link: https://github.com/openwrt/openwrt/pull/17352
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
This reverts commit 3bd2cee9bf.
Based on my testing, we are unable to control USB power supply
through some third-party tools (e.g. uhubctl) or sysfs in userspace.
Signed-off-by: Shiji Yang <yangshiji66@qq.com>
Link: https://github.com/openwrt/openwrt/pull/17339
Signed-off-by: Robert Marko <robimarko@gmail.com>
set macaddress correctly for board
Signed-off-by: Florian Maurer <f.maurer@outlook.de>
Link: https://github.com/openwrt/openwrt/pull/17305
Signed-off-by: Robert Marko <robimarko@gmail.com>
The blocksize was too high, resulting in forgetting the config on sysupgrade
It is not needed for SPI-NOR.
Signed-off-by: Florian Maurer <f.maurer@outlook.de>
Link: https://github.com/openwrt/openwrt/pull/17305
Signed-off-by: Robert Marko <robimarko@gmail.com>
The boards where renamed, but BUILD_DEVICES was not adapted. This
variable points to the board name. Without this change the u-boot
binaries are not selected in the configuration.
Copy the u-boot binaries under the BUILD_DEVICES name as it is expected
by the image scripts.
Fixes: 33e23e8922 ("build: d1: add SUPPORTED_DEVICES")
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
No need to pass the option no-warn-rwx-segments. Since v2.12, TF-A
automatically selects it if needed.
A patch is added to revert commit 03a581e2 ("feat(stm32mp1-fdts): remove
RTC clock configuration").
This commit removed RTC clock configuration, as it assumed that it was done
correctly by OPTEE.
But it is not the case. Without this patch the RTC is in a bad state,
consequently the wifi module cannot be initialized.
stm32_rtc 5c004000.rtc: rtc_ck is slow
stm32_rtc 5c004000.rtc: Can't enter in init mode. Prescaler config failed.
stm32_rtc: probe of 5c004000.rtc failed with error -110
sdio mmc1:0001:1: Direct firmware load for brcm/brcmfmac43430-sdio.st,stm32mp135f-dk.bin failed with error -2
brcmfmac: brcmf_sdio_htclk: HT Avail timeout (1000000): clkctl 0x50
Tested on STM32MP135F-DK.
Signed-off-by: Thomas Richard <thomas.richard@bootlin.com>
Link: https://github.com/openwrt/openwrt/pull/17243
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
