That's what @Gemnon pulled and sent me, i'm going to trust him and add his dts to the branch so I can test it, he has both devices to test
Compared to other devices there are lots of diffs in the dts. The device uses alot of different drivers compared to the SDK.
> That's what @Gemnon pulled and sent me, i'm going to trust him and add his dts to the branch so I can test it, he has both devices to test
Compared to other devices there are lots of diffs in the dts. The device uses alot of different drivers compared to the SDK.
he also sent me some firmware that will probably let me extract that dts from uboot
/dts-v1/;/{timestamp=<0x662b6bf8>;description="IOWRT FIT upgrade image";#address-cells=<0x01>;id="IOPSYS";created="20240910-113626";model="PIXLY";iopsys_version="heimgardos-7.3.1.17.2.0-20240910-113626-airoha-en7523-pixly";images{bootloader{data-size=<0x80000>;data-offset=<0x00>;description="bootloader";compression="none";iopsys_bootloader_version="4004";hash-1{value=<0x1bab61650xea59d02e0x46395bb00x9692bc890x4afa97120x251c68230x7b9d2e840x167d198d>;algo="sha256";};};boot{data-size=<0x372052>;data-offset=<0x80000>;description="boot";compression="none";hash-1{value=<0xa1e138b0x74b386ac0x8369ef430x1fe8c0b40x1100e89a0x53e22baa0x1eb561d10x78e2099>;algo="sha256";};};rootfs{data-size=<0x1ae1eb4>;data-offset=<0x3f2054>;description="rootfs";compression="none";hash-1{value=<0x1282c6590xb090f1f90x750cfd4b0x2df9b3910xecdbd8d30x27a7349d0xceaea3700x618cad32>;algo="sha256";};};u-boot-env{data-size=<0x160a>;data-offset=<0x1ed3f08>;description="u-boot-env";compression="none";hash-1{value=<0x607b3c810x7f9699380xf57bc30d0x718c7f2e0x20d1b1dd0x34bd4f540x62f863a60xa943bcb6>;algo="sha256";};};upgrade_bundle{data-size=<0xce7>;data-offset=<0x1ed5514>;description="upgrade_bundle";created="20240910-113626";type="multi";compression="none";hash-1{value=<0x7342db210x3db66b20x983e62dc0x6a84cbde0x755097720x462f77770x3e6376b90x6a585ea3>;algo="sha256";};};};};
he also sent me some firmware that will probably let me extract that dts from uboot
```dts
/dts-v1/;
/ {
timestamp = <0x662b6bf8>;
description = "IOWRT FIT upgrade image";
#address-cells = <0x01>;
id = "IOPSYS";
created = "20240910-113626";
model = "PIXLY";
iopsys_version = "heimgardos-7.3.1.17.2.0-20240910-113626-airoha-en7523-pixly";
images {
bootloader {
data-size = <0x80000>;
data-offset = <0x00>;
description = "bootloader";
compression = "none";
iopsys_bootloader_version = "4004";
hash-1 {
value = <0x1bab6165 0xea59d02e 0x46395bb0 0x9692bc89 0x4afa9712 0x251c6823 0x7b9d2e84 0x167d198d>;
algo = "sha256";
};
};
boot {
data-size = <0x372052>;
data-offset = <0x80000>;
description = "boot";
compression = "none";
hash-1 {
value = <0xa1e138b 0x74b386ac 0x8369ef43 0x1fe8c0b4 0x1100e89a 0x53e22baa 0x1eb561d1 0x78e2099>;
algo = "sha256";
};
};
rootfs {
data-size = <0x1ae1eb4>;
data-offset = <0x3f2054>;
description = "rootfs";
compression = "none";
hash-1 {
value = <0x1282c659 0xb090f1f9 0x750cfd4b 0x2df9b391 0xecdbd8d3 0x27a7349d 0xceaea370 0x618cad32>;
algo = "sha256";
};
};
u-boot-env {
data-size = <0x160a>;
data-offset = <0x1ed3f08>;
description = "u-boot-env";
compression = "none";
hash-1 {
value = <0x607b3c81 0x7f969938 0xf57bc30d 0x718c7f2e 0x20d1b1dd 0x34bd4f54 0x62f863a6 0xa943bcb6>;
algo = "sha256";
};
};
upgrade_bundle {
data-size = <0xce7>;
data-offset = <0x1ed5514>;
description = "upgrade_bundle";
created = "20240910-113626";
type = "multi";
compression = "none";
hash-1 {
value = <0x7342db21 0x3db66b2 0x983e62dc 0x6a84cbde 0x75509772 0x462f7777 0x3e6376b9 0x6a585ea3>;
algo = "sha256";
};
};
};
};
```
Cant be sure extracted it from the firmware. And the boxes i have behave strange there are CX30-s that try to get ip address via dhcp, others that listens on 192.168.39.1, some boxes (HX30, HX40) listen on 192.168.1.1, some of the same on 192.168.5.1. Some not even have model number, still trying to figure out the things (but which ones i tried just fails with decompression error when i try to boot any initramfs image from here). Iam pretty sure they used to be development boxes as some had github version of iowrt thing on them.
Also by default ssh password auth disabled, i can enable it, some boxes let me change the burnt in root password (which is the same from the hash on all boxes but i have no idea what) via luci (and for some reason it only change ssh root password not the device password), for some luci just gives an error and have to backup config, alter /etc/shadow reload config for access. Serial port can be accessed on all via holes on the back of the boxes, uboot has bootstopkeysha256 set to something (will have to brute-force it) but that can be overriden in env.
This one is the (Unknown Model) which i could not even get paired with a product at Genexis/Heimgard:
I0M0QCCXF (laxy) Have a pon, 2 fxs, 4 gigabit lan port, 256M flash, 512M memory, MT7916D
mtd layout looks like pretty much the same as other ones:
dev: size erasesize name
mtd0: 00100000 00020000 "u-boot"
mtd1: 0ff00000 00020000 "ubi"
mtd2: 0001f000 0001f000 "env1"
mtd3: 0001f000 0001f000 "env2"
mtd4: 00383000 0001f000 "boot1"
mtd5: 00383000 0001f000 "boot2"
mtd6: 01b9c000 0001f000 "rootfs1"
mtd7: 021e8000 0001f000 "rootfs2"
mtd8: 0141a000 0001f000 "overlay1"
mtd9: 0141a000 0001f000 "overlay2"
mtd10: 00193000 0001f000 "art"
mtd11: 0041e000 0001f000 "boarddata"
mtd12: 07d17000 0001f000 "container"
> > Stock dts
>
> Is it really ? :)
Cant be sure extracted it from the firmware. And the boxes i have behave strange there are CX30-s that try to get ip address via dhcp, others that listens on 192.168.39.1, some boxes (HX30, HX40) listen on 192.168.1.1, some of the same on 192.168.5.1. Some not even have model number, still trying to figure out the things (but which ones i tried just fails with decompression error when i try to boot any initramfs image from here). Iam pretty sure they used to be development boxes as some had github version of iowrt thing on them.
Also by default ssh password auth disabled, i can enable it, some boxes let me change the burnt in root password (which is the same from the hash on all boxes but i have no idea what) via luci (and for some reason it only change ssh root password not the device password), for some luci just gives an error and have to backup config, alter /etc/shadow reload config for access. Serial port can be accessed on all via holes on the back of the boxes, uboot has bootstopkeysha256 set to something (will have to brute-force it) but that can be overriden in env.
@Sirherobrine23
This one is the (Unknown Model) which i could not even get paired with a product at Genexis/Heimgard:
I0M0QCCXF (laxy) Have a pon, 2 fxs, 4 gigabit lan port, 256M flash, 512M memory, MT7916D
mtd layout looks like pretty much the same as other ones:
dev: size erasesize name
mtd0: 00100000 00020000 "u-boot"
mtd1: 0ff00000 00020000 "ubi"
mtd2: 0001f000 0001f000 "env1"
mtd3: 0001f000 0001f000 "env2"
mtd4: 00383000 0001f000 "boot1"
mtd5: 00383000 0001f000 "boot2"
mtd6: 01b9c000 0001f000 "rootfs1"
mtd7: 021e8000 0001f000 "rootfs2"
mtd8: 0141a000 0001f000 "overlay1"
mtd9: 0141a000 0001f000 "overlay2"
mtd10: 00193000 0001f000 "art"
mtd11: 0041e000 0001f000 "boarddata"
mtd12: 07d17000 0001f000 "container"
All have Heimgard logo and using heimgard ota so i think they are from the time before Genexis acquired them...
So the listing what i have with the corresponding ota images i could find:
CX20: G0A0ACKW0 (arcee):
https://firmware.homecontrol-cloud.com/heimgard/G0A0ACKW0-HOv1/v1-heimgard/host/arcee-7.3.1.17.2.0/last.itb
CX30: I0A0NCKX0 (pixly)
https://firmware.homecontrol-cloud.com/heimgard/I0A0NCKX0-HOv1/v1-heimgard/host/pixly-7.3.1.17.2.0/last.itb
"Unknown model": I0M0QCCXF (laxy)
https://firmware.homecontrol-cloud.com/heimgard/I0M0QCCXF-HOv1/v1-heimgard/host/laxy-7.3.1.17.2.0/last.itb
HX30: I0ABGCUFD (zephyr)
https://firmware.homecontrol-cloud.com/heimgard/I0ABGCUFD-HOv1/v1-heimgard/host/zephyr-7.3.1.17.2.0/last.itb
HX40: G0A0GCUXD (rodimus)
https://firmware.homecontrol-cloud.com/heimgard/G0A0GCUXD-HOv1/v1-heimgard/host/rodimus-7.3.1.17.2.0/last.itb
All have Heimgard logo and using heimgard ota so i think they are from the time before Genexis acquired them...
@Gemnon could create a branch and add the devices. i currently working with the u-boot for Mikrotik and the xx230v/xx530v and i need to review the merbanan changes for the WED and NPU. You have code write, and for now it's easy to add the devices, look at these commits for example:
@Gemnon could create a branch and add the devices. i currently working with the u-boot for Mikrotik and the xx230v/xx530v and i need to review the merbanan changes for the WED and NPU. You have code write, and for now it's easy to add the devices, look at these commits for example:
- 63edd204022cd48fae562c8d2f54753e64bbd72b
- 19910446b7d48ed634304111ca9555c665d237f3
Uncompressing Kernel Image
lzma compressed: uncompress error 7
Must RESET board to recover
resetting ...
Also when i try to extract it with imxtract it sais:
=> imxtract 0x90000000 kernel-1 0x80008000
## Copying 'kernel-1' subimage from FIT image at 90000000 ...
crc32+ sha1+ Unimplemented compression type 3
Not sure if it has anything to do with uncompress error 7 though (it does with the stock firmware image too so that has nothing to do with it for sure)
@Sirherobrine23
Same result:
```
Uncompressing Kernel Image
lzma compressed: uncompress error 7
Must RESET board to recover
resetting ...
```
Also when i try to extract it with imxtract it sais:
```
=> imxtract 0x90000000 kernel-1 0x80008000
## Copying 'kernel-1' subimage from FIT image at 90000000 ...
crc32+ sha1+ Unimplemented compression type 3
```
Not sure if it has anything to do with uncompress error 7 though (it does with the stock firmware image too so that has nothing to do with it for sure)
It can extract and load lzma heimgard_cx30-kernel.bin, well until "Starting linux..." where it locks up, so lzma works in u-boot, just initramfs image too big for it.
Pretty sure the reason is the size.
It can extract and load lzma heimgard_cx30-kernel.bin, well until "Starting linux..." where it locks up, so lzma works in u-boot, just initramfs image too big for it.
Was changed load address (0x80208000) like it is on other boxes recompiled. And i have lines... And lockup at npu thing (well no initramfs so no wonder it cant load it). That is a progress. At least it can boot stuff. :)
Think have to shrink down initramfs somehow, to be able to boot it. Or move to nand boot.
Tried to boot via seperate ramdisk, but failed most likely because i have no idea how to do it. (ram_try.log)
Tried to boot via seperate uncompressed ramdisk because i figured kernel not able to uncompress it, now it booted fully. (booted.log)
Ethernet works, after copying over /lib/firmware/mediatek from iowrt and reload mt7915e module wifi works too, at least shows up by iw.
Was changed load address (0x80208000) like it is on other boxes recompiled. And i have lines... And lockup at npu thing (well no initramfs so no wonder it cant load it). That is a progress. At least it can boot stuff. :)
Think have to shrink down initramfs somehow, to be able to boot it. Or move to nand boot.
Tried to boot via seperate ramdisk, but failed most likely because i have no idea how to do it. (ram_try.log)
Tried to boot via seperate uncompressed ramdisk because i figured kernel not able to uncompress it, now it booted fully. (booted.log)
Ethernet works, after copying over /lib/firmware/mediatek from iowrt and reload mt7915e module wifi works too, at least shows up by iw.
Was changed load address (0x80208000) like it is on other boxes recompiled. And i have lines... And lockup at npu thing (well no initramfs so no wonder it cant load it). That is a progress. At least it can boot stuff. :)
I recently noticed that the NPU log isn't showing up, i'll look into that, but it should be working. Interestingly, the initramfs loads with less than 10Mb of size, if you're basing the build on my configuration, the files will exceed 22MB because it includes some testing tools and additional things for testings
Ethernet works, after copying over /lib/firmware/mediatek from iowrt and reload mt7915e module wifi works too, at least shows up by iw.
The only things that need firmware are the NPU and the mt7916e, I couldn't find the eeprom for mt7916e in the volume art, but that might be my mistake in how i extracted the firmware from the dump you sent me
> Was changed load address (0x80208000) like it is on other boxes recompiled. And i have lines... And lockup at npu thing (well no initramfs so no wonder it cant load it). That is a progress. At least it can boot stuff. :)
I recently noticed that the NPU log isn't showing up, i'll look into that, but it should be working. Interestingly, the initramfs loads with less than 10Mb of size, if you're basing the build on my configuration, the files will exceed 22MB because it includes some testing tools and additional things for testings
> Ethernet works, after copying over /lib/firmware/mediatek from iowrt and reload mt7915e module wifi works too, at least shows up by iw.
The only things that need firmware are the NPU and the mt7916e, I couldn't find the eeprom for mt7916e in the volume art, but that might be my mistake in how i extracted the firmware from the dump you sent me
I recently noticed that the NPU log isn't showing up, i'll look into that, but it should be working. Interestingly, the initramfs loads with less than 10Mb of size, if you're basing the build on my configuration, the files will exceed 22MB because it includes some testing tools and additional things for testings
Well for me it seems it breaks on decompression because of its size, i think it hits a protected range (or there is some iowrt u-boot limit).
if i just upload the kernel to 0x9600000 and a cpio ramdisk image to 0x86000000 set bootargs to make root=/dev/ram0 and
bootm 0x96000000 0x86000000 it works (and not stuck on npu binary) if i try to boot the generated initramfs it fails with decompression error... But lzma working because lzmad kernel alone booting just fine.
> I recently noticed that the NPU log isn't showing up, i'll look into that, but it should be working. Interestingly, the initramfs loads with less than 10Mb of size, if you're basing the build on my configuration, the files will exceed 22MB because it includes some testing tools and additional things for testings
Well for me it seems it breaks on decompression because of its size, i think it hits a protected range (or there is some iowrt u-boot limit).
if i just upload the kernel to 0x9600000 and a cpio ramdisk image to 0x86000000 set bootargs to make root=/dev/ram0 and
bootm 0x96000000 0x86000000 it works (and not stuck on npu binary) if i try to boot the generated initramfs it fails with decompression error... But lzma working because lzmad kernel alone booting just fine.
Load a uboot chainload and try booting from new uboot
https://sirherobrine23.com.br/airoha_an7523/builds/releases/download/v49/openwrt-airoha-en7523-tplink_xx230v_v1-chainload-uboot.bin
Since the 230 has a header, just skip the 512B and it will start normally:
`tftpboot 0x90000000 192.168.1.100:<230 chainload file> && bootm 0x90000200`
Yes, this one loads your initramfs image just fine (had some errors with ubi but think that is because it is not for this device).
> Load a uboot chainload and try booting from new uboot
>
> https://sirherobrine23.com.br/airoha_an7523/builds/releases/download/v49/openwrt-airoha-en7523-tplink_xx230v_v1-chainload-uboot.bin
>
> Since the 230 has a header, just skip the 512B and it will start normally:
>
> `tftpboot 0x90000000 192.168.1.100:<230 chainload file> && bootm 0x90000200`
Yes, this one loads your initramfs image just fine (had some errors with ubi but think that is because it is not for this device).
We'll have to replace the U-boot, as is the case on all devices. This is bad, but what can you do? I don't know if your device will be able to replace the U-boot, Since it doesn't have a security key, it's probably possible to replace the uboot on your router
We'll have to replace the U-boot, as is the case on all devices. This is bad, but what can you do? I don't know if your device will be able to replace the U-boot, Since it doesn't have a security key, it's probably possible to replace the uboot on your router
We'll have to replace the U-boot, as is the case on all devices. This is bad, but what can you do? I don't know if your device will be able to replace the U-boot, Since it doesn't have a security key, it's probably possible to replace the uboot on your router
What is missing from the vendor uboot ?
> We'll have to replace the U-boot, as is the case on all devices. This is bad, but what can you do? I don't know if your device will be able to replace the U-boot, Since it doesn't have a security key, it's probably possible to replace the uboot on your router
What is missing from the vendor uboot ?
had some errors with ubi but think that is because it is not for this device
I will replace the uboot from the xx230v, xx530v, and ex530v with the mainstream uboot, i just need to change a few things, what's holding me back right now is my Mikrotik with the nor and nand with spi driver
> had some errors with ubi but think that is because it is not for this device
I will replace the uboot from the xx230v, xx530v, and ex530v with the mainstream uboot, i just need to change a few things, what's holding me back right now is my Mikrotik with the `nor` and `nand` with spi driver
Is there even nor support? I know the old stuff has nor support but not the new. Wont we need a driver for it?
A small change in the dts and NOR is detected, but not NAND. In the new Airoha driver, they are supporting NOR at the same time, this can be seen in the Mikrotik dts
> Is there even nor support? I know the old stuff has nor support but not the new. Wont we need a driver for it?
A small change in the dts and NOR is detected, but not NAND. In the new Airoha driver, they are supporting NOR at the same time, this can be seen in the Mikrotik dts
For it to work well you need the proper CS selected via pinmux. That might not have the needed plumbing.
I tried using NFi with spi_ctrl, but it either timed out or returned a zered ID for nand:
[ 1.279611] spi-nor spi0.0: probe with driver spi-nor failed with error -110
[ 1.537006] spi-nand spi0.1: probe with driver spi-nand failed with error -110
[ 18.796241] 3 fixed-partitions partitions found on MTD device spi0.0
[ 18.802623] Creating 3 MTD partitions on "spi0.0":
[ 18.807444] 0x000000000000-0x000000000800 : "preload"
[ 18.813321] 0x000000000800-0x0000000fe000 : "u-boot"
[ 18.818452] mtd: partition "u-boot" doesn't start on an erase/write block boundary -- force read-only
[ 18.828706] 0x0000000fe000-0x000000100000 : "factory"
[ 18.836563] spi-nand spi0.1: unknown raw ID 0000000000
[ 18.841727] spi-nand spi0.1: probe with driver spi-nand failed with error -95
> For it to work well you need the proper CS selected via pinmux. That might not have the needed plumbing.
I tried using NFi with spi_ctrl, but it either timed out or returned a zered ID for nand:
```
[ 1.279611] spi-nor spi0.0: probe with driver spi-nor failed with error -110
[ 1.537006] spi-nand spi0.1: probe with driver spi-nand failed with error -110
```
```
[ 18.796241] 3 fixed-partitions partitions found on MTD device spi0.0
[ 18.802623] Creating 3 MTD partitions on "spi0.0":
[ 18.807444] 0x000000000000-0x000000000800 : "preload"
[ 18.813321] 0x000000000800-0x0000000fe000 : "u-boot"
[ 18.818452] mtd: partition "u-boot" doesn't start on an erase/write block boundary -- force read-only
[ 18.828706] 0x0000000fe000-0x000000100000 : "factory"
[ 18.836563] spi-nand spi0.1: unknown raw ID 0000000000
[ 18.841727] spi-nand spi0.1: probe with driver spi-nand failed with error -95
```
We'll have to replace the U-boot, as is the case on all devices. This is bad, but what can you do? I don't know if your device will be able to replace the U-boot, Since it doesn't have a security key, it's probably possible to replace the uboot on your router
What is missing from the vendor uboot ?
Well iopsys u-boot version 4004 dies loading initramfs with "lzma compressed: uncompress error 7" (it loads just a kernel fine so think it is due to size), the one @Sirherobrine23 sent boots the very same image just fine.
> > We'll have to replace the U-boot, as is the case on all devices. This is bad, but what can you do? I don't know if your device will be able to replace the U-boot, Since it doesn't have a security key, it's probably possible to replace the uboot on your router
>
> What is missing from the vendor uboot ?
Well iopsys u-boot version 4004 dies loading initramfs with "lzma compressed: uncompress error 7" (it loads just a kernel fine so think it is due to size), the one @Sirherobrine23 sent boots the very same image just fine.
Thanks, will test it. Just ordered another 10 (so there are many available now to brick :)), iam pretty sure that they sent theese units to be destroyed in Romania and they ended up on Vinted instead for some reason. Sold as Genexis but all Heimgard units before acquisition. Genexis as far as i know sells only to isps and those isps pretty sure lock those boxes up (lamely mostly :) ).
So i would call them Heimgard hardware as no idea if there was any hardware changes by Genexis (do not think so) since they acquired them.
> I will add this router to u-boot
Thanks, will test it. Just ordered another 10 (so there are many available now to brick :)), iam pretty sure that they sent theese units to be destroyed in Romania and they ended up on Vinted instead for some reason. Sold as Genexis but all Heimgard units before acquisition. Genexis as far as i know sells only to isps and those isps pretty sure lock those boxes up (lamely mostly :) ).
So i would call them Heimgard hardware as no idea if there was any hardware changes by Genexis (do not think so) since they acquired them.
Honestly I dont know. But the devices should have the recovery option available by pulling gpio0 down, so writing something new to the device should be possible. One must not wipe the art partition though because the calibration data is stored there for the wifi cards.
Honestly I dont know. But the devices should have the recovery option available by pulling gpio0 down, so writing something new to the device should be possible. One must not wipe the art partition though because the calibration data is stored there for the wifi cards.
Heimgard CX30 is non, does not have usb ports, it has usb-c for power (did not tested but doubt if it can be use via hub as an usb port too), CX20 do not have it eigher, HX30/40 and laxy has usb port i think they are 2.0 from:
> @Gemnon this device is USB2 or USB3?
Heimgard CX30 is non, does not have usb ports, it has usb-c for power (did not tested but doubt if it can be use via hub as an usb port too), CX20 do not have it eigher, HX30/40 and laxy has usb port i think they are 2.0 from:
https://dk-store.netceed.com/upload_dir/shop/75000.pdf
and
https://dk-store.netceed.com/upload_dir/shop/75001.pdf
@Gemnon, the xx230v's uboot chainload, you have tftpput which you can use to perform a full NAND dump, in case you need to recover it
read with mtd and send to tftp server with tftput
@Gemnon, the xx230v's uboot chainload, you have `tftpput` which you can use to perform a full NAND dump, in case you need to recover it
read with `mtd` and send to tftp server with `tftput`
@Gemnon, the xx230v's uboot chainload, you have tftpput which you can use to perform a full NAND dump, in case you need to recover it
read with mtd and send to tftp server with tftput
Thanks did that already on CX30, iopsys u-boot also has tftpput so have full backup for the cx30 iam testing on.
Will get the same for the other boxes.
Need that file? (can upload it tommorow, as it is on my laptop i left at work)
> @Gemnon, the xx230v's uboot chainload, you have `tftpput` which you can use to perform a full NAND dump, in case you need to recover it
>
> read with `mtd` and send to tftp server with `tftput`
Thanks did that already on CX30, iopsys u-boot also has tftpput so have full backup for the cx30 iam testing on.
Will get the same for the other boxes.
Need that file? (can upload it tommorow, as it is on my laptop i left at work)
if you want to send it, I can locate the mt76 EEPROM in the ART file, because that UBI dump you sent is strange
I might be wrong but what fixed my boot was copying over /lib/firmware from iowrt.
Is it not /lib/firmware/mediatek/mt7916_eeprom.bin in rootfs ? I attached the stuff extracted on ota.
Not sure if it is but copying that folder to ramfs seems to made it recognize wifi. (also i think all combo is there not just the one for pixly (under /lib/firmware/mediatek/airoha) but i might be wrong)
> if you want to send it, I can locate the mt76 EEPROM in the ART file, because that UBI dump you sent is strange
I might be wrong but what fixed my boot was copying over /lib/firmware from iowrt.
Is it not /lib/firmware/mediatek/mt7916_eeprom.bin in rootfs ? I attached the stuff extracted on ota.
Not sure if it is but copying that folder to ramfs seems to made it recognize wifi. (also i think all combo is there not just the one for pixly (under /lib/firmware/mediatek/airoha) but i might be wrong)
I might be wrong but what fixed my boot was copying over /lib/firmware from iowrt.
Is /lib/firmware/mediatek/mt7916_eeprom.bin in rootfs
Normally, the EEPROM is located in the ART part and in the rootfs. if it's not in the ART part, we should use the one in the rootfs, but we always prefer the ART part
> I might be wrong but what fixed my boot was copying over /lib/firmware from iowrt.
> Is /lib/firmware/mediatek/mt7916_eeprom.bin in rootfs
Normally, the EEPROM is located in the ART part and in the rootfs. if it's not in the ART part, we should use the one in the rootfs, but we always prefer the ART part
Will send it tommorow.
The dump i found home is what i made from luci in heimgardos, and there art partition and boarddata is just full of ff , most likely luci could not back it off (or was protected) but it should be in nand backup.
Will send it tommorow.
The dump i found home is what i made from luci in heimgardos, and there art partition and boarddata is just full of ff , most likely luci could not back it off (or was protected) but it should be in nand backup.
fw_printenv shows those but not as hex but plaintext.
base_macaddr=B0DD74AF93E0
ethaddr=B0:DD:74:AF:93:E0
That is why i think it is taken from there.
> nothing in any part of nand dump
>
> ```
> ╭─sirherobrine23@matheus-note ~/gitwork/airoha_en7523/dumps/gene
> ╰─$ dd status=none if=cx30_dump.bin | xxd -c32 | grep 'b0dd 74af 93e0' 1 ↵
> ╭─sirherobrine23@matheus-note ~/gitwork/airoha_en7523/dumps/gene
> ╰─$ 1 ↵
> ```
fw_printenv shows those but not as hex but plaintext.
base_macaddr=B0DD74AF93E0
ethaddr=B0:DD:74:AF:93:E0
That is why i think it is taken from there.
Will test on cx30, should I try to flash that uboot, or load it from uboot and then initramfs and then use sysupgrade image?
> my last build with changes 76ea660aef
Will test on cx30, should I try to flash that uboot, or load it from uboot and then initramfs and then use sysupgrade image?
I'll have to revert the Ethernet driver to see why I'm getting a kernel panic on Ethernet, as this is only happening with the en7529 SoC; it's working normally with the en7562
I'll have to revert the Ethernet driver to see why I'm getting a kernel panic on Ethernet, as this is only happening with the en7529 SoC; it's working normally with the en7562
> Regarding the backups you're making, using chainload to back up the NAND probably won't give you the same error as with the vendor u-boot
It gives the very same error on laxy, and mtd list incorrect but that might be because it is not uboot for this.
```
List of MTD devices:
* spi-nand0
- device: nand@0
- parent: spi@1fa10000
- driver: spi_nand
- path: /spi@1fa10000/nand@0
- type: NAND flash
- block size: 0x20000 bytes
- min I/O: 0x800 bytes
- OOB size: 128 bytes
- OOB available: 56 bytes
- ECC strength: 8 bits
- ECC step size: 512 bytes
- bitflip threshold: 6 bits
- 0x000000000000-0x000010000000 : "spi-nand0"
- 0x000000000000-0x000000080000 : "uboot-atf"
- 0x000000080000-0x000000120000 : "romfile"
- 0x0000000a0000-0x000000240000 : "misc"
- 0x0000001a0000-0x0000002a0000 : "chainloader"
- 0x0000002a0000-0x000007500000 : "ubi"
- 0x000007500000-0x000008000000 : "reserved_bmt"
U-Boot> mtd read spi-nand0 $loadaddr 0x0 0x10000000
Reading 268435456 byte(s) (131072 page(s)) at offset 0x00000000
WARNING at drivers/mtd/nand/spi/core.c:144/spinand_select_target()!
Failure while reading at offset 0x10000000
Read on spi-nand0 failed with error -22
```
It should have (think all have 256Mb size, and 512 RAM except c20 as that have half of RAM that is why i had to backup it half) and it only not reads the last 2MB maybe error on NAND (sadly only have one laxy box so cant test it on another)
U-Boot 2023.07.02-OpenWrt-r23630-842932a63d-00059-g88f17672a-dirty (Sep 10 2024 - 12:34:36 +0200)
CPU: Airoha EN7529DT
DRAM: 496 MiB
Core: 7 devices, 6 uclasses, devicetree: fit
spi-nand: spi_nand spinand@0: Winbond SPI NAND was found.
spi-nand: spi_nand spinand@0: 256 MiB, block size: 128 KiB, page size: 2048, OOB size: 128
+ 0x000000000000-0x000010000000 : "spi-nand0"
+ 0x000000000000-0x000000100000 : "u-boot"
+ 0x000000100000-0x000010000000 : "ubi"
In: serial@1fbf0000
Out: serial@1fbf0000
Err: serial@1fbf0000
Net: eth0: en75xx-eth
Loading Environment from UBI... ubi0: default fastmap pool size: 100
ubi0: default fastmap WL pool size: 50
ubi0: attaching mtd2
ubi0: scanning is finished
ubi0: attached mtd2 (name "ubi", size 255 MiB)
ubi0: PEB size: 131072 bytes (128 KiB), LEB size: 126976 bytes
ubi0: min./max. I/O unit sizes: 2048/2048, sub-page size 2048
ubi0: VID header offset: 2048 (aligned 2048), data offset: 4096
ubi0: good PEBs: 2022, bad PEBs: 18, corrupted PEBs: 0
ubi0: user volume: 11, internal volumes: 1, max. volumes count: 128
ubi0: max/mean erase counter: 3/1, WL threshold: 4096, image sequence number: 1985439503
ubi0: available PEBs: 14, total reserved PEBs: 2008, PEBs reserved for bad PEB handling: 22
Read 126976 bytes from volume env1 to 9eaf6280
Read 126976 bytes from volume env2 to 9eb152c0
> Is your NAND have 256Mb of size?
It should have (think all have 256Mb size, and 512 RAM except c20 as that have half of RAM that is why i had to backup it half) and it only not reads the last 2MB maybe error on NAND (sadly only have one laxy box so cant test it on another)
```
U-Boot 2023.07.02-OpenWrt-r23630-842932a63d-00059-g88f17672a-dirty (Sep 10 2024 - 12:34:36 +0200)
CPU: Airoha EN7529DT
DRAM: 496 MiB
Core: 7 devices, 6 uclasses, devicetree: fit
spi-nand: spi_nand spinand@0: Winbond SPI NAND was found.
spi-nand: spi_nand spinand@0: 256 MiB, block size: 128 KiB, page size: 2048, OOB size: 128
+ 0x000000000000-0x000010000000 : "spi-nand0"
+ 0x000000000000-0x000000100000 : "u-boot"
+ 0x000000100000-0x000010000000 : "ubi"
In: serial@1fbf0000
Out: serial@1fbf0000
Err: serial@1fbf0000
Net: eth0: en75xx-eth
Loading Environment from UBI... ubi0: default fastmap pool size: 100
ubi0: default fastmap WL pool size: 50
ubi0: attaching mtd2
ubi0: scanning is finished
ubi0: attached mtd2 (name "ubi", size 255 MiB)
ubi0: PEB size: 131072 bytes (128 KiB), LEB size: 126976 bytes
ubi0: min./max. I/O unit sizes: 2048/2048, sub-page size 2048
ubi0: VID header offset: 2048 (aligned 2048), data offset: 4096
ubi0: good PEBs: 2022, bad PEBs: 18, corrupted PEBs: 0
ubi0: user volume: 11, internal volumes: 1, max. volumes count: 128
ubi0: max/mean erase counter: 3/1, WL threshold: 4096, image sequence number: 1985439503
ubi0: available PEBs: 14, total reserved PEBs: 2008, PEBs reserved for bad PEB handling: 22
Read 126976 bytes from volume env1 to 9eaf6280
Read 126976 bytes from volume env2 to 9eb152c0
```
Meanwhile let a box on the internet to test the Heimgard app and it worked (after a few restart), noticed ota url changed to hmg-ota.genexis.eu but they have the very same last version for theese as Heimgard. (because hmg most likely means Heimgard anyway :))
New release still locks up same place.
Meanwhile let a box on the internet to test the Heimgard app and it worked (after a few restart), noticed ota url changed to hmg-ota.genexis.eu but they have the very same last version for theese as Heimgard. (because hmg most likely means Heimgard anyway :))
In my xx230v it was the same thing, I must have reverted to some problematic commit I was having last month, I'll have to check my local Git repository to find the working driver, since I got this one from the server
In my xx230v it was the same thing, I must have reverted to some problematic commit I was having last month, I'll have to check my local Git repository to find the working driver, since I got this one from the server
Regarding the MT76, we must be having some problem with the EEPROM load because if it's a raw MTD it can initialize correctly, but when it's in the UBI it can't load. I must be doing something wrong
Regarding the MT76, we must be having some problem with the EEPROM load because if it's a raw MTD it can initialize correctly, but when it's in the UBI it can't load. I must be doing something wrong
The Ethernet driver hasn't changed at all in the latest commits 🤔, But only the SPI driver was working, so I'm not really sure what it is. Did you test it with the latest build from the automatic build releases?
The Ethernet driver hasn't changed at all in the latest commits 🤔, But only the SPI driver was working, so I'm not really sure what it is. Did you test it with the latest build from the automatic build releases?
I'm migrating my email server, so I won't be able to check it tomorrow, probably today i make solution for this, but my xx230v and Mikrotik not having this problem in Ethernet driver
I'm migrating my email server, so I won't be able to check it tomorrow, probably today i make solution for this, but my xx230v and Mikrotik not having this problem in Ethernet driver
The Ethernet driver hasn't changed at all in the latest commits 🤔, But only the SPI driver was working, so I'm not really sure what it is. Did you test it with the latest build from the automatic build releases?
> The Ethernet driver hasn't changed at all in the latest commits 🤔, But only the SPI driver was working, so I'm not really sure what it is. Did you test it with the latest build from the automatic build releases?
Used the one from:
https://sirherobrine23.com.br/airoha_an7523/builds/releases/download/v59/openwrt-airoha-an7523-genexis_cx30-initramfs-kernel.bin
Just redownloaded to be sure but same error. :)
If v58 or v54 works, it's probably due to some change I made in the DTS file when I renamed it from en7523 to an7523, so I'll have to look into that.
Thanks, will check on the older ones, is it not possible that the cause it is bigger than mine (i stripped down the first test to 8m because vendor uboot unpack) and while cross uboot can extract it but as it is not protecting reserved memory blocks (i think stock uboot fails extracting because of it) later it cause issues?
> If v58 or v54 works, it's probably due to some change I made in the DTS file when I renamed it from en7523 to an7523, so I'll have to look into that.
Thanks, will check on the older ones, is it not possible that the cause it is bigger than mine (i stripped down the first test to 8m because vendor uboot unpack) and while cross uboot can extract it but as it is not protecting reserved memory blocks (i think stock uboot fails extracting because of it) later it cause issues?
If you're writing to the NAND, the initramfs file doesn't alter anything in the NAND, it only reads the MAC addresses and nothing else, because everything is literally in RAM
It would be a problem if you were to record on the NAND because you would have to do some things in ubi for write Openwrt with new ubi partitions mapping
If you're writing to the NAND, the initramfs file doesn't alter anything in the NAND, it only reads the MAC addresses and nothing else, because everything is literally in RAM
It would be a problem if you were to record on the NAND because you would have to do some things in ubi for write Openwrt with new ubi partitions mapping
If you're writing to the NAND, the initramfs file doesn't alter anything in the NAND, it only reads the MAC addresses and nothing else, because everything is literally in RAM
It would be a problem if you were to record on the NAND because you would have to do some things in ubi for write Openwrt with new ubi partitions mapping
Iam just booting from tftpboot right now via chainloader, what i compiled from v51 source (april 18th) back then boots, but v51 from the releases page locks up. Also stock uboot capable to but sysupgrade images kernel (and locks up the very same place), so not get to the point to find out if there is no rootfs. (so it must be size its issue)
> If you're writing to the NAND, the initramfs file doesn't alter anything in the NAND, it only reads the MAC addresses and nothing else, because everything is literally in RAM
>
> It would be a problem if you were to record on the NAND because you would have to do some things in ubi for write Openwrt with new ubi partitions mapping
Iam just booting from tftpboot right now via chainloader, what i compiled from ~~v51~~ source (april 18th) back then boots, but v51 from the releases page locks up. Also stock uboot capable to but sysupgrade images kernel (and locks up the very same place), so not get to the point to find out if there is no rootfs. (so it must be size its issue)
This is only happening with the cx30, is laxy having the same problem? because this only happened to me once when I was messing with the SPI where the entire boot system would hang, worrying about the mac address of the ram or flash memory.
This is only happening with the cx30, is laxy having the same problem? because this only happened to me once when I was messing with the SPI where the entire boot system would hang, worrying about the mac address of the ram or flash memory.
You let it boot completely because it was still loading the kernel modules. Try again and wait a bit, as the drivers loaded directly
If it loaded, it's because the MAC address was being mapped by uboot env
You let it boot completely because it was still loading the kernel modules. Try again and wait a bit, as the drivers loaded directly
If it loaded, it's because the MAC address was being mapped by uboot env
You let it boot completely because it was still loading the kernel modules. Try again and wait a bit, as the drivers loaded directly
If it loaded, it's because the MAC address was being mapped by uboot env
I did, some reason eth0 gets into bridge. But i have lan1, lan2 and wan , once i added those to bridge and removed eth0 ping works so ethernet works .If i remember good in dsa that is how it should be, but i might be wrong. But you right it is working once i get the interfaces to bridge.
> You let it boot completely because it was still loading the kernel modules. Try again and wait a bit, as the drivers loaded directly
>
> If it loaded, it's because the MAC address was being mapped by uboot env
I did, some reason eth0 gets into bridge. But i have lan1, lan2 and wan , once i added those to bridge and removed eth0 ping works so ethernet works .If i remember good in dsa that is how it should be, but i might be wrong. But you right it is working once i get the interfaces to bridge.
CX30 works fine with v60 , will test other devices and figure out how to get to flash. Genexis/iopsys image is fit in fit, in the fit there is an "uncompressed" boot fit image that contains kernel and dtb and gets imxtracted to kernel partition, and then there is rootfs what it imxtract to rootfs. Openwrt sysupgrade image is all in one.
CX30 works fine with v60 , will test other devices and figure out how to get to flash. Genexis/iopsys image is fit in fit, in the fit there is an "uncompressed" boot fit image that contains kernel and dtb and gets imxtracted to kernel partition, and then there is rootfs what it imxtract to rootfs. Openwrt sysupgrade image is all in one.
The base MAC address is the same for Ethernet? Because if so, I'll add the script to correct this for the Genexis devices, but for now the devices will use the 2.4GHz wireless address for the network switch until I add the script if necessary
The base MAC address is the same for Ethernet? Because if so, I'll add the script to correct this for the Genexis devices, but for now the devices will use the 2.4GHz wireless address for the network switch until I add the script if necessary
The base MAC address is the same for Ethernet? Because if so, I'll add the script to correct this for the Genexis devices, but for now the devices will use the 2.4GHz wireless address for the network switch until I add the script if necessary
Nope they actually different even wifi mac is. And it is the same with openwrt on zephyr and pixly (i mean they get the very same mac), there is a base_macaddr environment variable what they use for ethernet in AABBCCDDEEFF format. Also on zephyr 2.5gbit ports not work yet not even sure there is driver for it right now :) Attaching logs, orig is the original iowrt boot, openwrt is v60. Going to check the other device in a bit.
For rodimus ethernet seems to be broken there is lan1 in openwrt but non of the ports seem to work. Also rodimus and zephyr have zigbee thing on usb (not even used in stock firmware so no idea why it is there) and never used zigbee so not even know what it is good for.
:
Arcee has a phantom wan (it has just two lan ports :)) but otherwise seem to work just mac address have nothing to do with the devices (like the others)
Laxy there is lan1,lan2,lan3 (it has 4 ports and pon) could not get link on any but that boots too.
Added a few pics from the back of the boxes ,for port numbers (rodimus-hx30) have 3 buttons on the side (Wlan/WPS/reset).
> The base MAC address is the same for Ethernet? Because if so, I'll add the script to correct this for the Genexis devices, but for now the devices will use the 2.4GHz wireless address for the network switch until I add the script if necessary
Nope they actually different even wifi mac is. And it is the same with openwrt on zephyr and pixly (i mean they get the very same mac), there is a base_macaddr environment variable what they use for ethernet in AABBCCDDEEFF format. Also on zephyr 2.5gbit ports not work yet not even sure there is driver for it right now :) Attaching logs, orig is the original iowrt boot, openwrt is v60. Going to check the other device in a bit.
For rodimus ethernet seems to be broken there is lan1 in openwrt but non of the ports seem to work. Also rodimus and zephyr have zigbee thing on usb (not even used in stock firmware so no idea why it is there) and never used zigbee so not even know what it is good for.
:
```
echo -e "version\r" | socat -T 2 - /dev/ttyUSB0,b115200,ra
w,echo=0
stack ver. [6.10.2 GA build 267]
211018_xG21_Z3S_V5V61020>
```
Arcee has a phantom wan (it has just two lan ports :)) but otherwise seem to work just mac address have nothing to do with the devices (like the others)
Laxy there is lan1,lan2,lan3 (it has 4 ports and pon) could not get link on any but that boots too.
Added a few pics from the back of the boxes ,for port numbers (rodimus-hx30) have 3 buttons on the side (Wlan/WPS/reset).
I’d like to ask you to provide a short list so I can create the dts files correctly:
Number of Ethernet ports, along with the port names
Whether any external PHYs are used
USB ports and their type (USB 2.0 or 3.0)
MAC address offset or a dump of each device and the MAC addresses so I can locate each one
In addition to the devices that use/have an xPON port
Also rodimus and zephyr have zigbee thing on usb (not even used in stock firmware so no idea why it is there) and never used zigbee so not even know what it is good for.
ZigBee is used to create a network to control IoT devices without relying on Wi-Fi to create that network. If I'm not mistaken, the W1700k also has a ZigBee component, which it seems wasn't used either
there is a base_macaddr environment variable what they use for ethernet in AABBCCDDEEFF format.
What was causing the chaotic startup was me trying to use the MAC address as the stored base for the u-boot environment, I think something is missing in the kernel, that's why OpenWRT wasn't starting directly, I'll have to create a script to correct the addresses that are extracted from the u-boot environment later
I’d like to ask you to provide a short list so I can create the dts files correctly:
- Number of Ethernet ports, along with the port names
- Whether any external PHYs are used
- USB ports and their type (USB 2.0 or 3.0)
- MAC address offset or a dump of each device and the MAC addresses so I can locate each one
- In addition to the devices that use/have an xPON port
> Also rodimus and zephyr have zigbee thing on usb (not even used in stock firmware so no idea why it is there) and never used zigbee so not even know what it is good for.
ZigBee is used to create a network to control IoT devices without relying on Wi-Fi to create that network. If I'm not mistaken, the W1700k also has a ZigBee component, which it seems wasn't used either
> there is a base_macaddr environment variable what they use for ethernet in AABBCCDDEEFF format.
What was causing the chaotic startup was me trying to use the MAC address as the stored base for the u-boot environment, I think something is missing in the kernel, that's why OpenWRT wasn't starting directly, I'll have to create a script to correct the addresses that are extracted from the u-boot environment later
> MAC address offset or a dump of each device and the MAC addresses so I can locate each one
If you are going to submit it, please send it to this repository: https://sirherobrine23.com.br/airoha_en7523/off-topics-private/issues
I’d like to ask you to provide a short list so I can create the dts files correctly:
Number of Ethernet ports, along with the port names
Whether any external PHYs are used
USB ports and their type (USB 2.0 or 3.0)
MAC address offset or a dump of each device and the MAC addresses so I can locate each one
In addition to the devices that use/have an xPON port
laxy: 4x 1gbit ports lan1,lan2,lan3,lan4, has 1xUSB 3.0 (at least iowrt shown 5000 speed when connected my usb drive to it and it was recognized on usb2) it has xPON port, also have 2 fxs port (i do not think there is external PHYs)
rodimus: 4x 1gbit ports lan1,lan2,lan3,lan4, 1 2.5gbit wan port wan, has 1xUSB 2.0 (device got recognized on usb1) and 1 fxs port, no xPON. (i do not think there is external PHYs)
zephyr: 3x1gbit port ( lan1, lan2, lan3), 2x 2.5gbit port (lan4,wan) has 1xUSB 2.0 (device got recognized on usb1), 1 fxs port and i think lan4 is an external phy:
arcee: 2x1gbit port (lan1,lan2), no USB
pixly: 3x1gbit port (lan1,lan2,wan) no USB
there is a base_macaddr environment variable what they use for ethernet in AABBCCDDEEFF format.
What was causing the chaotic startup was me trying to use the MAC address as the stored base for the u-boot environment, I think something is missing in the kernel, that's why OpenWRT wasn't starting directly, I'll have to create a script to correct the addresses that are extracted from the u-boot environment later
Will try to find those, but as we chainload foreign uboot to be able to boot initramfs kernel wont get envs from that? :)
> I’d like to ask you to provide a short list so I can create the dts files correctly:
>
> - Number of Ethernet ports, along with the port names
> - Whether any external PHYs are used
> - USB ports and their type (USB 2.0 or 3.0)
> - MAC address offset or a dump of each device and the MAC addresses so I can locate each one
> - In addition to the devices that use/have an xPON port
laxy: 4x 1gbit ports lan1,lan2,lan3,lan4, has 1xUSB 3.0 (at least iowrt shown 5000 speed when connected my usb drive to it and it was recognized on usb2) it has xPON port, also have 2 fxs port (i do not think there is external PHYs)
rodimus: 4x 1gbit ports lan1,lan2,lan3,lan4, 1 2.5gbit wan port wan, has 1xUSB 2.0 (device got recognized on usb1) and 1 fxs port, no xPON. (i do not think there is external PHYs)
zephyr: 3x1gbit port ( lan1, lan2, lan3), 2x 2.5gbit port (lan4,wan) has 1xUSB 2.0 (device got recognized on usb1), 1 fxs port and i think lan4 is an external phy:
arcee: 2x1gbit port (lan1,lan2), no USB
pixly: 3x1gbit port (lan1,lan2,wan) no USB
> > there is a base_macaddr environment variable what they use for ethernet in AABBCCDDEEFF format.
>
> What was causing the chaotic startup was me trying to use the MAC address as the stored base for the u-boot environment, I think something is missing in the kernel, that's why OpenWRT wasn't starting directly, I'll have to create a script to correct the addresses that are extracted from the u-boot environment later
Will try to find those, but as we chainload foreign uboot to be able to boot initramfs kernel wont get envs from that? :)
Sorry still going to test new builds just previous weeks was busy , could get iowrt stuff to at least build packages I need for the devices. But still would prefer openwrt. Got some more boxes. Well can be sure CX20 is arcee and all this stuff are engineering boxes.
Sorry still going to test new builds just previous weeks was busy , could get iowrt stuff to at least build packages I need for the devices. But still would prefer openwrt. Got some more boxes. Well can be sure CX20 is arcee and all this stuff are engineering boxes.
Sorry if we don't pay much attention to some things, I'm fixing some drivers, and others are probably busy with other things.
Well can be sure CX20 is arcee and all this stuff are engineering boxes.
Interesting, if the PCB has any revisions it will be interesting, some devices have undergone minor changes on my part, but most people in this organization can open PRs and/or commit, if they want to send something to a repository.
Sorry if we don't pay much attention to some things, I'm fixing some drivers, and others are probably busy with other things.
> Well can be sure CX20 is arcee and all this stuff are engineering boxes.
Interesting, if the PCB has any revisions it will be interesting, some devices have undergone minor changes on my part, but most people in this organization can open PRs and/or commit, if they want to send something to a repository.
Sorry if we don't pay much attention to some things, I'm fixing some drivers, and others are probably busy with other things.
Well can be sure CX20 is arcee and all this stuff are engineering boxes.
Interesting, if the PCB has any revisions it will be interesting, some devices have undergone minor changes on my part, but most people in this organization can open PRs and/or commit, if they want to send something to a repository.
No worries iam glad what you do! Iam not much a developer sadly, just removed their call home stuff from rootfs and made sure on reset boxes not cause havoc, can't even build iowrt as device profiles needs GitHub login :) Not sure how GPL compatible is this. But then even could not find Heimgard repo or at least the last source tarball before it went Genexis :)
> Sorry if we don't pay much attention to some things, I'm fixing some drivers, and others are probably busy with other things.
>
>
> > Well can be sure CX20 is arcee and all this stuff are engineering boxes.
>
> Interesting, if the PCB has any revisions it will be interesting, some devices have undergone minor changes on my part, but most people in this organization can open PRs and/or commit, if they want to send something to a repository.
No worries iam glad what you do! Iam not much a developer sadly, just removed their call home stuff from rootfs and made sure on reset boxes not cause havoc, can't even build iowrt as device profiles needs GitHub login :) Not sure how GPL compatible is this. But then even could not find Heimgard repo or at least the last source tarball before it went Genexis :)
Tested the latest build on pixly mac addresses good now and both ethernet/wifi working. :)
Have a lot of lines of this though (but think that just some debug info)
[ 1051.971303] airoha-spi 1fa10000.spi: exec_op cs=0 opcode=13 tx=1 addr=3 dummy=0 data=0 dir=0
Tested the latest build on pixly mac addresses good now and both ethernet/wifi working. :)
Have a lot of lines of this though (but think that just some debug info)
[ 1051.971303] airoha-spi 1fa10000.spi: exec_op cs=0 opcode=13 tx=1 addr=3 dummy=0 data=0 dir=0
@Gemnon I added some changes for a few devices today (47fbe84c61)
I'm not sure which Ethernet chips some of these devices use (likely the same as the e650) so please wait for the build in github so I can update the dts, 01_leds, and 02_network.
Rodimus R1
Zephyr
@Gemnon I added some changes for a few devices today (47fbe84c61a2b471e42a8df55c45b3166c1cc845)
I'm not sure which Ethernet chips some of these devices use (likely the same as the e650) so please wait for the build in [github](https://github.com/Sirherobrine23/airoha_en7523_builds/actions/runs/28039480967) so I can update the dts, 01_leds, and 02_network.
- Rodimus R1
- Zephyr
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Is it really ? :)
That's what @Gemnon pulled and sent me, i'm going to trust him and add his dts to the branch so I can test it, he has both devices to test
and also manage to get the NPU firmware, if it works with the v2 driver, i'll use it for testing, and it's already in the NPU repository
Compared to other devices there are lots of diffs in the dts. The device uses alot of different drivers compared to the SDK.
he also sent me some firmware that will probably let me extract that dts from uboot
Device: Heimgard Pixlyto Device: Heimgard/Genexis CX30ubi names:
env1env2boot1boot2rootfs1rootfs2overlay1overlay2artboarddatacontainerCant be sure extracted it from the firmware. And the boxes i have behave strange there are CX30-s that try to get ip address via dhcp, others that listens on 192.168.39.1, some boxes (HX30, HX40) listen on 192.168.1.1, some of the same on 192.168.5.1. Some not even have model number, still trying to figure out the things (but which ones i tried just fails with decompression error when i try to boot any initramfs image from here). Iam pretty sure they used to be development boxes as some had github version of iowrt thing on them.
Also by default ssh password auth disabled, i can enable it, some boxes let me change the burnt in root password (which is the same from the hash on all boxes but i have no idea what) via luci (and for some reason it only change ssh root password not the device password), for some luci just gives an error and have to backup config, alter /etc/shadow reload config for access. Serial port can be accessed on all via holes on the back of the boxes, uboot has bootstopkeysha256 set to something (will have to brute-force it) but that can be overriden in env.
@Sirherobrine23
This one is the (Unknown Model) which i could not even get paired with a product at Genexis/Heimgard:
I0M0QCCXF (laxy) Have a pon, 2 fxs, 4 gigabit lan port, 256M flash, 512M memory, MT7916D
mtd layout looks like pretty much the same as other ones:
dev: size erasesize name
mtd0: 00100000 00020000 "u-boot"
mtd1: 0ff00000 00020000 "ubi"
mtd2: 0001f000 0001f000 "env1"
mtd3: 0001f000 0001f000 "env2"
mtd4: 00383000 0001f000 "boot1"
mtd5: 00383000 0001f000 "boot2"
mtd6: 01b9c000 0001f000 "rootfs1"
mtd7: 021e8000 0001f000 "rootfs2"
mtd8: 0141a000 0001f000 "overlay1"
mtd9: 0141a000 0001f000 "overlay2"
mtd10: 00193000 0001f000 "art"
mtd11: 0041e000 0001f000 "boarddata"
mtd12: 07d17000 0001f000 "container"
So the listing what i have with the corresponding ota images i could find:
CX20: G0A0ACKW0 (arcee):
https://firmware.homecontrol-cloud.com/heimgard/G0A0ACKW0-HOv1/v1-heimgard/host/arcee-7.3.1.17.2.0/last.itb
CX30: I0A0NCKX0 (pixly)
https://firmware.homecontrol-cloud.com/heimgard/I0A0NCKX0-HOv1/v1-heimgard/host/pixly-7.3.1.17.2.0/last.itb
"Unknown model": I0M0QCCXF (laxy)
https://firmware.homecontrol-cloud.com/heimgard/I0M0QCCXF-HOv1/v1-heimgard/host/laxy-7.3.1.17.2.0/last.itb
HX30: I0ABGCUFD (zephyr)
https://firmware.homecontrol-cloud.com/heimgard/I0ABGCUFD-HOv1/v1-heimgard/host/zephyr-7.3.1.17.2.0/last.itb
HX40: G0A0GCUXD (rodimus)
https://firmware.homecontrol-cloud.com/heimgard/G0A0GCUXD-HOv1/v1-heimgard/host/rodimus-7.3.1.17.2.0/last.itb
All have Heimgard logo and using heimgard ota so i think they are from the time before Genexis acquired them...
@Gemnon could create a branch and add the devices. i currently working with the u-boot for Mikrotik and the xx230v/xx530v and i need to review the merbanan changes for the WED and NPU. You have code write, and for now it's easy to add the devices, look at these commits for example:
63edd2040219910446b7Allso adding some dmesg outputs the boxes i have with me now, will attach the other two later if it helps.
Device: Heimgard/Genexis CX30to Device: Genexis devicestry this build @Gemnon
@Sirherobrine23
Same result:
Also when i try to extract it with imxtract it sais:
Not sure if it has anything to do with uncompress error 7 though (it does with the stock firmware image too so that has nothing to do with it for sure)
Pretty sure the reason is the size.
It can extract and load lzma heimgard_cx30-kernel.bin, well until "Starting linux..." where it locks up, so lzma works in u-boot, just initramfs image too big for it.
Was changed load address (0x80208000) like it is on other boxes recompiled. And i have lines... And lockup at npu thing (well no initramfs so no wonder it cant load it). That is a progress. At least it can boot stuff. :)
Think have to shrink down initramfs somehow, to be able to boot it. Or move to nand boot.
Tried to boot via seperate ramdisk, but failed most likely because i have no idea how to do it. (ram_try.log)
Tried to boot via seperate uncompressed ramdisk because i figured kernel not able to uncompress it, now it booted fully. (booted.log)
Ethernet works, after copying over /lib/firmware/mediatek from iowrt and reload mt7915e module wifi works too, at least shows up by iw.
I recently noticed that the NPU log isn't showing up, i'll look into that, but it should be working. Interestingly, the initramfs loads with less than 10Mb of size, if you're basing the build on my configuration, the files will exceed 22MB because it includes some testing tools and additional things for testings
The only things that need firmware are the NPU and the mt7916e, I couldn't find the eeprom for mt7916e in the volume art, but that might be my mistake in how i extracted the firmware from the dump you sent me
Well for me it seems it breaks on decompression because of its size, i think it hits a protected range (or there is some iowrt u-boot limit).
if i just upload the kernel to 0x9600000 and a cpio ramdisk image to 0x86000000 set bootargs to make root=/dev/ram0 and
bootm 0x96000000 0x86000000 it works (and not stuck on npu binary) if i try to boot the generated initramfs it fails with decompression error... But lzma working because lzmad kernel alone booting just fine.
Load a uboot chainload and try booting from new uboot
https://sirherobrine23.com.br/airoha_an7523/builds/releases/download/v49/openwrt-airoha-en7523-tplink_xx230v_v1-chainload-uboot.bin
Since the 230 has a header, just skip the 512B and it will start normally:
tftpboot 0x90000000 192.168.1.100:<230 chainload file> && bootm 0x90000200Yes, this one loads your initramfs image just fine (had some errors with ubi but think that is because it is not for this device).
We'll have to replace the U-boot, as is the case on all devices. This is bad, but what can you do? I don't know if your device will be able to replace the U-boot, Since it doesn't have a security key, it's probably possible to replace the uboot on your router
Did you modify the DTS file and the image/en7523.mk file? If so, could you send me the changes so I can merge
What is missing from the vendor uboot ?
I will replace the uboot from the xx230v, xx530v, and ex530v with the mainstream uboot, i just need to change a few things, what's holding me back right now is my Mikrotik with the
norandnandwith spi driverIs there even nor support? I know the old stuff has nor support but not the new. Wont we need a driver for it?
His uboot vendor is limiting the size of the initramfs boot
A small change in the dts and NOR is detected, but not NAND. In the new Airoha driver, they are supporting NOR at the same time, this can be seen in the Mikrotik dts
Ahh, yeah Nand and Nor protocols are ontop of the SPI driver. Forgot about that.
For it to work well you need the proper CS selected via pinmux. That might not have the needed plumbing.
pinmux in pinctrl?
I tried using NFi with spi_ctrl, but it either timed out or returned a zered ID for nand:
Move this discussion to airoha_an7523/openwrt#3 (comment)
Well iopsys u-boot version 4004 dies loading initramfs with "lzma compressed: uncompress error 7" (it loads just a kernel fine so think it is due to size), the one @Sirherobrine23 sent boots the very same image just fine.
I will add this router to u-boot
Thanks, will test it. Just ordered another 10 (so there are many available now to brick :)), iam pretty sure that they sent theese units to be destroyed in Romania and they ended up on Vinted instead for some reason. Sold as Genexis but all Heimgard units before acquisition. Genexis as far as i know sells only to isps and those isps pretty sure lock those boxes up (lamely mostly :) ).
So i would call them Heimgard hardware as no idea if there was any hardware changes by Genexis (do not think so) since they acquired them.
@merbanan u confirm have changes in devices hardware?
Honestly I dont know. But the devices should have the recovery option available by pulling gpio0 down, so writing something new to the device should be possible. One must not wipe the art partition though because the calibration data is stored there for the wifi cards.
@Gemnon this device is USB2 or USB3?
Heimgard CX30 is non, does not have usb ports, it has usb-c for power (did not tested but doubt if it can be use via hub as an usb port too), CX20 do not have it eigher, HX30/40 and laxy has usb port i think they are 2.0 from:
https://dk-store.netceed.com/upload_dir/shop/75000.pdf
and
https://dk-store.netceed.com/upload_dir/shop/75001.pdf
@Gemnon, the xx230v's uboot chainload, you have
tftpputwhich you can use to perform a full NAND dump, in case you need to recover itread with
mtdand send to tftp server withtftputreplace => 0x8000000 in mtd read with size of nand
Thanks did that already on CX30, iopsys u-boot also has tftpput so have full backup for the cx30 iam testing on.
Will get the same for the other boxes.
Need that file? (can upload it tommorow, as it is on my laptop i left at work)
if you want to send it, I can locate the mt76 EEPROM in the ART file, because that UBI dump you sent is strange
I might be wrong but what fixed my boot was copying over /lib/firmware from iowrt.
Is it not /lib/firmware/mediatek/mt7916_eeprom.bin in rootfs ? I attached the stuff extracted on ota.
Not sure if it is but copying that folder to ramfs seems to made it recognize wifi. (also i think all combo is there not just the one for pixly (under /lib/firmware/mediatek/airoha) but i might be wrong)
Normally, the EEPROM is located in the ART part and in the rootfs. if it's not in the ART part, we should use the one in the rootfs, but we always prefer the ART part
Will send it tommorow.
The dump i found home is what i made from luci in heimgardos, and there art partition and boarddata is just full of ff , most likely luci could not back it off (or was protected) but it should be in nand backup.
@Gemnon
mac addr: ?
mt76 eeprom (probably): art: off
0x000c0000, size:0x1000:Yes seem to be like that is it.
cx20 art is same addresses, i make dtsi for genexis devices
they are in the EEPROM, so there's a chance the Ethernet ports have different addresses.
000c0000:1679 0000-b0dd 74af 93e2-b0dd 74af 93e3-0000 0000 0000 0000 0000 0000 0000 0000 .y....t.....t...................ethernet is B0:DD:74:AF:93:E0
It might get it from u-boot env as there it is set :
base_macaddr=B0DD74AF93E0
ethaddr=B0:DD:74:AF:93:E0
nothing in any part of nand dump
fw_printenv shows those but not as hex but plaintext.
base_macaddr=B0DD74AF93E0
ethaddr=B0:DD:74:AF:93:E0
That is why i think it is taken from there.
img-702916436_vol-env1.ubifs:
img-702916436_vol-env2.ubifs:
img-702916436_vol-overlay1.ubifs:
img-702916436_vol-overlay2.ubifs:
my last build with changes 76ea660aef
Will test on cx30, should I try to flash that uboot, or load it from uboot and then initramfs and then use sysupgrade image?
don't replace uboot for now
For now, use the xx230v chainload for testing, until I consider uboot stable and update the dts files between openwrt and uboot
Well this one panicked after getting to ubi line:
I'll have to revert the Ethernet driver to see why I'm getting a kernel panic on Ethernet, as this is only happening with the en7529 SoC; it's working normally with the en7562
Regarding the backups you're making, using chainload to back up the NAND probably won't give you the same error as with the vendor u-boot
It gives the very same error on laxy, and mtd list incorrect but that might be because it is not uboot for this.
Only the mapping is wrong, but the size is correct, this won't prevent the dump from being done
Is your NAND have 256Mb of size?
It should have (think all have 256Mb size, and 512 RAM except c20 as that have half of RAM that is why i had to backup it half) and it only not reads the last 2MB maybe error on NAND (sadly only have one laxy box so cant test it on another)
New release still locks up same place.
Meanwhile let a box on the internet to test the Heimgard app and it worked (after a few restart), noticed ota url changed to hmg-ota.genexis.eu but they have the very same last version for theese as Heimgard. (because hmg most likely means Heimgard anyway :))
I need to do a rollback of the ethernet driver and it will probably NPU drive, to solve this problem
Build based in this commit ec2fb58a49f491218af7d01d2c091775c5fd9379
No panic this time, no ethernet and no wifi. Will try on the box i usuallly test in a few hour (this is a cx30 too but at home)
Second box same, cx20 same (no ethernet/wifi but boots), hx40 boots (no ethernet/wifi), hx30 crash at npu.
In my xx230v it was the same thing, I must have reverted to some problematic commit I was having last month, I'll have to check my local Git repository to find the working driver, since I got this one from the server
Regarding the MT76, we must be having some problem with the EEPROM load because if it's a raw MTD it can initialize correctly, but when it's in the UBI it can't load. I must be doing something wrong
now
CONFIG_MTD_UBI_NVMEMis enabled in kernel config (needs to get eeprom in ubi)dts from firmware
It cx30 locked now on NPU (at least that was the last line)
@Gemnon try this builds
Latest build boot log on cx30.
The Ethernet driver hasn't changed at all in the latest commits 🤔, But only the SPI driver was working, so I'm not really sure what it is. Did you test it with the latest build from the automatic build releases?
I'm migrating my email server, so I won't be able to check it tomorrow, probably today i make solution for this, but my xx230v and Mikrotik not having this problem in Ethernet driver
Used the one from:
https://sirherobrine23.com.br/airoha_an7523/builds/releases/download/v59/openwrt-airoha-an7523-genexis_cx30-initramfs-kernel.bin
Just redownloaded to be sure but same error. :)
If v58 or v54 works, it's probably due to some change I made in the DTS file when I renamed it from en7523 to an7523, so I'll have to look into that.
Thanks, will check on the older ones, is it not possible that the cause it is bigger than mine (i stripped down the first test to 8m because vendor uboot unpack) and while cross uboot can extract it but as it is not protecting reserved memory blocks (i think stock uboot fails extracting because of it) later it cause issues?
If you're writing to the NAND, the initramfs file doesn't alter anything in the NAND, it only reads the MAC addresses and nothing else, because everything is literally in RAM
It would be a problem if you were to record on the NAND because you would have to do some things in ubi for write Openwrt with new ubi partitions mapping
Iam just booting from tftpboot right now via chainloader, what i compiled from
v51source (april 18th) back then boots, but v51 from the releases page locks up. Also stock uboot capable to but sysupgrade images kernel (and locks up the very same place), so not get to the point to find out if there is no rootfs. (so it must be size its issue)I'll take a look, I still haven't been able to mount the rootfs either with xx230v, I also attached the last log (initramfs) which was from the 06th
@Gemnon test this build
It crash:
`[ 80.667443] Hardware name: Generic DT based system
[ 80.667448] PC is at handle_softirqs+0x90/0x268
[ 80.667467] LR is at 0xed4720
[ 80.667474] pc : [] lr : [<00ed4720>] psr: 40000113
[ 80.667480] sp : c0013fa0 ip : 00000000 fp : c0085280
[ 80.667485] r10: c1ea6d00 r9 : c1ea6d00 r8 : c1f03d40
[ 80.667491] r7 : 0000000a r6 : 00000040 r5 : 00000282 r4 : c0085280
[ 80.667497] r3 : 1daae000 r2 : 00000000 r1 : c1ea6d00 r0 : cfe9e868
[ 80.667505] Flags: nZcv IRQs on FIQs on Mode SVC_32 ISA ARM Segment user
[ 80.667513] Control: 10c5383d Table: 835d806a DAC: 00000055
[ 80.667517] Call trace:
[ 80.667526] handle_softirqs from irq_exit+0x98/0xc4
[ 80.667548] irq_exit from call_with_stack+0x18/0x20
[ 80.667568] call_with_stack from __irq_svc+0x6c/0x7c
[ 80.667579] Exception stack(0xc005bcd0 to 0xc005bd18)
[ 80.667588] bcc0: c3268400 0000000f ffff0000 c3268794
[ 80.667598] bce0: c3268400 00000001 0000000f e029834c 0000000f 00000af0 e0298348 c1fc2d10
[ 80.667606] bd00: 00055e15 c005bd20 c05ecb08 c06c5194 a0000013 ffffffff
[ 80.667613] __irq_svc from page_pool_destroy+0x20/0x18c
[ 80.667628] page_pool_destroy from airoha_qdma_cleanup+0x27c/0x368
[ 80.667645] airoha_qdma_cleanup from airoha_probe+0x2150/0x25b8
[ 80.667659] airoha_probe from platform_probe+0x5c/0xb4
[ 80.667677] platform_probe from really_probe+0xc8/0x2c8
[ 80.667691] really_probe from __driver_probe_device+0x88/0x1a0
[ 80.667703] __driver_probe_device from driver_probe_device+0x38/0xe4
[ 80.667714] driver_probe_device from __driver_attach+0x90/0x140
[ 80.667725] __driver_attach from bus_for_each_dev+0x60/0x94
[ 80.667744] bus_for_each_dev from bus_add_driver+0xd0/0x1fc
[ 80.667760] bus_add_driver from driver_register+0x80/0x11c
[ 80.667771] driver_register from do_one_initcall+0x48/0x27c
[ 80.667785] do_one_initcall from kernel_init_freeable+0x230/0x284
[ 80.667801] kernel_init_freeable from kernel_init+0x1c/0x12c
[ 80.667818] kernel_init from ret_from_fork+0x14/0x38
[ 80.667830] Exception stack(0xc005bfb0 to 0xc005bff8)
[ 80.667838] bfa0: 00000000 00000000 00000000 00000000
[ 80.667847] bfc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
[ 80.667855] bfe0: 00000000 00000000 00000000 00000000 00000013 00000000
[ 80.888695] sched: DL replenish lagged too much
`
Which SoC is this? EN7529?
CPU: Airoha EN7529DT
Based on vendor uboot. Stock kernel sais EN7523 though:
the xx230v is EN7529CT, in theory is same SoC
This my latest build with npu disable to genexis dts
This is what happened:
This is only happening with the cx30, is laxy having the same problem? because this only happened to me once when I was messing with the SPI where the entire boot system would hang, worrying about the mac address of the ram or flash memory.
try this build
Theese booted ethernet not work though. :)
You let it boot completely because it was still loading the kernel modules. Try again and wait a bit, as the drivers loaded directly
If it loaded, it's because the MAC address was being mapped by uboot env
I did, some reason eth0 gets into bridge. But i have lan1, lan2 and wan , once i added those to bridge and removed eth0 ping works so ethernet works .If i remember good in dsa that is how it should be, but i might be wrong. But you right it is working once i get the interfaces to bridge.
I haven't yet added these devices to the 02_network, so the bridge isn't loading on the network interfaces.
CX30 works fine with v60 , will test other devices and figure out how to get to flash. Genexis/iopsys image is fit in fit, in the fit there is an "uncompressed" boot fit image that contains kernel and dtb and gets imxtracted to kernel partition, and then there is rootfs what it imxtract to rootfs. Openwrt sysupgrade image is all in one.
The base MAC address is the same for Ethernet? Because if so, I'll add the script to correct this for the Genexis devices, but for now the devices will use the 2.4GHz wireless address for the network switch until I add the script if necessary
Nope they actually different even wifi mac is. And it is the same with openwrt on zephyr and pixly (i mean they get the very same mac), there is a base_macaddr environment variable what they use for ethernet in AABBCCDDEEFF format. Also on zephyr 2.5gbit ports not work yet not even sure there is driver for it right now :) Attaching logs, orig is the original iowrt boot, openwrt is v60. Going to check the other device in a bit.
For rodimus ethernet seems to be broken there is lan1 in openwrt but non of the ports seem to work. Also rodimus and zephyr have zigbee thing on usb (not even used in stock firmware so no idea why it is there) and never used zigbee so not even know what it is good for.
:
Arcee has a phantom wan (it has just two lan ports :)) but otherwise seem to work just mac address have nothing to do with the devices (like the others)
Laxy there is lan1,lan2,lan3 (it has 4 ports and pon) could not get link on any but that boots too.
Added a few pics from the back of the boxes ,for port numbers (rodimus-hx30) have 3 buttons on the side (Wlan/WPS/reset).
I’d like to ask you to provide a short list so I can create the dts files correctly:
ZigBee is used to create a network to control IoT devices without relying on Wi-Fi to create that network. If I'm not mistaken, the W1700k also has a ZigBee component, which it seems wasn't used either
What was causing the chaotic startup was me trying to use the MAC address as the stored base for the u-boot environment, I think something is missing in the kernel, that's why OpenWRT wasn't starting directly, I'll have to create a script to correct the addresses that are extracted from the u-boot environment later
So the LEDs are working on all devices?
If you are going to submit it, please send it to this repository: https://sirherobrine23.com.br/airoha_en7523/off-topics-private/issues
laxy: 4x 1gbit ports lan1,lan2,lan3,lan4, has 1xUSB 3.0 (at least iowrt shown 5000 speed when connected my usb drive to it and it was recognized on usb2) it has xPON port, also have 2 fxs port (i do not think there is external PHYs)
rodimus: 4x 1gbit ports lan1,lan2,lan3,lan4, 1 2.5gbit wan port wan, has 1xUSB 2.0 (device got recognized on usb1) and 1 fxs port, no xPON. (i do not think there is external PHYs)
zephyr: 3x1gbit port ( lan1, lan2, lan3), 2x 2.5gbit port (lan4,wan) has 1xUSB 2.0 (device got recognized on usb1), 1 fxs port and i think lan4 is an external phy:
arcee: 2x1gbit port (lan1,lan2), no USB
pixly: 3x1gbit port (lan1,lan2,wan) no USB
Will try to find those, but as we chainload foreign uboot to be able to boot initramfs kernel wont get envs from that? :)
Sorry still going to test new builds just previous weeks was busy , could get iowrt stuff to at least build packages I need for the devices. But still would prefer openwrt. Got some more boxes. Well can be sure CX20 is arcee and all this stuff are engineering boxes.
Sorry if we don't pay much attention to some things, I'm fixing some drivers, and others are probably busy with other things.
Interesting, if the PCB has any revisions it will be interesting, some devices have undergone minor changes on my part, but most people in this organization can open PRs and/or commit, if they want to send something to a repository.
No worries iam glad what you do! Iam not much a developer sadly, just removed their call home stuff from rootfs and made sure on reset boxes not cause havoc, can't even build iowrt as device profiles needs GitHub login :) Not sure how GPL compatible is this. But then even could not find Heimgard repo or at least the last source tarball before it went Genexis :)
Tested the latest build on pixly mac addresses good now and both ethernet/wifi working. :)
Have a lot of lines of this though (but think that just some debug info)
[ 1051.971303] airoha-spi 1fa10000.spi: exec_op cs=0 opcode=13 tx=1 addr=3 dummy=0 data=0 dir=0
That's good, that SPI log was a debug log I forgot to remove
I was testing some things for Mikrotik hEX S 2025, but I'll have to do it in uboot and not in the kernel
@Gemnon I added some changes for a few devices today (
47fbe84c61)I'm not sure which Ethernet chips some of these devices use (likely the same as the e650) so please wait for the build in github so I can update the dts, 01_leds, and 02_network.
They use Airoha EN8811.