[ 20.900878] device eth0.1 entered promiscuous mode
[ 20.905758] device eth0 entered promiscuous mode
[ 20.940478] br-wlan: port 1(eth0.1) entered forwarding state
[ 20.946310] br-wlan: port 1(eth0.1) entered forwarding state
[ 22.939289] br-wlan: port 1(eth0.1) entered forwarding state
[ 24.809528] Kernel panic - not syncing: stack-protector: Kernel stack is corrupted in: 8701d944
[ 24.823293] Rebooting in 3 seconds..
This Kernel panic is a known bug of the closed source MediaTek WiFi driver. Now we know it can not manage if there is a WiFi AP in the neighborhood with some Unicode character in its SSID.
Onion @administrators deeply kept silent about this bug.
It seems Onion's new WARP CORE driver (also closed source ???) can manage this situation.
Unfortunately Onion @administrators do not say too much about this bug feature (as usual ;-).
@György-Farkas kernel was corrupted for sure, that's something I can tell you with honesty. The device was literally dead, only after inserting a microSD card with a new firmware my device was able to update itself and become functional.
Well... no one but you should know if this correct (or not ;-)
@Lazar-Demin Hello, i'm trying to solve this problem too, can you explain more specifically where the root of the problem is and what modules it affects?
I debugged and find in arch/mips/mm/dma-default.c struct dma_map_ops what consist NULL, do u fix it?
power OFF your Omega2+
press and hold down the Reset button (the momentary push button between the FW_RST pin and the +3.3V)
switch the Omega's power ON
and after this you should see the UBoot menu on the serial console...
Just adding myself to the message chain as I also would like to be able to run openvpn, but I don't really want to jump through all the hoops to compile kernels or the openvpn if it's all going to get broken by an Onion update. I would just rather opkg it and have it run :-)
Relatively unlikely to make a difference, by now that you've done the upgrade you might try booting it without the USB stick plugged in - it represents some additional load on the power circuitry, and adds some complexity of software possibilities.
Generally speaking there are three categories of things that could be wrong
You could have installed a bad image, or at least one that's wrong for the board. Always failing in the same place in the boot sequence would be a hint of this, but it would seemingly not occur with a different image especially one that previously worked.
You could have a power issue that's causing memory corruption rather than an immediate reboot. Always failing in a part o the boot process that involves wifi would be indicative of this
You could theoretically have a badly soldered board or one with bad components, not so bad that memory detection fails right away in U-Boot as has happened in a small number of cases, but more "flaky". However if a boad-level issue this would seem likely to cause more random failures than specific ones. An actual bad memory cell could cause a consistent failure but seems yet less likely.
@Chris-Stratton The problem is, that it does not reboot. It stucks with that message!
There's a chance the CPU did reset, but was unable to read the flash chip because on boot it would be sending 3-byte addresses while the flash chip may still be expecting 4-byte ones from the settings in use before the crash.