RX-AX56U Consistent Kernel Panics after upgrade

[sputnikk]

Hi all,

After a recent upgrade to 386.2 (I have also tried 386.3) - my router keeps getting kernel panics.

Nothing seems to stop working except my WAN - which is PPPoE based. Hard resetting the modem OR flipping the "Internet Connection" toggle on "Network Map" restores it.
I can't seem to find a trigger either.

Not sure where to go from here. Could someone lend a pointer please?

Aug 18 01:03:00 kernel: CPU: 1 PID: 26636 Comm: dcd Tainted: P           O    4.1.52 #1
Aug 18 01:03:00 kernel: Hardware name: Generic DT based system
Aug 18 01:03:00 kernel: task: d5ef7000 ti: d07d6000 task.ti: d07d6000
Aug 18 01:03:00 kernel: PC is at 0xb6c9239c
Aug 18 01:03:00 kernel: LR is at 0x1e808
Aug 18 01:03:00 kernel: pc : [<b6c9239c>]    lr : [<0001e808>]    psr: 600f0010
Aug 18 01:03:00 kernel: sp : bed84958  ip : 000a3050  fp : b5dff024
Aug 18 01:03:00 kernel: r10: 000a33c8  r9 : b5dff680  r8 : 000a387c
Aug 18 01:03:00 kernel: r7 : b5dff6b8  r6 : 000a3876  r5 : 00000000  r4 : b5dff664
Aug 18 01:03:00 kernel: r3 : 00000000  r2 : 88b48e00  r1 : 0007e101  r0 : 00000000
Aug 18 01:03:00 kernel: Flags: nZCv  IRQs on  FIQs on  Mode USER_32  ISA ARM  Segment user
Aug 18 01:03:00 kernel: Control: 10c5387d  Table: 0cc4c04a  DAC: 00000015
Aug 18 01:03:00 kernel: CPU: 1 PID: 26636 Comm: dcd Tainted: P           O    4.1.52 #1
Aug 18 01:03:00 kernel: Hardware name: Generic DT based system
Aug 18 01:03:00 kernel: [<c0026fe0>] (unwind_backtrace) from [<c0022c38>] (show_stack+0x10/0x14)
Aug 18 01:03:00 kernel: [<c0022c38>] (show_stack) from [<c048e07c>] (dump_stack+0x8c/0xa0)
Aug 18 01:03:00 kernel: [<c048e07c>] (dump_stack) from [<c003ac88>] (get_signal+0x490/0x558)
Aug 18 01:03:00 kernel: [<c003ac88>] (get_signal) from [<c00221d0>] (do_signal+0xc8/0x3ac)
Aug 18 01:03:00 kernel: [<c00221d0>] (do_signal) from [<c0022658>] (do_work_pending+0x94/0xa4)
Aug 18 01:03:00 kernel: [<c0022658>] (do_work_pending) from [<c001f4cc>] (work_pending+0xc/0x20)



Aug 18 01:08:57 kernel: CPU: 1 PID: 32067 Comm: dcd Tainted: P           O    4.1.52 #1
Aug 18 01:08:57 kernel: Hardware name: Generic DT based system
Aug 18 01:08:57 kernel: task: cda34c00 ti: cd070000 task.ti: cd070000
Aug 18 01:08:57 kernel: PC is at 0xb6ccd39c
Aug 18 01:08:57 kernel: LR is at 0x1e808
Aug 18 01:08:57 kernel: pc : [<b6ccd39c>]    lr : [<0001e808>]    psr: 600f0010
Aug 18 01:08:57 kernel: sp : beca6958  ip : 000a3050  fp : b5dff024
Aug 18 01:08:57 kernel: r10: 000a33c8  r9 : b5dff680  r8 : 000a387c
Aug 18 01:08:57 kernel: r7 : b5dff6b8  r6 : 000a3876  r5 : 00000000  r4 : b5dff664
Aug 18 01:08:57 kernel: r3 : 00000000  r2 : 35019300  r1 : 0007e101  r0 : 00000000
Aug 18 01:08:57 kernel: Flags: nZCv  IRQs on  FIQs on  Mode USER_32  ISA ARM  Segment user
Aug 18 01:08:57 kernel: Control: 10c5387d  Table: 0cae404a  DAC: 00000015
Aug 18 01:08:57 kernel: CPU: 1 PID: 32067 Comm: dcd Tainted: P           O    4.1.52 #1
Aug 18 01:08:57 kernel: Hardware name: Generic DT based system
Aug 18 01:08:57 kernel: [<c0026fe0>] (unwind_backtrace) from [<c0022c38>] (show_stack+0x10/0x14)
Aug 18 01:08:57 kernel: [<c0022c38>] (show_stack) from [<c048e07c>] (dump_stack+0x8c/0xa0)
Aug 18 01:08:57 kernel: [<c048e07c>] (dump_stack) from [<c003ac88>] (get_signal+0x490/0x558)
Aug 18 01:08:57 kernel: [<c003ac88>] (get_signal) from [<c00221d0>] (do_signal+0xc8/0x3ac)
Aug 18 01:08:57 kernel: [<c00221d0>] (do_signal) from [<c0022658>] (do_work_pending+0x94/0xa4)
Aug 18 01:08:57 kernel: [<c0022658>] (do_work_pending) from [<c001f4cc>] (work_pending+0xc/0x20)

 

[Volt]

That's a well-known message (that can be safely ignored) related to Ai Protection. It is very unlikely that it affects WAN.

 

[sputnikk]

Thanks. I see the article relating to it now.

I would expect disabling AIProtection would stop these panics, but it has not

 

[Volt]

Try clicking "Withdraw" on the "Privacy" tab, and then reboot the router.

 

[sputnikk]

Will do.

In trying to chase down the WAN PPPoE connectivity, I only found this in GUI logs:

Aug 19 22:15:27 WAN_Connection: Fail to connect with some issues.


Flipping the WAN switch on/off fixes it (or "service restart_wan"). I'm surprised the router doesn't detect this condition and reset WAN on it's own. Internet detection set to PPP Echo.

I was blaming the bridged modem, now I blame the ASUS.

I'm afraid my only option now is to write a crontab script to check connectivity and service restart_wan for me.

 

[sputnikk]

After multiple reflashes, making custom reset scripts, updating firmware, toggling various things - disabling (or rather, not using) CakeQOS has made it stable. Could be coincidence with something else, but its been stable since.

I kept enabling it having seen that it went from an ATM install to right in the GUI. I liked the feature set given my low bandwidth but Wifi6 clients. I can live without. Perhaps CakeQOS is memory intensive and AX56U doesn't provide enough RAM? (though I have swap on ext2 enabled).

I also wish there was more debug logging coming from the stack as to the cause of my WAN disconnects.

I'm stable now. Hopefully this helps someone.