Jetpack 5.1 kernel panic on reboot

[seeky15]

We had this issue in 5.0.2 already.

Is there a fix for this kernel panic already for the 5.1?

 reboot
[  164.497466] Trying to unregister non-registered hwtime source
[  164.499435] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000000
[  164.499707] Mem abort info:
[  164.499809]   ESR = 0x86000004
[  164.499881]   EC = 0x21: IABT (current EL), IL = 32 bits
[  164.500009]   SET = 0, FnV = 0
[  164.500158]   EA = 0, S1PTW = 0
[  164.500233] user pgtable: 4k pages, 48-bit VAs, pgdp=0000000116b4f000
[  164.500367] [0000000000000000] pgd=0000000000000000, p4d=0000000000000000
[  164.500537] Internal error: Oops: 86000004 [#1] PREEMPT SMP
[  164.500654] Modules linked in: lzo_rle lzo_compress realtek zram aes_ce_blk crypto_simd cryptd aes_ce_cipher ghash_ce sha2_ce sha256_arm64 sha1_ce leds_gpio max77620_thermal tegra_bpmp_thermal input_leds userspace_alert tegra210_adma spi_tegra114 overlay binfmt_misc ina3221 pwm_fan nvgpu nvmap
[  164.501357] CPU: 3 PID: 153 Comm: kworker/u12:3 Tainted: G           OE     5.10.104-tegra #1
[  164.501532] Hardware name: Unknown NVIDIA Jetson Xavier NX Developer Kit/NVIDIA Jetson Xavier NX Developer Kit, BIOS 2.1-32413640 01/24/2023
[  164.504602] Workqueue: events_power_efficient phy_state_machine
[  164.510635] pstate: 60c00009 (nZCv daif +PAN +UAO -TCO BTYPE=--)
[  164.516580] pc : 0x0
[  164.518942] lr : phy_link_change+0x40/0x70
[  164.523139] sp : ffff80001155bd20
[  164.526380] x29: ffff80001155bd20 x28: ffffc3073cfa6000
[  164.531893] x27: ffff8000108fbca8 x26: ffff495f9715c648
[  164.537145] x25: 0000000000000000 x24: ffff495e451f0000
[  164.542658] x23: ffff495e5b4b3cf8 x22: ffff495e44c16580
[  164.548428] x21: 0000000000000000 x20: ffff495e451f0000
[  164.553851] x19: ffff495e5b4b3800 x18: 0000000000000000
[  164.559295] x17: 0000000000000007 x16: 000000000000000e
[  164.564532] x15: 0000000000000004 x14: 0000000000000019
[  164.570303] x13: 0000000000000000 x12: ffff495e57e5c8c8
[  164.575399] x11: ffff495e57e5c6d8 x10: 0000000000000a80
[  164.580807] x9 : ffff495e44670d7c x8 : fefefefefefefeff
[  164.586402] x7 : 0000000000000018 x6 : ffff495e44670d6c
[  164.591570] x5 : 0000746e65696369 x4 : 000000000000002f
[  164.597253] x3 : 0000000000000000 x2 : ffffc3073b6de450
[  164.602590] x1 : 0000000000000000 x0 : ffff495e451f0000
[  164.607670] Call trace:
[  164.610380]  0x0
[  164.611961]  phy_state_machine+0x184/0x220
[  164.616244]  process_one_work+0x1c4/0x4a0
[  164.620268]  worker_thread+0x54/0x430
[  164.623942]  kthread+0x148/0x170
[  164.626939]  ret_from_fork+0x10/0x24
[  164.630694] Code: bad PC value
[  164.633578] ---[ end trace dc750f5d02f3052d ]---
[  164.638378] Kernel panic - not syncing: Oops: Fatal exception
[  164.643721] SMP: stopping secondary CPUs
[  164.647750] Kernel Offset: 0x43072b520000 from 0xffff800010000000
[  164.653864] PHYS_OFFSET: 0xffffb6a2c0000000
[  164.657635] CPU features: 0x8240002,03802a30
[  164.662177] Memory Limit: none
[  164.665077] ---[ end Kernel panic - not syncing: Oops: Fatal exception ]---

[kayccc]

I can’t comment whether it’s fixed or not as there qre couple issues fixed in new release.
Please get it a try with JetPack 5.1.

Thanks

[seeky15]

Oh I mean the text is from 5.1

Wayne had a patch for the kernel in 5.0.2
Do you have one for 5.1 already or do we have to deal with kernel panics on reboot again?

[WayneWWW]

Hi,

Where does this come from? From the ethernet? or something you are not sure either?

[seeky15]

Hey @WayneWWW I am not sure, but as you pointed out the phy_link_change would point to ethernet, yep.

[WayneWWW]

Oh. If it is that issue which happened intermittently, then it is not resolved yet. Sorry about it.

The workaround patch from 5.0.2 is still needed.

[seeky15]

Seems I’ve missed that. I’ve only got one applied for the panic caused by the display port. Can you link me the patch for the ethernet panic?

[WayneWWW]

Please try to remove below in your kernel-dts/t19x-common-platforms/tegra194-platforms-eqos.dtsi

  interrupt-parent = <&tegra_main_gpio>;
  interrupts = <TEGRA194_MAIN_GPIO(G, 4) IRQ_TYPE_LEVEL_LOW>;

The side effect is wake on LAN will not work after it.