Odroid N2 | Hard lockup on USB drive insert/remove #4188
Labels
Milestone
Comments
|
Many thanks for your report. Are there other kernel errors, before the actual crash happens? Just in case some kernel files are broken, please try to reinstall it: |
|
Hi @MichaIng , Thanks for the quick response! I reinstalled the kernel as requested. Rebooted & plugged in the USB drive and waited about six hours before the unit locked up again. I was running Kernel errors: |
|
Hmm, I don't like kernels which print "expected" error messages as now it's very hard to know if there is an actual issue or not... One thing could be muted by disabling CEC support. I did just create/update the boot.ini for N2: https://github.com/MichaIng/DietPi/blob/dev/boot_n2.ini Then CEC can be disabled: Remembering the initial reason for the boot.ini update: Is it actually an Odroid N2 or N2+? The latter was not really supported but with this new boot.ini it is as it loads a different device tree. Maybe this is even the reason for all those kernel errors. |
|
Hi @MichaIng, Thanks for that. The devices are definitely an N2, not N2+. I bought two of them over a year ago (and both are affected - which sort of rules out faulty hardware/PSU issues). Followed your instructions above (copy/pasted both & confirmed they worked), unfortunately the system is still crashing. Had Logfile attached. Click to expand!
Logfile attached. Click to expand! |
|
CEC errors are still there, which makes sense as CEC is not disabled by boot.ini but only not actively enabled. I.e. it was disabled before and I guess the error is coming from a kernel module which cannot find the "hdmirx" option. The other kernel errors can all be found in various other cases on Odroid and Armbian in combination with the Hardkernel kernel. Not beautiful (IMO) but likely not related . Can you try to clock down the CPUs a bit? I don't think it helps, but it does not hurt to try it out: |
|
Hi @MichaIng, Applied the above as requested.
Noticed that I can trigger the lockup if I insert the USB card reader slowly (presumably not a 'clean' insert). Not sure if this is related or another issue? Surely we shouldn't have a crash if a USB device has some issues on insert?!?
Logfile attached. Click to expand! |
|
Also, another interesting observation, I implemented a hardware watchdog as per https://wiki.odroid.com/odroid-n2/application_note/software/watchdog_timer. This works for the test scenarios however does not work for the above (system still dies & remains unavailable). |
|
Dammit, I had a typo in the boot.ini fix 😄. Please try again with: The last command, after a reboot, should then show a max CPU frequency on all cores lowered to 1608 MHz. However, I do not expect this to help, just to rule it out. @meveric |
|
I remember that right at the beginning the N2 had terrible USB support, whit lots of similar issues as you describe. Since this is very close to the hardware itself, USB controller, Kernel, etc. it's probably best to ask HardKernel directly. I've seen similar reports on the coreelec forum, so it doesn't seem to be that uncommon. |
|
There are a few other USB-related issues already, but none that matches this case good enough IMO. @TheOriginalMrWolf please report the issue at: https://github.com/hardkernel/linux/issues |
|
Hi @MichaIng, Looks like something still not right with boot.ini - I'm expecting to see Max Freq == 1608 across the board:
|
|
|
|
Okay, frequencies are valid. Two ideas:
|
|
Hi @MichaIng - the 'ancient' syntax worked. Unfortunately this didn't help with the crashing. |
|
Have you tried switching the CPU governor? |
|
Just using the default so far. |
|
Default is ondemand on DietPi btw. And @meveric you were right about using |
MichaIng
changed the title
|
@MichaIng |
|
I guess it is a persistent issue with some USB drives and N2, though I personally didn't run into any with my N2+ so far. Experimenting with mainline kernel is on my ToDo around new years eve when I'm back home (currently family visits). |
MichaIng
added
Testing/testers required 🔽
Solution available 🥂
|
New image ready for testing based on mainline kernel: https://dietpi.com/downloads/images/testing/ |
|
Referring to the thread for testing the new image and closing this one: #5039 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Hi,
Running on Odroid N2, I experience hard lockups (console unresponsive, no response to ping on LAN nor any other network traffic, SSH connections dropped) when using USB drives.
This can occur either randomly on install/remove of a drive, or after the drive has simply been plugged in for a few hours/days. Occurs if drive is either plugged directly into SBC, or via a powered Pluggable USB3-Hub7C.
Was connected via SSH & running
journalctl -fwhen this happened and saw:Is there any way to better capture debug information?
Required Information
DietPi version |
cat /boot/dietpi/.versionG_DIETPI_VERSION_CORE=7
G_DIETPI_VERSION_SUB=0
G_DIETPI_VERSION_RC=2
G_GITBRANCH='master'
G_GITOWNER='MichaIng'
Distro version |
echo $G_DISTRO_NAMEorcat /etc/debian_versionBuster / 10.8
Kernel version |
uname -aLinux NewPi 4.9.241-arm64 #1 SMP PREEMPT Thu Feb 25 18:56:07 CET 2021 aarch64 GNU/LinuxSBC model |
echo $G_HW_MODEL_NAMEor (EG: RPi3)Odroid N2 (aarch64)
Power supply used | (EG: 5V 1A RAVpower)
Odroid supplied PSU
SDcard used | (EG: SanDisk ultra)
Sandisk Ultra in Kingston MobileLite G4 card reader
Samsung T5 Portable SSD
Additional Information (if applicable)
Steps to reproduce
or
Expected behaviour
Actual behaviour
Extra details
The text was updated successfully, but these errors were encountered: