Duet crash when refreshing WebGUI

krzaku

I just had the most unpleasant experience. I lost a 15 hour print at hour 12 due to an apparent bug in the firmware that caused the Duet to crash when I refreshed the WebGUI page. Admittedly I might have refreshed the page one too many times but this should not cause the firmware to crash. I have the DuetWiFi with the firmware version 1.19 installed. I hope this will get fixed. I know this kind of issues are not easy to diagnose but it is a 120 GBP device after all.

dc42

I'm sorry you lost your print, but I have never heard of that happening before. Did you take a M122 dump immediately afterwards? The software reset data might give a clue about what happened.

krzaku

Unfortunately not, I didn't think about it at all. One other thing is the file I was printing was quite large, 96 MB in size, not sure if that might have had anything to do with my issue.

Motalw

Hi Dave I also had the duet crashing few times and the only way to get it to connect to the web interface again is powering it off and on again. usually it happens if I try to probe the bed when the probe is already triggered. I'm using yours DC42 IR Probe.

krzaku

I managed to crash it again by refreshing the WebGUI repeatedly while the printer was in simulation mode and "printing a file" (not sure if the simulation mode is connected to the crash in any way). Here's the log data you requested David:

Last reset 00:00:19 ago, cause: software
Last software reset reason: User, spinning module Network, available RAM 3224 bytes (slot 0)
Software reset code 0x5001, HFSR 0x00000000, CFSR 0x00000000, ICSR 0x0440f80f, BFAR 0xe000ed38, SP 0x2001f9b4
Stack: 00434fcb 00008096 ffffffe9 00051051 00081111 40080000 00008096 00000000 0042fceb 004338b0 61000200 000000d1 00000000 00000000 00000000 00000000 00000000 4c5693a5
Error status: 0

Full log: https://pastebin.com/LfV4e6qe

On a side note, it would be a good idea for the firmware to run M122 on it's own every time it is turned on, and if it detects an abnormal reset then it would output the M122 data (or whatever else you would need when debugging) to the SD card.

krzaku

I got another crash, this time I wasn't playing with the WebGUI. I restarted the printer, and it crashed right after it booted with this error in M122:

Last reset 00:00:15 ago, cause: software
Last software reset reason: User, spinning module Network, available RAM 7440 bytes (slot 2)
Software reset code 0x5001, HFSR 0x00000000, CFSR 0x00000000, ICSR 0x0040080f, BFAR 0xe000ed38, SP 0x2001f9b4
Stack: 00434fcb 00008196 ffffffe9 00051051 00081111 40080000 00008196 00000000 0042fceb 004338a2 21000200 00000000 00000000 00000000 00000000 00000000 00000000 00000000
Error status: 0

Any ideas @dc42?

dc42

Please try with firmware 1.19.2 (both DuetWiFiFirmware and DuetWiFiServer), or alternatively with the 1.20beta1 firmware that I expect to release tomorrow.

krzaku

As 1.20beta1 is still not available, I have tried 1.20alpha7. It's still crashing, here's the log:

Last reset 00:00:31 ago, cause: software
Last software reset reason: Stuck in spin loop, spinning module Network, available RAM 7272 bytes (slot 3)
Software reset code 0x4041, HFSR 0x00000000, CFSR 0x00000000, ICSR 0x0040080f, BFAR 0xe000ed38, SP 0x2001f9bc
Stack: 004269e1 00008096 004358e3 00008096 ffffffe9 00001000 00081111 40080000 00008096 20002c30 004306d3 0043426e 41030200 476a6000 430c7168 3edb6db7 b7b4d800 3e178897
Error status: 0

krzaku

David, are there any news on this?

dc42

I'm sorry for the delay in responding. I read your messages on a portable device, so it wasn't easy to flag them for looking into.

The status report shows that it spent too long in the spin loop in the Network module, so the software watchdog timed it out. Do you have the USB port connected to anything?

krzaku

No, I never used the USB port.

dc42

I am putting some more debug into the next 1.20 beta release to help pin this down. I have tried to reproduce it, without success.

krzaku

Thanks, will try it and report back.

krzaku

I've been away from 3d printing for a while, I've just come back and I tried reproducing this bug again, and it's still there. Although I can't see any additional debug info you're talking about in the M122 output: https://pastebin.com/9GYuUWFF

dc42

Thanks for your report. The critical part of the pastebin report is:

[[language]]
RepRapFirmware for Duet WiFi version 1.20RC1 running on Duet WiFi 1.0
...
Last reset 00:00:16 ago, cause: software
Last software reset reason: Stuck in spin loop, spinning module Network, available RAM 7944 bytes (slot 0)
Software reset code 0x4041, HFSR 0x00000000, CFSR 0x00000000, ICSR 0x0441f80f, BFAR 0xe000ed38, SP 0x2001f92c
Stack: 0042c809 00008196 0043bffb 00008196 ffffffe9 00001000 00081111 40080000 00008196 20002dfc 00436fff 0043a914 610f0200 407042a1 00000000 3edb6db7 b7254200 3e178897

I have it on my list to look into next week. However, as nobody else has reported this issue, it may indicate a hardware problem.

dc42

Are you able to reproduce this if you refresh the web interface while a print is running in simulation mode?

dc42

I've looked at the stack trace, and it suggests that the processor was waiting forever for the SD card interface to become ready. Please confirm that you are not using an SD extender cable. I suggest you try a new SD card.

krzaku

Using a new SD card seems to have fixed it, I'm not 100% sure though because of the semi-random nature of this problem. Let's say it's resolved for now, if I encounter it again with the new card I'll report it. Thanks for the help in debugging this.