01-01-2020, 02:08 PM
(This post was last modified: 01-01-2020, 02:21 PM by Twix166.
Edit Reason: Corrected some type-os
)
I have also been seeing frequent crashes in the past day. I have been using my PBP all day today doing Audiobook proof listening. So I have VLC running and a PDF reader open alongside Chromium to edit my Google sheet in which I make the edit notes.
Initially I was seeing the WiFi drop-outs and was able to stabilise this by setting managed=true for the NetworkManager. WiFi is now behaving correctly but the crashes have kept happening. Sometimes as frequently as every 10 minutes. I also noticed that when the crash happens, it does take about 10 seconds for the system to reboot. Which led me to believe that this is a kernel crash. As mentioned earlier, the 10 second timeout is one of the kernel options that is specified.
With this information, I decided to tail /var/log/kern.log so that I could keep an eye on it. I also monitored CPU temperature by using 'watch -n 5 cat /sys/class/thermal/thermal_zone0/temp'. The temperature readout hovered around 36500 (assuming this means 36.5 degr C) most of the time. But, upon the moment of the crash, it had risen to around 45000. This is still not very hot in my opinion but, it may be related. It might be that whatever causes the kernel to crash causes the temp to rise.
The crash happens after 'endless' mention of 'wl_run_escan'. I am not sure what this means but perhaps somebody else can use this information. I did do a quick search relating to 'wl_run_escan' and found that there is a related heap overflow vulnerability. It is documented here: https://packetstormsecurity.com/files/14...rflow.html
<<UPDATE>> It looks like the vulnarability may have been fixed:
https://review.lineageos.org/c/LineageOS.../+/184549/.
I don't know if this made it into the Debian/Mate build though. I am afraid I am starting to run out of skill at this point.
I have not had a chance to further investigate this so this may take me on a wild goose chase. I thought I'd share rather than sit on this so that other can maybe confirm this behaviour.
I will continue to monitor the logs and temps.
Initially I was seeing the WiFi drop-outs and was able to stabilise this by setting managed=true for the NetworkManager. WiFi is now behaving correctly but the crashes have kept happening. Sometimes as frequently as every 10 minutes. I also noticed that when the crash happens, it does take about 10 seconds for the system to reboot. Which led me to believe that this is a kernel crash. As mentioned earlier, the 10 second timeout is one of the kernel options that is specified.
With this information, I decided to tail /var/log/kern.log so that I could keep an eye on it. I also monitored CPU temperature by using 'watch -n 5 cat /sys/class/thermal/thermal_zone0/temp'. The temperature readout hovered around 36500 (assuming this means 36.5 degr C) most of the time. But, upon the moment of the crash, it had risen to around 45000. This is still not very hot in my opinion but, it may be related. It might be that whatever causes the kernel to crash causes the temp to rise.
Code:
Jan 1 19:30:01 eggsy kernel: [ 4644.137134] wl_run_escan: LEGACY_SCAN sync ID: 43, bssidx: 0
Jan 1 19:32:01 eggsy kernel: [ 4764.140080] wl_run_escan: LEGACY_SCAN sync ID: 44, bssidx: 0
Jan 1 19:34:01 eggsy kernel: [ 4884.139105] wl_run_escan: LEGACY_SCAN sync ID: 45, bssidx: 0
Jan 1 19:36:01 eggsy kernel: [ 5004.138227] wl_run_escan: LEGACY_SCAN sync ID: 46, bssidx: 0
Jan 1 19:38:01 eggsy kernel: [ 5124.141598] wl_run_escan: LEGACY_SCAN sync ID: 47, bssidx: 0
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^$
Jan 1 19:39:57 eggsy kernel: [ 0.000000] Initializing cgroup subsys cpuset
Jan 1 19:39:57 eggsy kernel: [ 0.000000] Initializing cgroup subsys cpu
Jan 1 19:39:57 eggsy kernel: [ 0.000000] Initializing cgroup subsys cpuacct
Jan 1 19:39:57 eggsy kernel: [ 0.000000] Linux version 4.4.190 (adam@adam-vb) (gcc version 7.1.1 20170707 (Linaro GCC 7.1-201$
Jan 1 19:39:57 eggsy kernel: [ 0.000000] Boot CPU: AArch64 Processor [410fd034]
Jan 1 19:39:57 eggsy kernel: [ 0.000000] earlycon: Early serial console at MMIO32 0xff1a0000 (options '')
Jan 1 19:39:57 eggsy kernel: [ 0.000000] bootconsole [uart0] enabled<<UPDATE>> It looks like the vulnarability may have been fixed:
https://review.lineageos.org/c/LineageOS.../+/184549/.
I don't know if this made it into the Debian/Mate build though. I am afraid I am starting to run out of skill at this point.
I have not had a chance to further investigate this so this may take me on a wild goose chase. I thought I'd share rather than sit on this so that other can maybe confirm this behaviour.
I will continue to monitor the logs and temps.
