Re: Firefox and Thunderbird freeze system causing a reboot

On Thu, 7 Nov 2024, Philipp Ost wrote:

> Hi everyone,
>
> I have a curious situation with my system since updating to 
> 14.2-STABLE on Tuesday, 5.11.'24: both Firefox and Thunderbird cause 
> the system to freeze which then triggers and automatic reboot. This 
> doesn't happen immediately, though. It takes at most 2:15 minutes 
> until this freeze and reboot cycle happens. No user-interaction is 
> taking place during that time.
>
> The machine is running FreeBSD 14.2-STABLE 
> stable/14-n269500-cdffbea57c0d  amd64.
> Firefox is at version firefox-132.0.1_2,2 and Thunderbird at 
> thunderbird-128.4.1_1.
>
> Other gtk3-based programs do not show this behaviour, i.e., Inkscape, 
> Darktable, GIMP, EOG, Filezilla, Gnumeric.
>
> I (re-)built both Firefox and Thunderbird with debugging symbols. 
> Console output of both Firefox and Thunderbird (when being started 
> from an xterm) can be found here:
> philippost.de/firefox-debug-console-output
> philippost.de/thunderbird-debug-console-output
>
> In addition, I tried running them under truss(1), but from skimming 
> through both logs nothing immediately stood out at what could cause 
> this.
>
> Also, there's nothing to be found in the logs at all.
>
> Did anyone else observed this? How do I go about debugging this?
>
> I am at a loss at how to proceed and appreaciate any help in trying to 
> figure this out.

I have had my system crash a few times recently and just enabled crash 
dumps to try to track it down.  The only time that I collected a crash 
dump, I had just started an instance of Firefox and went to the Rocket 
League subreddit on Reddit when the system froze and eventually rebooted 
after saving the crash dump.

At the time of the crash, I had two Firefox profiles already open and 
was building something in Poudriere.  According to the panic, I received 
a general protection fault (Trap 9) with ccache.  Until it occurs more 
times, I will not be able to point at anything specific except 
undervolting.  Well, the other possibility is the revision I am running 
since I did not have the system crash at all prior to updating to it.

In my case, I am undervolting the CPU to control temperatures, so that 
could be the reason.  I have actually been applying less undervolting 
lately due to random programs dying albeit rarely.

My environment:
- FreeBSD 14.1-STABLE (revision 2ee5293d3a9c)
- Intel i7-14700K
- ASUS Rog Strix Z790-F WiFi
- Nvidia 4070 (driver:  550.127.05)
- 128GB memory (running at 4800MHz)
- No errors in MemTest86 after multitudes of tests.

Various things to consider:
- Run MemTest86 on your system for a few passes to make sure your memory
   is stable.  If you undervolt too much, this will fail.  However, it
   could be a sign of bad memory or not enough current to the Memory
   Controller (MC).  The highest level of XMP for my memory will knock
   the voltage of the MC to a low level causing major system instability.

- I doubt it is the act of Firefox or Thunderbird dying that is causing
   the system to crash.  They could be tickling something to cause that
   such as a driver.

- If you have a 13th or 14th Gen Intel CPU, it could have issues due to
   Intel and/or the motherboard manufacturers not getting the voltage
   correct[1].  I am concerned for myself on this one.

- Check /var/log/messages for any memory errors or core dumps.  I got
   these when I tuned the undervolting too high:
       kernel: MCA: Bank 0, Status 0x8000004000050005
       ... exited on signal ...
   Run this command:
       zegrep 'MCA:|exited on signal' /var/log/messages*

   NOTE:  Ignore conftest panics as they happen even with good hardware
   when building packages.

- Check in /var/crash for any panics.  That could be why it takes a
   couple of minutes to reboot; it is writing the panic to disk.

- Enable crash dumps in /etc/rc.conf.  See rc.conf(5) and dumpon(8) for
   information.

- Update your system BIOS if possible.

- Follow PR 281700[2] since that sounds similar.

That is about all I can advise at the moment.  Hopefully, it gives you 
more to try.

Sean
   1. https://community.intel.com/t5/Blogs/Tech-Innovation/Client/Intel-Core-13th-and-14th-Gen-Desktop-Instability-Root-Cause/post/1633239
   2. https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=281700
-- 
scf@FreeBSD.org