Re: HP Laptop freezes while using Xorg sporadically.

From: Edward Sanford Sutton, III <mirror176_at_hotmail.com>
Date: Mon, 14 Jul 2025 15:36:49 UTC
On 7/13/25 18:43, Mauricio wrote:
> I have an HP 255-G7 Laptop running FreeBSD 14.3 RELEASE which every now 
> and then freezes entirely while using Xorg. When the laptop freezes, the 
> only thing that gives me a response is mouse and graphic tablet movement.

   Have you used other versions of FreeBSD or another operating system 
without issue or is this a new-to-you computer with no known history 
good or bad?
   Have you tried different versions of graphics/drm-*-kmod packages? I 
assume you installed them from the kmods repository but is that correct? 
Trying some versions requires an older or newer FreeBSD version but it 
might still help track down if it is a graphics driver bug that is 
introduced or fixed in a newer version.
   Is your BIOS up to date? HP's changelog of BIOS updates is too vague 
to know if there are relevant fixes but you may want to look at 
https://support.hp.com/us-en/drivers/swdetails/hp-255-g7-notebook-pc/model/24381328/swItemId/ob-345901-1 
(if I got the right machine) or even reach out to them for further 
update changes.

> At that point, I can't jump to a virtual console with CTRL+ALT+FUNCTIONKEY.
> I can't even shut the computer down by pressing the shutdown button for 
> less than a second as I usually do.
> But trying to do both things makes the disk activity light indicator 
> start blinking a lot for a moment.

   Do you have swap configured? If so, how much? I've seen systems get 
very slow and even hang when the was no swap and too much memory 
pressure instead of just closing out processes as expected but haven't 
tested such scenarios on 14.3.

> The laptop doesn't seem to get back to work even after half an hour of 
> waiting or more.
> Something that i have been noticing is that the freeze seems to be 
> triggered by an increase of resources use. I noticed that because the 
> freezing starts usually after opening a program like a terminal emulator 
> while there are already some programs running in other windows. It also 
> happened one time while trying to use my laptop while copying 23GBs of 
> data in a virtual console.

   If you can narrow it down to a repeatable sequence then you could 
more easily collect data from the occurrence. You could also try to push 
the system into an expected bad state and preemptively use ctrl+alt+f1 
to see if anything shows up on the terminal during failure.
   Are you able to reproduce without X running?
   FreeBSD has a tests suite which runs through a series of checks. It 
can take a long time to run, some tests will likely normally fail, and 
some won't run until additional packages are installed. You can have it 
run multiple tests in parallel which saves a massive amount of time but 
then test results become more inconsistent. I normally reboot after 
running it as it some some of the tests leave the system in a strange 
state. Keeping logs of the results could help identify regressions with 
new FreeBSD versions (tests and the OS both get reworked; its not always 
clear at a glance if it is a new but broken test vs an old test started 
failing that caused error count to go up) or help spot intermittent and 
special case hardware failures.
   There are other non-FreeBSD hardware tests that may be worth running. 
Some are stand alone tools that boot separately like memtest86 but 
others may need a Windows/Linux install to run from. You can look into 
tests that overclockers use to verify system stability; overclock or 
not, an unstable system may still show with the tools. These tools will 
stress the system in ways that the FreeBSD test suite does not. I even 
more recently found a system was intermittently but regularly crashing 
while restoring a ZFS backup (resolved by slightly lowering clock 
frequencies but most laptops don't offer such control).
   Maybe useful to monitor temperatures too if the issue is not 
repeatable. Physically cleaning even light dust off of hardware can also 
stabilize some intermittent hardware that isn't overheating.

> I leave some routine files here for you to check up if you feel the need 
> of it:
> 
> dmesg output: https://pastebin.com/C27M5K0F <https://pastebin.com/C27M5K0F>

   Probably just from incorrect shutdowns but I noticed disk mounting / 
listed pending error; could run through SMART tests/checks and make sure 
filesystem does fully clean up with correct shutdown or fsck.

> rc.conf: https://pastebin.com/9wX0sKsa <https://pastebin.com/9wX0sKsa>
> loader.conf: https://pastebin.com/wBrdL8NL <https://pastebin.com/wBrdL8NL>
> 
> As always, thanks to whoever that takes the time to help me with this.
>