Pop! os does not restart properly after freezing
 

Hei everyone,
I posted this issue on the forum of pop os few weeks ago but never got any answer. So I am trying my chance here.
So, sometimes when I play a game a bit demanding for my computer it will freeze. I guess it has something to do with the RAM. That thing is annoying but OK, I need to accept that my laptop isn't completely new anymore :)
The main problem comes after, when switching off and on the laptop by pressing the ON/OFF button, it does not work. More specifically, it starts, but it will stay stuck at the last black screen before the logging in screen. The solution I found is to start Pop os using the dongle, power off from the menu, remove the dongle, restart. Then it works... But it isn't very convenient.
I do not really know where I should look... I can say that when I used ubuntu before I did not have this issue after freezing. Also, I realised that since I installed pop os 20.10 it freezes too when I drag a file from the desktop to an open folder (Topic for another thread?). That means I can test any solution you give me with that (cause otherwise I would have to wait that it randomly freezes in a game).
I run pop os 20.10 (but had the same problem with 20.04) on a ACER aspire V nitro. Intel Core i7 and a GeForce GTX 960M graphic card.
Thanks a lot for your help.

Have you looked in the log files?

You may not find anything to help with the freezing itself--it may be freezing before it can log anything, though it's worth a look--but there's a good chance you will find some hints as to what's going wrong in the boot process. https://wiki.archlinux.org/index.php/Systemd/Journal

The Arch Wiki has a good article about how to use SystemD logs. You should also find log files in tradition format, as most distros I've seen use rsyslog to pipe SystemD log data into /var/log.

You don't tell us which games and which platform you play them on.

You say you suspect RAM yet you don't tell us how much you have there.

Which GPU driver are you using? I'm sure POP_OS has an "Additional drivers" utility (or sth similar) that will allow you to install the NVIDIA driver, which will be better for gaming.

Thanks for your answers. So here:
@ondoho
I usually play on steam. And it has happened on several games (Tomb raider, Hades, Desperados III,...). The reason why I suspect the memory is that it usually happens when it needs to load something, or when there is a visually demanding moment. The ventilation of the laptop goes crazy, and it freezes. I have 7.7 GiB and I already run the NVIDIA drivers.

@frankbell
Nope. I did not know such a thing existed :) But now I did. I went into var/log; and found several types of log. So I reproduced the crash using the dragging of a file from the desktop to a directory (it looks same, but I am actually not sure if it is the same crash). I attached the syslog (split in 3 files because too big), and the kernlog.
The timing is as such:
1423: Crash + switch off
1425: Switch on (does not work)
1426: switch off
1427: switch on, boot on dongle. Power off from USB-based pop os. Remove dongle.
1428: Switch on (works).
As you can see there is nothing in the log between 1423 and 1428... It seems that there is some info about the freezing itself though. But I have no idea what it means.

The last 4 lines of your kernlog would suggest that Teams does not play well with Steam.

You mean these ones?
I see the link to teams (written in it) but how do you know that steam is involved? On top of that steam is not being opened at startup, so it should not be involved in anything at that stage.

Because you wrote it happens when you play (steam) games.

Disable Teams autostarting after login.

I am really sceptical that teams has any role into this. I had that issue before using teams at all.

Disabling teams did not change anything.

MeinzBeur,

If you did not have this problem with Ubuntu, why not use Ubuntu instead?

Because there are many other features from pop os that are better than ubuntu (I know this is subjective).

Looking at your kernel log in post #4, the output you posted in post #6 is talking about a segfault (Segmentation Fault) that's occurring with the SGI_video_sync extension, that appears to be caused by "teams". It's a software bug, not a hardware problem with your RAM, although a segfault does mean that a program has attempted to access memory it shouldn't be trying to access. So the only way to fix it, is with a software fix, so you'll have to report it to the relevant developers and have them fix the code responsible and release an update you can install that fixes it. Or use Ubuntu as beachboy2 has said if you didn't have the problem on that.

Is the freezing occurring exactly when the segfaults happen?

EDIT: If it's happening with more than just one program, it's possible it might be a driver bug too. Are you using NVIDIA's proprietary drivers, or the open-source driver? Which driver were you using in Ubuntu?

You can check with the following command at the terminal:

Thanks for the detailed answer. So:
* Ok, that segfault thingy is something new to me. I get the issue, but I don't think there is anyway to get it fix, as teams full name is "Microsoft teams". Wish me good luck to get the developers to fix it :)
* No, the segfaults seems non related to the freezing; for at least 2 reasons: (i) Segfault happens when teams opens, regardless of the computer freezing before or after and (ii) I had freezing issues before I even installed teams on my computer.

As much as it puzzles me I don't think the segfault creates any issue in the running of my computer, and I am forced to use it for my work. I don't think I can get microsoft to change their software.

So seems to me that we don't have a solution to my issue. Maybe the attached picture can help. When I restart after the freeze, it sometimes blocks on this screen instead of a black screen. I guess that could mean that there is actually more than reason why my computer freezes.

EDIT: just saw your edit to the previous post. I use NVIDIA drivers 455. I had NVIDIA also when I was on Ubuntu, not sure which version though. I moved to pop os in february this year, so I guess the latest NVIDIA primers at that time.


NEW EDIT: Wait, no. Actually this is maybe weird:
* In the pop shop I can see that NVIDIA drivers 455 are installed.
* In the NVIDIA X server settings it tells me that 440.95.01 is in use.
* and the output of lspi is:
That does not seem like NVIDIA to me, right?

From the research I've done, I'm not sure that the segfault and the error messages in the screenshot you posted are related at all, and therefore I'm thinking that they are two different problems at this point. Although when I replied before, I wasn't sure if you were talking about Microsoft Teams or something else called "teams", but I did find this when I replied before, so it seems they already know about the bug.

The reason the lspci output says Intel graphics while you have NVIDIA graphics drivers installed is probably because you have a NVIDIA Optimus setup. In relation to the error messages [in your posted screenshot] from ACPI, my research suggests it maybe that your graphics hardware/setup may have problems with KMS (Kernel Mode Setting). You can try disabling KMS and see it that helps.

Steps to do that:

1. When your machine boots and you see the GRUB menu asking you to select a choice (eg. a normal boot, recovery mode, etc), select whatever the normal boot option you use to start the system and press "e".
2. Use the down arrow key on your keyboard to scroll down to the line that starts with the word "linux".
3. Use the right arrow key to move the cursor to the end of that same line, then press the spacebar key.
4. Type in nomodeset then press CTRL+X (not including the plus sign).

Although, note that this will only disable KMS for that session, and not permanently.

For example:

The above line will look a little different on your machine, particularly the version of your kernel and the UUID for your root partition.

Yes, I agree with you regarding the segfault issue. Well, obviously they know about the issue since some months... and obviously it will remain for a while more :)

Ok. I had never heard about this Optimus setup before.

For the whole KMS setting stuff. I had already looked into it a while ago. There are a bunch of forum posts on the topic. I can't remember exactly what I did but it obviously did not work at that time. "Luckily" my computer crashed while playing yesterday evening, so I tested your fix this morning:

I attached the screen after pressing "e" in GRUB. Maybe you can find some info there.
I added `nomodeset` in the line and pressed ctrl + X.
The screen gets back to the one in the previous screenshot (platform device creation failed: -16 + other things) and quickly after this line is appended to the screen:
Then I get to a next screen where there is only a lower dash blinking, and after 1 min maybe this appears:
Not sure if this is linked to the problem.Probably not (https://ubuntuforums.org/showthread.php?t=2437409). In any case nothing happened in the next 5 mins, so I just rebooted the computer on the pop os dongle to get around the issue (as described in earlier posts).

Does that trigger any guess?