[SOLVED] Debugging freezing laptop

JJK · 2016-12-29 11:20:47

Hello everybody,

My laptop has recently (about two weeks ago) started freezing sometimes, screen does not update, sound gets stuck in a short loop, cannot move mouse, does not react to keyboard input.

The crash happened most often while I was programming in Intellij. I noticed that disabling the vim shortcuts plugin made the crash less often. But it has also happened while intellij was not running. Firefox is also almost always on.

What I tried:
Look at the output of journalctl of the previous boot and see if there are any lines that are the same and look suspicious. The only thing I found was the line for starting daemon for power management. Further, the line "starting cleanup of temporary directories" occurred some times. I included the journalctl logs, maybe you can find other things.

Output top to file, then after the crash I looked at the top file to see if there was high memory or cpu usage. I could not see any of the two, just normal usage.

Test hardware using the builtin hardware diagnostics tool in the bios. No errors.

Relevant things about my system:
It is a dell latitude E5450
Most important applciations: i3, firefox, urxvt, zsh
I have enabled TLP

I am currently testing if TLP is the culprit, further I will check if it matters if I have the laptop on battery or on power line.

I hope you can help me debug this, in my opinion random freezing issues are the hardest.

I can provide other log files if you need them.

Log files:
(journalctl -xn 100 from 3 boots;
journalctl -kn100 from 3 boots (dmesg)
and pacman.log | tail -n 100)
https://drive.google.com/open?id=0B9Ru5 … WRMSkg0bms

Last edited by JJK (2017-06-28 05:20:07)

ewaller · 2016-12-29 18:26:02

Those logs seem strange and don't have the information for which I was looking. First off, is this an Intel based machine? and have you installed and configured your processor's microcode updates?

JJK · 2016-12-30 08:48:07

The laptop has an intel i5 5200U processor. I had never heard of configuring processor's microcode updates, but I checked the wiki, I will see if installing those helps anything. I checked the logs again, they don't look strange to me, though boot1_dmesg and boot2_dmesg are quite short, I might have printed it with n=20 instead of n=100. But the relevant information is there anyway.

Last edited by JJK (2016-12-30 08:49:29)

honestly · 2017-01-03 22:03:39

I have the same issue! It started a few days ago, since I'm lazy with system upgrades, I'd assume we have the same issue and I guess this means it's probably not (directly) a hardware problem.

My Dell Latitude E5450 (5450-6693, Core i5-5300U) that had been running fine since I bought it a year ago, dual-booting arch and windows 7 (uefi), developed random freezes that manifest with the same symptoms as yours. Additionally, the hard drive audibly powers off and when I reboot I get orphaned inodes in fsck.

Last edited by honestly (2017-01-03 22:07:25)

JJK · 2017-01-04 06:49:06

It might be the same issue, can you describe the symptoms that you noticed so that we can compare, further, do you have some useful logs to add, like the logs I posted. I'm particularly interested in your pacman.log since you recently downloaded the update that started this.

ngoonee · 2017-01-04 07:00:13

Do either of you get symptoms as with my thread https://bbs.archlinux.org/viewtopic.php … 5#p1680965 ?

In particular if you could check journalctl (and not the ones you've both posted which are very abbreviated) for errors?

JJK · 2017-01-07 14:51:27

@ngoonee I see you have a similar issue indeed, but I cannot find either of the errors you described. Also you mention that you think the issue is related to tmux/vim. I use both programs, vim more than tmux, but my laptop has frozen more often without vim running than with.

Also I have now installed microcode updates, let's see if the problem persists.

Btw, could It be an issue that I haven't configured a swap partition?

Last edited by JJK (2017-01-07 15:01:15)

honestly · 2017-01-07 20:32:39

For my issues see my post here: https://bbs.archlinux.org/viewtopic.php?id=221474

Here is my entire pacman.log: https://gist.github.com/duk3luk3/de0dd0 … 1f348e86d0

So switching from i3wm to openbox fixed the issue it seems... no idea how that works.

EDIT: So, it's working fine with i3wm and linux-lts. It must be a kernel issue that is only triggered by i3wm.

Last edited by honestly (2017-01-08 14:13:09)

Chocopuff · 2017-01-10 00:55:02

I have experienced the same problem over here as well. Kernel 4.8.13 stock arch and 4.9 custom compiled.
Each time, it has happened on XMonad when I switch workspaces. Each time that I am playing music, the music gets stuck in a repeating loop and I can't recover out of the freeze.

I don't remember if I used a compositor at the time of each crash. I use compton if that helps, but not all the time.

TLP might be the culprit, but more of I had powersave governor settings compiled into 4.9, but I didn't have that problem on 4.8.13 which did not have any weird settings.

CPU: Intel Core i5-5257U CPU @ 3.1GHz
8GB RAM, 8GB Swap

I'm not sure how to get the correct logs yet. Will post when I have.

I believe I see "starting cleanup of temporary directories" as well like JJK.

JJK · 2017-01-16 10:45:11

The problem still happens with TLP off. It happens a lot less when I'm not using intellij, so it might be related to resource-heavy appliciations. I added boot4 logs to the logs directory (see original post). I'm currently a bit out of options to test, will try with compositor off (I use xcompmgr).

JJK · 2017-02-02 19:58:33

'LTS kernel fixed the problem. This is a known issue with the latest kernels.

honestly · 2017-03-09 11:40:39

linux-lts has now been upgraded to 4.9, bringing this exciting issue to everyone using LTS.

And I've also confirmed that it still exists in 4.11.

Last edited by honestly (2017-03-09 16:47:37)

honestly · 2017-03-28 06:45:41

Someone should try to use kdump https://wiki.archlinux.org/index.php/Kdump to get kernel panic info out. I haven't had time to try it...

lilydjwg · 2017-03-28 06:56:37

honestly wrote:

Someone should try to use kdump https://wiki.archlinux.org/index.php/Kdump to get kernel panic info out. I haven't had time to try it...

I tried it. The kernel didn't crash for me.

honestly · 2017-04-05 18:51:10

Well, did you expect it to crash? Do you have the same problem? Also, how did you compile the kdump kernel? I haven't been able to get it to work...

lilydjwg · 2017-04-06 01:25:20

honestly wrote:

Well, did you expect it to crash? Do you have the same problem? Also, how did you compile the kdump kernel? I haven't been able to get it to work...

Not really, but unless it crashes, kdump won't work.

Mine is a Macbook Pro. It freezes from time to time, without any log, and doesn't response to sysrq requests.

I didn't notice my kernel support kdump until I tried it. If you want, you can get my kernel from the [archlinuxcn] repo (name: linux-lily).

honestly · 2017-04-06 06:59:31

Did you test whether kdump works by forcing a panic? (I'd recommend doing SysRq+REISUC to do that)

EDIT: So today I found a docking station with serial and set up serial console... of course it doesn't crash anymore.

Last edited by honestly (2017-04-06 11:40:21)

tlvince · 2017-04-07 22:07:51

JJK wrote:

'LTS kernel fixed the problem. This is a known issue with the latest kernels.

What was the last known-good kernel version that doesn't have this issue? Are there upstream reports?

honestly · 2017-04-08 12:20:21

I'm running 4.4.52-1-lts which is fine. The bug starts somewhere around kernel 4.8. There's no upstream bug report because nobody has been able to get a kernel panic or any other debug info out.

tlvince · 2017-04-08 13:36:47

Good to know. This is indeed a tricky one indeed.

honestly · 2017-04-09 10:43:55

I've been running it on a docking station so I can get a serial console for two days now - no freezes on 4.10.8-1-ARCH kernel. Of course, running without the docking station, I get the freeze within minutes.

It's like someone intentionally made this undebuggable...

If someone has an AMT-enabled E5450, they could use the AMT virtual serial port.

Last edited by honestly (2017-04-09 10:57:22)

honestly · 2017-04-18 15:24:07

I'm now trying to bisect the kernel using the linux-mainline aur package. I've successfully reproduced the freeze with 4.10-1 (as a control) which I had no reason to think would not freeze; as well as 4.6-1. That narrows the problem down significantly.

Could anyone please try to independently reproduce that they get the freeze on a Dell Latitude E5450 with kernel 4.6?

Last edited by honestly (2017-04-18 15:26:34)

stoelpi · 2017-04-19 18:32:00

I can reproduce the freeze on a Dell Latitude E6500 with Manjaro Linux 17.0.1 kernel 4.4.61-1

honestly · 2017-04-19 21:58:11

Well I'm running 4.4.52-1-lts (been running that for over a month without a problem), would be a funny coincidence if the bug was introduced that soon after... thanks, I'll move up in smaller increments then.

richin13 · 2017-04-22 04:01:20

I downgraded the kernel to version 4.4.52-1-lts and can confirm that the bug is there. My laptop is showing the symptons already
discussed in this thread. I noticed that after a freeze laptop still control the screen brightness (using fn key), however, other controls as touchpad or music play/pause don't work. The mouse cursor can be moved freely but clicking produce no results at all. If there is music playing it'll continue playing normally until it reaches the end of the playlist.

Almost every time the laptop freezes, Mozilla Firefox is open, but opening Firefox does not necessarily freezes the system.
I also use vim (really often) and tmux (not so often) and I did not configured a SWAP partition. (That was also mentioned by someone earlier in this thread). I enabled microcodes, as suggested below

➔ dmesg | grep microcode
[    0.000000] microcode: CPU0 microcode updated early to revision 0x29, date = 2013-06-12
[    0.130566] microcode: CPU1 microcode updated early to revision 0x29, date = 2013-06-12
[    0.424808] microcode: CPU0 sig=0x206a7, pf=0x10, revision=0x29
[    0.424867] microcode: CPU1 sig=0x206a7, pf=0x10, revision=0x29
[    0.424883] microcode: CPU2 sig=0x206a7, pf=0x10, revision=0x29
[    0.424902] microcode: CPU3 sig=0x206a7, pf=0x10, revision=0x29
[    0.425042] microcode: Microcode Update Driver: v2.01 <tigran@aivazian.fsnet.co.uk>, Peter Oruba

Today, I suffered several freezes at morning and decided to downgrade the kernel to version LTS 4.4.52-1. It seemed to work just fine until around 20:42 (GMT-6) when the system freezed (like 5 minutes after I unlocked the screen). The log corresponding to the seconds before that 'crash' are shown in the following pastebin https://pastebin.com/mcT8zvLM

The first line -- Reboot -- corresponds to the hard reset (using the power button) I used to recover from the freeze.
The second line shows the text 'nouveau 0000:01:00.0: fifo: INTR 00800000' colored in red

After that reboot, the system crashed again at 20:52 (GMT-6). The log https://pastebin.com/E4nedvvC corresponding to the secods before, shows the exact same second line as the previous (nouveau 0000:01:00.0: fifo: INTR 00800000)

-----------------------------------

My system specs:

Laptop Dell Latitute E6420
CPU Intel Core i5-2520M
RAM 8 Gb
Graphics NVIDIA GF119M [NVS 4200] (Using driver nouveau)
DE Cinnamon 3.28
SHELL zsh

EDIT: typo

Last edited by richin13 (2017-04-22 04:03:36)

Arch Linux

#1 2016-12-29 11:20:47

[SOLVED] Debugging freezing laptop

#2 2016-12-29 18:26:02

Re: [SOLVED] Debugging freezing laptop

#3 2016-12-30 08:48:07

Re: [SOLVED] Debugging freezing laptop

#4 2017-01-03 22:03:39

Re: [SOLVED] Debugging freezing laptop

#5 2017-01-04 06:49:06

Re: [SOLVED] Debugging freezing laptop

#6 2017-01-04 07:00:13

Re: [SOLVED] Debugging freezing laptop

#7 2017-01-07 14:51:27

Re: [SOLVED] Debugging freezing laptop

#8 2017-01-07 20:32:39

Re: [SOLVED] Debugging freezing laptop

#9 2017-01-10 00:55:02

Re: [SOLVED] Debugging freezing laptop

#10 2017-01-16 10:45:11

Re: [SOLVED] Debugging freezing laptop

#11 2017-02-02 19:58:33

Re: [SOLVED] Debugging freezing laptop

#12 2017-03-09 11:40:39

Re: [SOLVED] Debugging freezing laptop

#13 2017-03-28 06:45:41

Re: [SOLVED] Debugging freezing laptop

#14 2017-03-28 06:56:37

Re: [SOLVED] Debugging freezing laptop

#15 2017-04-05 18:51:10

Re: [SOLVED] Debugging freezing laptop

#16 2017-04-06 01:25:20

Re: [SOLVED] Debugging freezing laptop

#17 2017-04-06 06:59:31

Re: [SOLVED] Debugging freezing laptop

#18 2017-04-07 22:07:51

Re: [SOLVED] Debugging freezing laptop

#19 2017-04-08 12:20:21

Re: [SOLVED] Debugging freezing laptop

#20 2017-04-08 13:36:47

Re: [SOLVED] Debugging freezing laptop

#21 2017-04-09 10:43:55

Re: [SOLVED] Debugging freezing laptop

#22 2017-04-18 15:24:07

Re: [SOLVED] Debugging freezing laptop

#23 2017-04-19 18:32:00

Re: [SOLVED] Debugging freezing laptop

#24 2017-04-19 21:58:11

Re: [SOLVED] Debugging freezing laptop

#25 2017-04-22 04:01:20

Re: [SOLVED] Debugging freezing laptop

Board footer