Today I got a total system freeze in Manjaro with kernel 4.19.
I’ve been having this random freezes for months already, but today I kind of have enough stamina to try to ask for help (again) to try to solve this.
This is the journalctl output (the important parts):
Aug 18 16:58:38 ruben-manjaro kwin_x11[1163]: qt.qpa.xcb: QXcbConnection: XCB error: 3 (BadWindow), sequence: 43389, resource id: 73400322, major code: 15 (QueryTree), minor code: 0
Aug 18 16:58:38 ruben-manjaro systemd[953]: app-mpv-162fba37af3440039387885610e619c4.scope: Consumed 57.014s CPU time.
Aug 18 16:58:42 ruben-manjaro rtkit-daemon[1273]: Supervising 6 threads of 3 processes of 1 users.
Aug 18 16:58:42 ruben-manjaro rtkit-daemon[1273]: Supervising 6 threads of 3 processes of 1 users.
Aug 18 16:58:55 ruben-manjaro plasmashell[1210]: libpng warning: iCCP: known incorrect sRGB profile
Aug 18 16:58:55 ruben-manjaro plasmashell[1210]: libpng warning: iCCP: too many profiles
Aug 18 16:59:23 ruben-manjaro kernel: NVRM: GPU at PCI:0000:01:00: GPU-4b9f2d81-7e36-49e8-cefc-16c0d5698972
Aug 18 16:59:23 ruben-manjaro kernel: NVRM: Xid (PCI:0000:01:00): 8, pid=1163, Channel 00000020
Aug 18 17:00:06 ruben-manjaro systemd-logind[599]: Power key pressed.
Aug 18 17:01:00 ruben-manjaro audit[2202899]: USER_ACCT pid=2202899 uid=0 auid=4294967295 ses=4294967295 msg='op=PAM:accounting grantors=pam_unix acct="ruben" exe="/usr/bin/smbd" hostname=192.168.0.11 addr=192.168.0.11 terminal=samba res=success'
Aug 18 17:01:00 ruben-manjaro smbd[2202899]: pam_unix(samba:session): session opened for user ruben(uid=1000) by (uid=0)
Aug 18 17:01:00 ruben-manjaro audit[2202899]: USER_START pid=2202899 uid=0 auid=4294967295 ses=4294967295 msg='op=PAM:session_open grantors=pam_unix acct="ruben" exe="/usr/bin/smbd" hostname=192.168.0.11 addr=192.168.0.11 terminal=smb/1838240609 res=success'
Aug 18 17:01:00 ruben-manjaro kernel: audit: type=1101 audit(1629298860.638:77): pid=2202899 uid=0 auid=4294967295 ses=4294967295 msg='op=PAM:accounting grantors=pam_unix acct="ruben" exe="/usr/bin/smbd" hostname=192.168.0.11 addr=192.168.0.11 terminal=samba res=success'
Aug 18 17:01:00 ruben-manjaro kernel: audit: type=1105 audit(1629298860.638:78): pid=2202899 uid=0 auid=4294967295 ses=4294967295 msg='op=PAM:session_open grantors=pam_unix acct="ruben" exe="/usr/bin/smbd" hostname=192.168.0.11 addr=192.168.0.11 terminal=smb/1838240609 res=success'
Aug 18 17:01:01 ruben-manjaro CROND[2202901]: (root) CMD (run-parts /etc/cron.hourly)
Aug 18 17:01:01 ruben-manjaro CROND[2202900]: (root) CMDEND (run-parts /etc/cron.hourly)
As you can see, I could use samba from my phone to connect with the PC, so it was not “absolutely” freezed, but I couldn’t use KDE connect, so it was not “only a visual problem”. Also the power button didn’t work, and neither did keyboard shortcuts.
The most important thing is probably the NVRM, which seems like a problem of Nvidia drivers (idk), and I have to say that the freeze was after opening a pip video on Firefox, it’s something that I can do without problem, but this time for some reason (maybe at random), my PC crashed after doing so.
I have this freezes with all kernels, so please don’t ask me to try kernel 5.10 or 5.xx because it’s the same.
So, anyone know what could be the problem or this is not enough information? is there any solution already or do I have to keep having this freezes “because Nvidia is the devil and that’s all”?.
Edit (Important): After some research I noticed that the problem could be that after closing mpv, the process “rtkit-daemon” started to change scheduling policy of some processes, and while doing so, is when I tried to pip on Firefox and then the crash happened, that’s why I didn’t get a “Successfully made thread…” message from rtkit-daemon. This would confirm the problem I already knew about the Arch kernel scheduler for Haswell CPUs, but it’s just a guess.
CPU: Intel i7-4790k Haswell
GPU: Nvidia GTX 1060 6gb
Video drivers: intel-nvidia Prime 470.57.02
DE: KDE Plasma 5.22.4 (X11)
Kernel: 4.19.201-1