You are not logged in.
Pages: 1
It's been a while, i think from 2018 on after a kernel update I guess, my system freezes completely at least once a day, mostly while browsing in the internet. Haven't noticed this in Wndows. Mouse not usable, hard drive LED turns immidiately off and i have to hard reset. I wouldn't want to install an old kernel, but will probably give it a try.
I have an old setup:
Intel Core 2 Quad Q9400
Nvidia 9600GT 512MB
Gigabyte GA-EP45-UD3LR
8 GB DDR2
and there is attached an M-Audio interface via USB which worked out of the box
I have tried the following:
- memtest OK
- changed nvidia's driver to noveau
- changing governor from schedutil to conservative
- booting with intel_idle.max_cstate=1
- moved from harddrive to ssd in the meantime
- Kernel is up2date
- intel-ucode (microcode) is up2date and loaded via grub
- there are no bios updates available since 2010
Journalctl doesn't really show anything special right before the crash. Crashed at 14:54
Last lines of boot:
Jul 01 14:37:32 home-pc kernel: audit: type=1334 audit(1593607052.695:71): prog-id=19 op=UNLOAD
Jul 01 14:37:32 home-pc kernel: audit: type=1334 audit(1593607052.695:72): prog-id=18 op=UNLOAD
Jul 01 14:47:31 home-pc dbus-daemon[548]: [session uid=1000 pid=548] Activating service name='org.freedesktop.thumbnails.Cache1' requested by ':1.19' (uid=1000 pid=1333 comm="thunar ")
Jul 01 14:47:31 home-pc tumblerd[6600]: Failed to load plugin "tumbler-ffmpeg-thumbnailer.so": libffmpegthumbnailer.so.4: cannot open shared object file: No such file or directory
Jul 01 14:47:31 home-pc tumblerd[6600]: Failed to load plugin "tumbler-raw-thumbnailer.so": libopenrawgnome.so.7: cannot open shared object file: No such file or directory
Jul 01 14:47:32 home-pc org.freedesktop.thumbnails.Cache1[6600]: Registered thumbnailer heif-thumbnailer -s %s %i %o
Jul 01 14:47:32 home-pc org.freedesktop.thumbnails.Cache1[6600]: Registered thumbnailer gsf-office-thumbnailer -i %i -o %o -s %s
Jul 01 14:47:32 home-pc org.freedesktop.thumbnails.Cache1[6600]: Registered thumbnailer /usr/bin/gdk-pixbuf-thumbnailer -s %s %u %o
Jul 01 14:47:32 home-pc org.freedesktop.thumbnails.Cache1[6600]: Registered thumbnailer /usr/bin/gdk-pixbuf-thumbnailer -s %s %u %o
Jul 01 14:47:32 home-pc dbus-daemon[548]: [session uid=1000 pid=548] Successfully activated service 'org.freedesktop.thumbnails.Cache1'
Jul 01 14:48:16 home-pc kernel: hrtimer: interrupt took 3756704 ns
Jul 01 14:49:46 home-pc audit[6703]: USER_AUTH pid=6703 uid=1000 auid=1000 ses=1 msg='op=PAM:authentication grantors=pam_unix acct="root" exe="/usr/bin/su" hostname=home-pc addr=? terminal=pts/2 res=success'
Jul 01 14:49:46 home-pc audit[6703]: USER_ACCT pid=6703 uid=1000 auid=1000 ses=1 msg='op=PAM:accounting grantors=pam_unix acct="root" exe="/usr/bin/su" hostname=home-pc addr=? terminal=pts/2 res=success'
Jul 01 14:49:46 home-pc su[6703]: (to root) anon on pts/2
Jul 01 14:49:46 home-pc audit[6703]: CRED_ACQ pid=6703 uid=1000 auid=1000 ses=1 msg='op=PAM:setcred grantors=pam_unix acct="root" exe="/usr/bin/su" hostname=home-pc addr=? terminal=pts/2 res=success'
Jul 01 14:49:46 home-pc audit[6703]: USER_START pid=6703 uid=1000 auid=1000 ses=1 msg='op=PAM:session_open grantors=pam_unix acct="root" exe="/usr/bin/su" hostname=home-pc addr=? terminal=pts/2 res=success'
Jul 01 14:49:46 home-pc kernel: audit: type=1100 audit(1593607786.531:73): pid=6703 uid=1000 auid=1000 ses=1 msg='op=PAM:authentication grantors=pam_unix acct="root" exe="/usr/bin/su" hostname=home-pc addr=? terminal=pts/2 res=success'
Jul 01 14:49:46 home-pc kernel: audit: type=1101 audit(1593607786.531:74): pid=6703 uid=1000 auid=1000 ses=1 msg='op=PAM:accounting grantors=pam_unix acct="root" exe="/usr/bin/su" hostname=home-pc addr=? terminal=pts/2 res=success'
Jul 01 14:49:46 home-pc kernel: audit: type=1103 audit(1593607786.531:75): pid=6703 uid=1000 auid=1000 ses=1 msg='op=PAM:setcred grantors=pam_unix acct="root" exe="/usr/bin/su" hostname=home-pc addr=? terminal=pts/2 res=success'
Jul 01 14:49:46 home-pc kernel: audit: type=1105 audit(1593607786.531:76): pid=6703 uid=1000 auid=1000 ses=1 msg='op=PAM:session_open grantors=pam_unix acct="root" exe="/usr/bin/su" hostname=home-pc addr=? terminal=pts/2 res=success'
Jul 01 14:49:46 home-pc su[6703]: pam_unix(su:session): session opened for user root by (uid=1000)Full boot log:
https://pastebin.com/BJTGce12
lsmod:
https://pastebin.com/2y0b57dF
lsusb + lspci:
https://pastebin.com/azBaTCHg
Last edited by d0ok (2020-07-01 14:41:44)
Offline
As loqs pointed out in another kernel related thread, but completely different hardware setup than yours (AMD Ryzen), try to disable the virtualbox kernel modules.
The vbox modules (vboxdrv, vboxnetadp, voxnetflt) taint the kernel and although some development happened with at least a module (vboxsf) now directly upstreamed and accepted into the kernel not so long ago, the rest is still externally developed by Oracle and could lead to crashes. In case you report anything upstream to the kernel developers, they would indeed point at the tainted state of the kernel, the vbox modules, and insist that you try to reproduce the issue without anything tainting the kernel.
It's been a while, i think from 2018 on after a kernel update I guess, my system freezes completely at least once a day, mostly while browsing in the internet.
So, basically all Arch linux kernels you ever installed since 2018 led to a frozen system?
*) Have you tried a different kernel flavor, lts or hardened?
*) Not sure about the switch from nvidia blob to nouveau leads to a more stable system, since nouveau is still under heavy development
I would suggest, that you update the kernel to latest version (linux 5.7.7.arch1-1), disable the virtualbox modules and also try to boot without Intel's microcode (intel-ucode).
If that does make no difference then try at least the Arch linux-lts kernel without vbox modules and no intel-ucode loaded.
[As a side note:]
It should make no difference to your issue, but deactivate support you do not need in your BIOS: Floppy, parport, etc. The full boot log revealed the floppy support ;-)
Offline
I now removed vbox-modules completely because at the moment I dont need it, I also changed clocksource to acpi_pm with the kernel parameter because of the log-message posted above. After a reboot it crashed again after just some minutes. I must admit though, that i haven't checked lsmod if it really wasn't loaded. But I highly doubt that it was loaded because right after i removed vbox I updated the kernel which forced depmod to run. I'll check it again tomorrow. Yesterday i was using the computer the whole day and noticed that two times everything was frozen for just about 3 to 7 seconds (beside the complete freezes which require a hard reset) and then runned again normal. Unfortunately there were no logs right after the freeze. Only these NMI logs in dmesg, which occured about 30 minutes before the freeze:
INFO: NMI handler (nmi_cpu_backtrace_handler) took too long to run: 1.786 msecs
INFO: NMI handler (perf_event_nmi_handler) took too long to run: 1.947 msecsBut I don't think that this has anything to do with it. Also the CPU fan runs faster after a complete freeze until i hard reset. Seems a bit that there is a cpu-deadlock or something like that.
So, basically all Arch linux kernels you ever installed since 2018 led to a frozen system?
I did not install every release. By far not. But yes, every installed kernel afterwards had this issue. I could not remeber when the problem startet to occur though. I've run kernel 4.19 LTS (with intel-ucode) today and got no problems at all so far. It even runs smoother than 5.*
I will try the next days or weeks to disable floppy and run the 5.* kernel without the microcode updates and will report afterwards.
Last edited by d0ok (2020-07-03 21:45:54)
Offline
I have a few Core 2 Quad systems and have found changing cpupower - https://wiki.archlinux.org/index.php/CP … cy_scaling - to use...
GOVERNOR="ondemand"and also changing to the hpet clocksource makes things much more stable.
Booting with the following...
clocksource=hpet intel_pstate=disablechanges the clocksource then makes cpu scaling fall back to use acpi-cpufreq with ondemand scaling.
Check after booting with...
cpupower frequency-infoand look at the "driver" line and also...
cat /sys/devices/system/clocksource/clocksource0/current_clocksourceto check the clocksource in use.
You could give that a go and see how it goes for your system.
Cheers
Paul.
Offline
I've tried all, but nothing worked. I reinstalled Arch now and it seems to work. Thanks.
Offline
Pages: 1