System hangs at randome times (mostly while writing a lot of text)

seth · 2017-02-20 07:37:09

There're an active grab by fluxbox, try to restart it:

export DISPLAY=:0 # you need to do that once per ssh login ;-)
fluxbox-remote restart

You could also try killing (and restarting) fluxbox, but the stock startfluxbox script will in turn quit the session.

The cause might be evdev error with your "gaming mouse" , which (also) registers as keyboard - do you have a cheap 3-button mouse flying around to see whether the problem vanishes with the G600?
Also please provide you /etc/X11/xorg.conf* stuff (the file itself and all .d directory files) - this might be due to a bad config blindly applied to all "keyboard" devices.

ku_respawned · 2017-02-20 13:12:42

xorg.conf:

Section "InputClass"
    Identifier             "Keyboard Defaults"
    MatchIsKeyboard        "yes"
    Option                 "XkbLayout" "de,ru(winkeys)"
    Option "XkbOptions" "grp:ctrl_shift_toggle,grp_led:scroll"
EndSection

10-evdev.conf:

Section "InputClass"
        Identifier "evdev pointer catchall"
        MatchIsPointer "on"
        MatchDevicePath "/dev/input/event*"
        Driver "evdev"
EndSection

Section "InputClass"
        Identifier "evdev keyboard catchall"
        MatchIsKeyboard "on"
        MatchDevicePath "/dev/input/event*"
        Driver "evdev"
EndSection

Section "InputClass"
        Identifier "evdev touchpad catchall"
        MatchIsTouchpad "on"
        MatchDevicePath "/dev/input/event*"
        Driver "evdev"
EndSection

Section "InputClass"
        Identifier "evdev tablet catchall"
        MatchIsTablet "on"
        MatchDevicePath "/dev/input/event*"
        Driver "evdev"
EndSection

Section "InputClass"
        Identifier "evdev touchscreen catchall"
        MatchIsTouchscreen "on"
        MatchDevicePath "/dev/input/event*"
        Driver "evdev"
EndSection

20-nvidia.conf

Section "Device"
        Identifier "Default nvidia Device"
#        Driver "nouveau"
        Driver "nvidia"
        VendorName "NVIDIA Corporation"
#        Option "NoLogo" "True"
EndSection

Or should this be nvidia-lts? I never found any info on setting up the lts driver...

I also have a 10-quirks.conf file, which should be created by default.

I do have a spare 5-button mouse, which I didn't like that much, because it's pretty slow moving. Would it be OK if I connect it when another input freeze happens?

seth · 2017-02-20 16:20:14

Nvidia won't have any impact on this, you need nvidia-lts if you're running the linux-lts kernel. The driver name is always nvidia (it's the xorg driver, not the kernel one, so this is unrelated to linux or linux-lts)

The idea is to attach the old mouse *now* and wait (and wait and wait and wait) and see whether the problem still occurs.

Try adding

NoMatchProduct "Logitech Gaming Mouse G600"

to your "Keyboard Defaults" section.

ku_respawned · 2017-02-24 19:39:12

seth wrote:

Try adding
NoMatchProduct "Logitech Gaming Mouse G600"
to your "Keyboard Defaults" section.

This didn't help. Just had another input freeze, Do you need another Xorg-log?

seth · 2017-02-24 20:49:28

Does it still have the "(EE) BUG: triggered 'if (axnum >= dev->valuator->numAxes)'" lines?
In case, try whether the problem stops w/ the 5 button legacy-rodant and prepare to have to file a bug against evdev ;-)

ku_respawned · 2017-02-25 16:41:36

Yes, it has. What exactly does it mean? Gonna switch the mice today, found the acceleration wiki.

EDIT: Didn't need accelerattion, BTW, I remembered about a bug I had with that Logitech mouse - the cursor sometimes jumped 3 inches up or down. I can't remember whether I ever had it happen with Windows. Gonna tell if it'll happen again with the new mouse.

Last edited by ku_respawned (2017-02-25 17:02:18)

seth · 2017-02-25 20:22:29

Means the error is not induced by the keyboard match and we cannot rule out the error (and likely device reset) to be the cause of the active pointer grab in fluxbox.

The error will hopefully not appear with the "ordinary" mouse and you will hopefully no longer see stale pointer grabs then either.
In that case fixing the evdev bug (what should happen anyway) will also kill your actual trouble.

Bugs go here: https://bugs.freedesktop.org - component is input/evdev.

You can also try to use libinput as input driver for the mouse (instead of evdev)

ku_respawned · 2017-02-26 12:54:51

I'm not sure if it's off-topic, but i just had (for the second time in the last few days) a crashing of everything shortly after resuming from suspend. I shortly saw some error messages, then everything blacked out. I was able to reboot with Ctrl+Alt+Del.
Here's the whole journalctl log (skipped mandb) since the PC resumed: https://bpaste.net/show/2f5b1faeaee6

BTW, how do I find out, what disk ATA7 is?

EDIT: OOPS, the link expires after 1 day, reuploaded it with a month to expire: https://bpaste.net/show/02a96bee32a7

Last edited by ku_respawned (2017-02-26 13:08:21)

seth · 2017-02-26 13:13:46

cat /proc/scsi/scsi

- sata6

The "crash" is due to "Feb 26 13:12:17 entheogenesis systemd[1]: systemd-logind.service: Watchdog timeout (limit 3min)!" which killed systemd-logind, in turn X11 and everything in your session.
The timeout *seems* to be related to the I/O of that device, yes.

ku_respawned · 2017-02-26 14:01:14

OK, it's a HDD with a SWAP partition on it, it's connected to a "bad" (a somewhat loose connector) SATA cable, probably need to exchange it.

ku_respawned · 2017-02-26 15:49:50

Just had another another input freeze with the new mouse. This time no axnum errors. New Xorg.0.log here: https://bpaste.net/show/b8c66d1bfd84
When I did "fluxbox-remote restart", the areas around the window titles (workspaces and date/time) blanked out, nothing else changed. After "killall -9 fluxbox && startfluxbox" everything blacked out and I had to reboot (maybe should've used startx instead of startfluxbox?).

After I rebooted, my keyboard didn't work. Found following in the journalctl:

Feb 26 16:27:13 entheogenesis kernel: usb 4-1.5.1: new low-speed USB device number 5 using ehci-pci
Feb 26 16:27:13 entheogenesis kernel: hidraw: raw HID events driver (C) Jiri Kosina
Feb 26 16:27:13 entheogenesis kernel: usbhid 4-1.5.1:1.0: can't add hid device: -32
Feb 26 16:27:13 entheogenesis kernel: usbhid: probe of 4-1.5.1:1.0 failed with error -32
Feb 26 16:27:13 entheogenesis kernel: usbcore: registered new interface driver usbhid
Feb 26 16:27:13 entheogenesis kernel: usbhid: USB HID core driver
Feb 26 16:27:13 entheogenesis kernel: input: Logitech USB Keyboard as /devices/pci0000:00/0000:00:1d.0/usb4/4-1/4-1.5/4-1.5.1/4-1.5.1:1.1/0003:046D:C31D.0001/input/input6
Feb 26 16:27:13 entheogenesis kernel: hid-generic 0003:046D:C31D.0001: input,hidraw0: USB HID v1.10 Device [Logitech USB Keyboard] on usb-0000:00:1d.0-1.5.1/input1

Also, maybe those "screen blackouts" appear because of wrongly configured VGA console?

Feb 26 16:27:13 entheogenesis kernel: NVRM: Your system is not currently configured to drive a VGA console
Feb 26 16:27:13 entheogenesis kernel: NVRM: on the primary VGA device. The NVIDIA Linux graphics driver
Feb 26 16:27:13 entheogenesis kernel: NVRM: requires the use of a text-mode VGA console. Use of other console
Feb 26 16:27:13 entheogenesis kernel: NVRM: drivers including, but not limited to, vesafb, may result in
Feb 26 16:27:13 entheogenesis kernel: NVRM: corruption and stability problems, and is not supported.

seth · 2017-02-26 16:16:53

fluxbox is holding an active grab, no idea why the restart didn't work as expected but in the default config, killing fluxbox quits the session and you'll face the DM (or nothing if you use startx, maybe even along an autologin)

After I rebooted, my keyboard didn't work.

Not at all or just in X11 - did eg ctrl+alt+f2 work? Does it work now again?
In general, may we face more wonky connectors here? ;-)

framebuffer consoles are not supported w/ nvidia and random funky things can happen, but I doubt it's the cause here.
Fix it anyway: https://wiki.archlinux.org/index.php/GR … ramebuffer

ku_respawned · 2017-02-27 14:00:00

The keyboard didn't work when i got to login (in the console, I normally start X manually). It worked after reboot and the usb cable is connected properly, and everything works fine on the second KVM channel at the same time.
I really doubt it could be the connectors, the only thing i didn't try yet would be switching the ports (cables) on the KVM switch with the notebook.

Also, I disabled the framebuffer in grub, but I have a following parameter for the kernel: vga=795
Is it wrong as well? I don't really want the text to get super-huge.

Last edited by ku_respawned (2017-02-27 14:11:20)

seth · 2017-02-27 15:36:09

No, the vga parameter is fine. So the keyboard issue was a singleton (and you've a kvm)?
The post-kill behavior can be expected.

You want to try fluxbox from git? (not the fluxbox-git package from aur which is terribly dated)
I had experienced stale grabs w/ 1.3.7 and in 1.4 the grabbing will have fundamentally changed (async -> sync) and several things cooked up (and were fixed) what *might* have resolved this (haven't seen a hanging grab since months)

chekuthan · 2017-02-28 14:34:52

Im having the same problem

seth · 2017-02-28 15:00:53

"Same problem" being confirmed by inspecting the grabs in xorg.conf?

@ku_respawned
What I forgot: does it also happen w/o the kvm?

ku_respawned · 2017-02-28 21:32:22

It's sorta strange, my mouse has a multicolor led, which I always turn off, but from time to time it begins to glow red again. I'm not sure if I'm pressing the color change button (didn't notice it, it has a loud tactile click) or the mouse turns off and on again, but nothing in dmesg.

Input freezes come much more often now, up to 4 a day, gonna try Fluxbox-git tomorrow.

@seth: I use my notebooks pretty often and haven't tried it w/o the KVM because of that, it works on the other channel (switch it between 2 Windows 10 notebooks) at the same time as it's unuseable in Fluxbox, what would make sense (and what i didn't yet come to) would be switching the physical channels. But I thought, it was caused by Fluxbox? Or aren't you sure?

Last edited by ku_respawned (2017-02-28 21:41:26)

seth · 2017-03-01 08:21:53

The active grab is in fluxbox - the question is why this happens (so often for you) - and the typical answer would be "it missed a button release event".

ku_respawned · 2017-03-07 18:39:28

Haven't had any time to try fluxbox-git lately, too much to do outside the PC realm. Will probably try this week.

I had another problem lately - the mouse wheel of the new mouse sometimes jumps way too far, 3-4 times farther than it's supposed to. I'm not sure if it's hardware-based, but I haven't had it happen with the old mouse to that extent. (EDIT: Also sometimes in the wrong direction)

EDIT: Nevermind, I exchanged the mouse and now it works properly. Gonna almost definitly install fluxbox-git tomorrow.

EDIT2: Just installed fluxbox-git, pacman has a very cool -Rd --assume-installed feature, haven't heard about it before. Also updated the whole system. Should input freezes reappear, I will write a new message.

Last edited by ku_respawned (2017-03-13 16:17:46)

ku_respawned · 2017-03-13 18:17:59

So, another input freeze with fluxbox-git

dmesg (necessary?): https://bpaste.net/show/02ac4836256e

Xorg.0.log: https://bpaste.net/show/1ee39ce03d40

Also, a couple of days ago my PC would't suspend, it said, some program wouldn't freeze for 20 seconds. Had to shutdown and restart to be able to suspend.

The next thing I'm gonna try, well I need a new WM/DE. Any suggestions? I don't know anything about other WMs, i used KDE shortly before switching to fluxbox (which I used for over 10 years).
Don't know what to choose. I love fluxbox and do stuff like maximizing/restoring, resizing and moving the windows with the keyboard shortcuts, I'd like to keep being able to do that. And possibly lightweight and easy to use.

EDIT: Since fluxbox-git produced 3 input freezes in 1.5 hours, I installed lxde. So far everything works, but I'm missing conky - it's not displayed on lxde's desktop. Gonna research on whether it's possible to fix that. Yesterday I had no input freezes with it, but 5 hours runtime don't tell much.

Last edited by ku_respawned (2017-03-14 11:58:40)

seth · 2017-03-14 15:25:17

Same condition as before, mouse juggles, then the pointer is actively grabbed by fluxbox.
I shall try to reproduce this later today (pulling and re-plugging the mouse cable while pressing a button or so)

seth · 2017-03-14 21:53:08

Ok, simply pulling the mouse while moving a window just releases the mouse button (and thus the grab) - did you ever encounter this with your cheap ordinary mouse?
Can you spot a pattern on when the freeze occurs (like "when I move a window and press a keyboard shortcut" or some such)?

ku_respawned · 2017-03-16 15:39:44

The input freezes came usually after typing and then clicking on something, not dragging. I had 3 mice - the G600 and two Speedlink 5-button (I exchanged it after I got wheel problems).
I haven't had an input freeze with LXDE, but I had another "2 processes refusing to freeze" and another disk-freezing leading to system freeze, even though I exchanged the SATA cable.
Also, doublecmd froze a couple of times while I was moving data around on my HDDs. I have installed doublecmd-git, which didn't help. It didn't happen before I updated the system (doublecmd also got updated), but I haven't moved around that much stuff before, now I was tidying up everything. Will see if it happens again after I try to downgrade doublecmd.

Arch Linux

#26 2017-02-20 07:37:09

Re: System hangs at randome times (mostly while writing a lot of text)

#27 2017-02-20 13:12:42

Re: System hangs at randome times (mostly while writing a lot of text)

#28 2017-02-20 16:20:14

Re: System hangs at randome times (mostly while writing a lot of text)

#29 2017-02-24 19:39:12

Re: System hangs at randome times (mostly while writing a lot of text)

#30 2017-02-24 20:49:28

Re: System hangs at randome times (mostly while writing a lot of text)

#31 2017-02-25 16:41:36

Re: System hangs at randome times (mostly while writing a lot of text)

#32 2017-02-25 20:22:29

Re: System hangs at randome times (mostly while writing a lot of text)

#33 2017-02-26 12:54:51

Re: System hangs at randome times (mostly while writing a lot of text)

#34 2017-02-26 13:13:46

Re: System hangs at randome times (mostly while writing a lot of text)

#35 2017-02-26 14:01:14

Re: System hangs at randome times (mostly while writing a lot of text)

#36 2017-02-26 15:49:50

Re: System hangs at randome times (mostly while writing a lot of text)

#37 2017-02-26 16:16:53

Re: System hangs at randome times (mostly while writing a lot of text)

#38 2017-02-27 14:00:00

Re: System hangs at randome times (mostly while writing a lot of text)

#39 2017-02-27 15:36:09

Re: System hangs at randome times (mostly while writing a lot of text)

#40 2017-02-28 14:34:52

Re: System hangs at randome times (mostly while writing a lot of text)

#41 2017-02-28 15:00:53

Re: System hangs at randome times (mostly while writing a lot of text)

#42 2017-02-28 21:32:22

Re: System hangs at randome times (mostly while writing a lot of text)

#43 2017-03-01 08:21:53

Re: System hangs at randome times (mostly while writing a lot of text)

#44 2017-03-07 18:39:28

Re: System hangs at randome times (mostly while writing a lot of text)

#45 2017-03-13 18:17:59

Re: System hangs at randome times (mostly while writing a lot of text)

#46 2017-03-14 15:25:17

Re: System hangs at randome times (mostly while writing a lot of text)

#47 2017-03-14 21:53:08

Re: System hangs at randome times (mostly while writing a lot of text)

#48 2017-03-16 15:39:44

Re: System hangs at randome times (mostly while writing a lot of text)

Board footer