Gnome desktop freezing

[Canuck62]

I am running Debian 8.11 and have encountered an issue. After login to Gnome (version 3.14.1), when attempting to run applications, the screen is corrupted by black sections and the session freezes after a short time. The mouse moves on the screen, but the mouse buttons won't drive anything. The keyboard won't respond and I can't break out. This symptoms are the same for multiple users. The only way to get the system back is to hard power down. The latest attempt led to a full system freeze (including mouse movement)

The symptoms developed after I was upgrading the system at the same time as running a backup (yes, this is a bad idea, but /home and /data are both separate partitions from / and I did not backup / until the update was complete. backup is cpio to external hdd).

When I boot to recovery mode, the system is stable and I can run diagnostics. I have run another apt update and apt install from recovery mode and the symptoms persist. I have checked aptitude and dpkg outputs and all seems clean. I don't wish to upgrade to stretch from a corrupt jessie system. I have sufficient packages installed (including some applications under Wine) that I am loathe to do a scratch install.
Here is the error from the syslog which seem to show a definitive error.

Code: Select all

Jun 26 18:13:18 svr-name kernel: [  213.304035] ------------[ cut here ]------------
Jun 26 18:13:18 svr-name kernel: [  213.304062] kernel BUG at /build/linux-oNEiVW/linux-3.16.68/drivers/gpu/drm/i915/i915_gem.c:2204!
Jun 26 18:13:18 svr-name kernel: [  213.304095] invalid opcode: 0000 [#1] SMP 
Jun 26 18:13:18 svr-name kernel: [  213.304117] Modules linked in: binfmt_misc snd_hrtimer snd_seq_midi snd_seq_midi_event snd_rawmidi snd_seq snd_seq_device cfg80211 rfkill nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc joydev snd_hda_codec_hdmi snd_hda_codec_realtek intel_powerclamp coretemp snd_hda_codec_generic kvm_intel snd_hda_intel snd_hda_controller snd_hda_codec i915 snd_hwdep snd_pcm_oss evdev kvm iTCO_wdt iTCO_vendor_support shpchp snd_mixer_oss snd_pcm pcspkr serio_raw lpc_ich mfd_core video wmi snd_timer drm_kms_helper snd soundcore drm i2c_algo_bit button acpi_cpufreq processor fuse parport_pc ppdev lp parport autofs4 ext4 crc16 mbcache jbd2 sg sd_mod crc_t10dif crct10dif_generic sr_mod cdrom crct10dif_common hid_generic usbhid hid usb_storage crc32c_intel ahci psmouse libahci i2c_i801 i2c_core libata ehci_pci ehci_hcd scsi_mod r8169 mii usbcore thermal usb_common thermal_sys
Jun 26 18:13:18 svr-name kernel: [  213.304623] CPU: 2 PID: 719 Comm: Xorg Not tainted 3.16.0-9-amd64 #1 Debian 3.16.68-2
Jun 26 18:13:18 svr-name kernel: [  213.304648] Hardware name: Acer Aspire M3910/Aspire M3910, BIOS P01-A3         05/17/2010
Jun 26 18:13:18 svr-name kernel: [  213.304679] task: ffff8800b4719410 ti: ffff8800b4484000 task.ti: ffff8800b4484000
Jun 26 18:13:18 svr-name kernel: [  213.304704] RIP: 0010:[<ffffffffa03fdf99>]  [<ffffffffa03fdf99>] i915_gem_object_move_to_inactive+0x149/0x150 [i915]
Jun 26 18:13:18 svr-name kernel: [  213.304756] RSP: 0018:ffff8800b4487c48  EFLAGS: 00010202
Jun 26 18:13:18 svr-name kernel: [  213.304775] RAX: ffff8800b737f800 RBX: ffff8800b7311928 RCX: ffff8800b4e71d68
Jun 26 18:13:18 svr-name kernel: [  213.304801] RDX: ffff8800b7311a08 RSI: ffff8800b4e719c0 RDI: ffff8800b4e71c80
Jun 26 18:13:18 svr-name kernel: [  213.304824] RBP: ffff8800b7311a08 R08: ffff880073a4b398 R09: ffff8800b7317af8
Jun 26 18:13:18 svr-name kernel: [  213.304847] R10: ffff8800b7310000 R11: 0000000000000000 R12: ffff8800b7311928
Jun 26 18:13:18 svr-name kernel: [  213.304873] R13: 00000000000026ae R14: ffff8800b737f800 R15: ffff880036e53a00
Jun 26 18:13:18 svr-name kernel: [  213.304896] FS:  00007f2d5150a980(0000) GS:ffff8800bb280000(0000) knlGS:0000000000000000
Jun 26 18:13:18 svr-name kernel: [  213.304925] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
Jun 26 18:13:18 svr-name kernel: [  213.304945] CR2: 000055c64d89d000 CR3: 00000000b447e000 CR4: 0000000000020770
Jun 26 18:13:18 svr-name kernel: [  213.304968] Stack:
Jun 26 18:13:18 svr-name kernel: [  213.304977]  ffffffffa03ff4fd ffff8800b7311a18 ffff88006346b000 0000000000000000
Jun 26 18:13:18 svr-name kernel: [  213.305012]  ffff8800b7311928 0000000000000020 ffff8800b737f800 ffff880036e53a00
Jun 26 18:13:18 svr-name kernel: [  213.305044]  ffffffffa046b863 ffff8800b4487da5 ffff8800b7317a08 ffff8800b4487e58
Jun 26 18:13:18 svr-name kernel: [  213.305078] Call Trace:
Jun 26 18:13:18 svr-name kernel: [  213.305097]  [<ffffffffa03ff4fd>] ? i915_gem_retire_requests_ring+0x5d/0x150 [i915]
Jun 26 18:13:18 svr-name kernel: [  213.305131]  [<ffffffffa046b863>] ? i915_gem_execbuffer_reserve+0x3d/0x2ee [i915]
Jun 26 18:13:18 svr-name kernel: [  213.305162]  [<ffffffffa03f818e>] ? i915_gem_do_execbuffer.isra.24+0x7fe/0x1470 [i915]
Jun 26 18:13:18 svr-name kernel: [  213.305197]  [<ffffffffa0404f29>] ? i915_gem_pwrite_ioctl+0x1a9/0x980 [i915]
Jun 26 18:13:18 svr-name kernel: [  213.305227]  [<ffffffffa03f92c1>] ? i915_gem_execbuffer2+0xb1/0x2b0 [i915]
Jun 26 18:13:18 svr-name kernel: [  213.305259]  [<ffffffffa02a18c3>] ? drm_ioctl+0x1d3/0x5c0 [drm]
Jun 26 18:13:18 svr-name kernel: [  213.305283]  [<ffffffff81092ef1>] ? enqueue_hrtimer+0x21/0xa0
Jun 26 18:13:18 svr-name kernel: [  213.305305]  [<ffffffff811ce9b3>] ? do_vfs_ioctl+0x2e3/0x4c0
Jun 26 18:13:18 svr-name kernel: [  213.305328]  [<ffffffff811cec11>] ? SyS_ioctl+0x81/0xa0
Jun 26 18:13:18 svr-name kernel: [  213.305355]  [<ffffffff8153e4dc>] ? system_call_fast_compare_end+0x1c/0x21
Jun 26 18:13:18 svr-name kernel: [  213.305382] Code: 01 00 00 00 00 00 00 c7 87 98 00 00 00 00 00 00 00 c7 87 40 01 00 00 00 00 00 00 74 0c f0 83 2f 01 74 01 c3 e9 e9 4e ea ff f3 c3 <0f> 0b 0f 0b 0f 1f 00 66 66 66 66 90 48 89 f8 48 8b bf 30 01 00 
Jun 26 18:13:18 svr-name kernel: [  213.305644] RIP  [<ffffffffa03fdf99>] i915_gem_object_move_to_inactive+0x149/0x150 [i915]
Jun 26 18:13:18 svr-name kernel: [  213.305679]  RSP <ffff8800b4487c48>
Jun 26 18:13:18 svr-name kernel: [  213.313651] ---[ end trace 01a49fab1673aa20 ]---
Jun 26 18:13:58 svr-name kernel: [  253.905059] SysRq : Keyboard mode set to system default

Any thoughts?

[stevepusser]

If you install inxi and give us the result of

Code: Select all

inxi -Fxxz

we would then know a lot more about your system and what drivers the hardware is using.

[Canuck62]

As an additional note, I have installed xfce and then logged into that as a shell. This shows similar corruption (missing letters in labels, icons invisible until mouse over) and the system crashed. This is making me think Xorg could be an issue.

Here is the inxi output

Code: Select all

System:    Host: Carinya-1 Kernel: 3.16.0-9-amd64 x86_64 (64 bit gcc: 4.9.2) Console: tty 
           Distro: Debian GNU/Linux 8 
Machine:   Mobo: Acer model: Aspire M3910 Bios: American Megatrends v: P01-A3 date: 05/17/2010
CPU:       Dual core Intel Core i3 550 (-HT-MCP-) cache: 4096 KB
           flags: (lm nx sse sse2 sse3 sse4_1 sse4_2 ssse3 vmx) bmips: 12768 
           Clock Speeds: 1: 1200 MHz 2: 1200 MHz 3: 1200 MHz 4: 1733 MHz
Graphics:  Card: Intel Core Processor Integrated Graphics Controller bus-ID: 00:02.0 chip-ID: 8086:0042
           Display Server: X.org 1.16.4 drivers: intel (unloaded: fbdev,vesa)
           tty size: 240x67 Advanced Data: N/A for root out of X
Audio:     Card Intel 5 Series/3400 Series High Definition Audio 
           driver: snd_hda_intel bus-ID: 00:1b.0 chip-ID: 8086:3b56 
           Sound: Advanced Linux Sound Architecture v: k3.16.0-9-amd64
Network:   Card: Realtek RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller
           driver: r8169 v: 2.3LK-NAPI port: e800 bus-ID: 02:00.0 chip-ID: 10ec:8168
           IF: eth0 state: up speed: 100 Mbps duplex: full mac: <filter>
Drives:    HDD Total Size: 1004.2GB (26.3% used)
           ID-1: /dev/sda model: ST500DM002 size: 500.1GB serial: W3T96LQL temp: 37C
           ID-2: /dev/sdb model: WDC_WD5000AAKS size: 500.1GB serial: WD-WCAWF9850815 temp: 44C
           ID-3: USB /dev/sdh model: USB_Disk size: 4.0GB serial: 37270552CF16434522243-0:0 temp: 0C
Partition: ID-1: / size: 30G used: 12G (43%) fs: ext4 dev: /dev/sda1 
           ID-2: /home size: 208G used: 45G (23%) fs: ext4 dev: /dev/sda2 
           ID-3: swap-1 size: 1.28GB used: 0.00GB (0%) fs: swap dev: /dev/sdb1 
Sensors:   System Temperatures: cpu: 30.0C mobo: N/A 
           Fan Speeds (in rpm): cpu: N/A 
Info:      Processes: 120 Uptime: 4 min Memory: 108.6/2946.9MB Init: systemd v: 215 runlevel: 1 Gcc sys: 4.9.2 
           Client: Shell (bash 4.3.301 running in tty ) inxi: 2.1.28 

[pylkko]

Anything in xorg or other log?

You could try upgrading to 9.9 to see if you still have similar problems.

[Wheelerof4te]

Switch to tty and upgrade to Stretch. There is no reason to use outdated, barely maintained Debian version when there is fully supported one available for more than 2 years.
Even more so if you're running a GNOME DE.

[GarryRicketson]

When I started having all kinds of trouble with Gnome, and it's use less "Oh no something went wrong" error messages, Freezes, and all sorts of other problem, well it was more then 5 years ago, it WAS a nice DE, but has mutated into something else. and is not nice, not user friendly, and not worth wasting time on it, since the developers will just continue doing what they want, and making it more bloated, less use full, and more buggy.
So I stopped using it, and the endless problems also stopped. I never have looked back, and it is a mystery to me why people keep trying to use it.

[None1975]

And is not nice, not user friendly, and not worth wasting time on it, since the developers will just continue doing what they want, and making it more bloated, less use full, and more buggy.

Excellent said!

So I stopped using it, and the endless problems also stopped. I never have looked back, and it is a mystery to me why people keep trying to use it.

I think the reason for this is advertising. In almost all major distributions, it is intrusive. Even in Debian it is the main DE. In addition, most people are lazy, don't like to analyze or try something new.

[Wheelerof4te]

When I started having all kinds of trouble with Gnome, and it's use less "Oh no something went wrong" error messages,

Those messages appear when you don't have a GPU that can do hardware acceleration. Honestly, even if Linux does strive to support all kinds of old hardware, using a GPU that can't do OpenGL accel in 2019th is just shooting yourself in the foot for no reason. But if it works for you, it's fine.

OP more likely has a problem with Jessie's most recent kernel, there was an security upgrade not long ago. I think upgrading to Stretch's kernel will solve OP's problem. Otherwise, it is likely a hardware issue.

EDIT: Let us not use every opportunity to attack GNOME and focus on solving the actual problem. GNOME should run well on 90% of hardware sold in the last 10 years.

[GarryRicketson]

Well, not attacking Gnome at every opportunity, but if one just looks at the number of posts, and how many are problems with Gnome,... does not take much logic to see something went wrong, and it started some years back, any way, I don't miss it at all and certainly am not going out and buying the latest hardware, just so I can use Gnome, why ? it is not needed at all.

Otherwise, it is likely a hardware issue.

My hardware is fine, it runs smoothly, reliable and reasonably fast "snappy" in fact,... as long as Gnome is not installed. I would say it is a software problem.
Many people just don't learn, and very few have any common sense ,... look at how many think that the next MS windows will be better, and keep on using that, ...just plain does not make sense to me, ...
I think the Debian developers are making a huge mistake in keeping Gnome as the default DE, Yes Debian, at least the base is pretty solid, and good, not as good as it was, but they wanted change, and made the changes, we all know that story, but on top of it , now, with Gnome, Debian is not "rock solid", but everyone thinks it should be. Get Gnome out of the picture, and Debian is still pretty reliable and solid. I know , the Gnome lovers/users, will not agree, and that is fine, I don't really care. Just like I don't really care if the MS windows lovers want to keep using that. One can try to suggest that there is something better, but the "mainstream" crowd , says follow us, we know this will get better some day, well that is fine, follow them, but not me, I will go the route that really is better, and am thank full all those problems are a thing of the past. So enjoy trying to fix the broken Gnome, repeating the same mistakes over and over, in a endless cycle of "Oh no, something went wrong".
=== Use some logic,....

Those messages appear when you don't have a GPU that can do hardware acceleration.

IF that is the case, then what kind of moron, would use "Oh no something went wrong" as the error message ? This just goes to show what kind of idiots are involved with developing Gnome, I mean, would it not be much more use full to write the code so that
IF GPU does not have hardware acceleration: error message: "You need to have hardware acceleration for this software to work", or something a little more clear, ? Something went wrong", gee, My screen is frozen, no keyboard access, ...I did not notice something is wrong, thank you for telling me,...

[Canuck62]

As per suggestions, I have raked through the other logs and there is nothing that looks like an error (or even suspicious).

Looks like an upgrade to 9.9 might be the next step.

[Wheelerof4te]

IF that is the case, then what kind of moron, would use "Oh no something went wrong" as the error message ? This just goes to show what kind of idiots are involved with developing Gnome, I mean, would it not be much more use full to write the code so that
IF GPU does not have hardware acceleration: error message: "You need to have hardware acceleration for this software to work",

True that. I myself think that error message is silly, as I have experienced it when I first tried using GNOME3 on my old PC. Reminds me of Windows 10 error messages, lol.

[pylkko]

There's people that spend a lot of time developing complicated software, for free, so that users can have features on linux desktop that most people are used to nowadays on any platform. And then there are people that go around calling people, that they've never met, "idiots" on internet forums. Call it "diversity".

[GarryRicketson]

I have sufficient packages installed (including some applications under Wine) that I am loathe to do a scratch install.

Just a suggestion here, for the OP, try installing a Window Manager, OpenBox is a good one to start with,..then when before you log in, use the selection tool, and select the WM, see if you still get freezes ,etc. It will take some getting use to, though.
DO NOT try to remove Gnome, just add the window manager, and give it a try. This could be a improvement for the meantime, maybe in future upgrades Gnome will work better, (I have my doubts).
It was Debian 6, when I was using Gnome, and at first on Debian 7 it was still ok, but by the time it was Debian 7.5 (or something like that), that is when I started noticing the deterioration of performance, and I added a WM, if I remember correctly it was Fluxbox, and I went to OpenBox later, any way, I really liked the WM, so I got the brainstorm to try to remove Gnome,... DO NOT try that, it left me with no login window, nor DE or WM, I was able to finally get things working again, but it was not easy, and rather long complicated story,..
If you prefer, you could also, try XFCE, Debian does support multi DE's or DE/WM,s ...
Any way, when Debian 8 was released, I tried it, but was not happy, so I kept using Debian 7 for some time,.. been working with Debian 9, lately, and more so Debian 10, but can not say I have tried Gnome on either, maybe I will, just to see if there has been any improvement, but I have my doubts on that,...I suppose maybe if I was using hardware that is newer, and with "bells and whistles", 8 or more gb of ram, etc, it might be ok, but still, for many other reasons, A simple WM, and terminal to me seems much more efficient
and easier to trouble shoot, if and when there is problems. So even if I did have newer hardware, I would keep using the WM and terminal, it is so much more user friendly then the more complex software, that get's like a plate full of spaghetti, and there is a worm somewhere in the spaghetti, well it get's kind of hard to find the worm in the plate of spaghetti, and I do not want to eat the worm, or like trying to find the maggot in a huge bowl of rice, ...
ERROR: "There is a bug some where, but we can't find it".
Yea, that is what it is , but where is it ?

[vbrummond]

Modern desktop environments in Linux fix a lot of issues with user experience and security. This comes at the expense of corner cases and unsupported hardware. It sounds to me like your GL drivers are not supported well, and this is only going to be fixed from code in a newer kernel or mesa.

[Canuck62]

Problem solved
Short answer: Upgraded to Ver 9.9 and desktop is now stable. Thanks to all who took the time to reply

Long answer:
I didn't think it was hardware as the system ran stable with a tty that was displaying the correct resolution for the monitor. Ran backups which should have pushed memory and i/o and the system was stable in terminal mode.
After backing up one night, on the next night followed instructions from https://www.debian.org/releases/stable/ ... ading.html.
Upgrade took about 4 hours, but I came back a few times to find prompts so there might have been delays waiting for my input.
selected gdm3 as the display manager - others were present so don't know if that caused the flakiness with the last Jessie update
other prompts were all typos or minor differences in config files
One minor hiccup as the desktop came up in xfce (which I had tried under Jessie to see if it was purely a gnome problem) Here I was thinking "gnome's change a lot" until I realised and switched back

An initial test of software is all positive except duplicity backup which has got in a tizzy. Will sort that myself
Thanks again