cinnamon 2.4.7+rebecca crashing

[apuntigam]

I run daily updates to Mint 17.1 Cinnamon (32-bit) on my laptop. Recently the cinnamon 2.4.7+rebecca package came up in the Update Manager, I installed it and rebooted. Since then, Cinnamon crashes every time I restart, without fail. Before this update, everything ran perfectly.

To test, I did a clean install of Mint 17.1 (32-bit). I installed every update except "cinnamon 2.4.7+rebecca" and rebooted, no problem. The only update left on the list after a refresh was cinnamon 2.4.7. I installed it, rebooted, and again Cinnamon is permanently crashing.

I have tried a number of kernels, doesn't seem to change anything. Currently on the latest 3.16 kernel in the update list, 3.16.0-33 I think. I have gone back to the recommended 3.13 kernel, no change.

Running on a laptop, old Acer Aspire 1414WLMi. Celeron M 350, 1.2GB RAM, 40GB HDD.

Can I give any additional info to help?

[clefebvre]

What output do you get when running Cinnamon from the command line with "cinnamon --replace"?

[clefebvre]

Also please check for crashes info in dmesg...

gdb backtrace for cinnamon would be ideal.

If you're able to compile Cinnamon from git, compile the 2.4-maintenance branch with the following commit reverted: 5c189e9

[apuntigam]

When I run "sudo cinnamon --replace", I get the following error message:
error: XDG_RUNTIME_DIR not set in the environmen.
Window manager error: Unable to open X display

[clefebvre]

not sudo..

[clefebvre]

just cinnamon --replace

[apuntigam]

Ah ok, if I just run "cinnamon --replace" I get:
Window manager error: Unable to open X display

[clefebvre]

You need to run that in Xorg, not in console though

[clefebvre]

Can you try version 2.4.8? We just updated the repositories.

From console you can "apt update; apt install cinnamon" to install version 2.4.8.

[apuntigam]

Installing 2.4.8 worked, thank you!

Do you want me to keep testing anything or just forget about it?

[clefebvre]

No it's fine.

Using a new feature in cogl broke Cinnamon for you. We went back to how we did things in the past here to fix this.

We can't reproduce that issue on our hardware though, and we want to move to that new cogl way of doing things (we believe it might fix freezing issues). We'll try to detect hardware better and do it only on newer hardware.

If you can, please make sure to take part in testing Mint 17.2 RC when it comes out, or Cinnamon 2.6 even before that.

Thanks.

[mtwebster]

Hi, I've created some test packages consisting of v2.4.8 + a new patch that hopefully both prevents this crash and fixes freeze issues for those gpu's that support some improved functions:

NOTE: These are only if you're running stable cinnamon from the repos (2.4.8) - nightlies already include the fix I'm testing here.

32-bit:
https://dl.dropboxusercontent.com/u/8201170/cinnamon_2.4.8_i386.deb
https://dl.dropboxusercontent.com/u/8201170/cinnamon-dbg_2.4.8_i386.deb
https://dl.dropboxusercontent.com/u/8201170/cinnamon-common_2.4.8_all.deb

64 bit:
https://dl.dropboxusercontent.com/u/8201170/cinnamon_2.4.8_amd64.deb
https://dl.dropboxusercontent.com/u/8201170/cinnamon-dbg_2.4.8_amd64.deb
https://dl.dropboxusercontent.com/u/8201170/cinnamon-common_2.4.8_all.deb

They are versioned the same as the current release, so if you run into issues, you can just upgrade to replace them.

I'd appreciate anyone who can test these, if you were experiencing this particular issue in 2.4.7, or have been experiencing freezes in < 2.4.7 and in 2.4.8 (but not 2.4.7). In fact I'd love as many testers as possible for a bit of data collection -

If you do test this, besides reporting whether everything is ok or not, and whether it solves/doesn't solve an issue, I'd like you to report:

• 32 or 64 bit
• CPU type (intel, amd etc..)
• Graphics (intel, nvidia, radeon, etc..)
• The output of ~/.xsession-errors, just one line:

St-Message: cogl npot texture sizes SUPPORTED
or:
St-Message: cogl npot texture sizes NOT supported

Thanks a lot!

[jonorthwash]

I'm running Cinnamon 2.5 / 2.5.1 (i.e., latest version from git, as of 2015-04-13). Is there a way to test these patches for those versions?

[mtwebster]

@jonorthwash This patch is also in git master as of April 12th, so you probably are already running with it. Check your .xsession-errors if you see that St-Message, you have it.

[ghost]

this happens for me since mint 17.1

the new updates were supposed to fix it and i actually thought they had for a while as i had noticed no stutter or lag whilst gaming for about an hour. but i actually just booted my machine and left it on the desktop for 1.20hrs with no activity and it was back with a vengeance.

troubleshoot --> restarting cinnamon fixes it temporarily. its a workaround but has been present for 3 months now.. it does seem better but not actually fully fixed rather delayed for longer. However the overall memory leak is definitley fixed ! so a big thanks for that.

My setup: Cinnamon 2.4.8 , Kernal happens on both the initial 3.13 and my current 3.16.0-34-generic
processor is AMD Trinity APU , GPU is Nvidia 7 series desktop GPU with latest beta 349.16 drivers.

everything else apart from the kernal update is standard, no tweaks no desklets, no applets, no extensions. no extra icons on the panel or desktop. Just a vanilla install with closed nvidia drivers. Its a fresh install only half a day old.

I tried the above patches but apparently im on a newer version so they didn’t work.

[mtwebster]

Not sure what your symptoms are exactly.. is it freezing? Did you happen to check the .xsession-errors message I mentioned above? Also are you 32 or 64 bit?

[ghost]

Im on 64bit ( sorry didn’t realise people were actually bothering with 32bit anymore )

the symptoms if you imagine moving your mouse across the screen with a grabbed window in hand horizontally there is a micro stutter that holds the window in place for 20 - 50ms chunks whilst the screen updates after a short period the thing pauses and then starts stuttering again, all fluidity is lose but tearing does NOT occur.
I think its a composting error. I would say that when it actually does work on first boot the mouse lag to window move is annoying ( im used to fluid window moves with instant response, like when vsync is off but with no tearing ) its probably that xserver is old & terrible or that the closed nvidia driver is just adds another layer of useless abstraction somehow.

anyways i can only add this:

Window manager warning: Log level 8: meta_end_modal_for_plugin: assertion 'compositor->modal_plugin == plugin' failed
Window manager warning: Log level 16: Called enable_unredirect_for_screen while unredirection is enabled.
Window manager warning: Log level 8: meta_end_modal_for_plugin: assertion 'compositor->modal_plugin == plugin' failed
Window manager warning: Log level 16: Called enable_unredirect_for_screen while unredirection is enabled.
Window manager warning: Log level 8: meta_end_modal_for_plugin: assertion 'compositor->modal_plugin == plugin' failed
Window manager warning: Log level 16: Called enable_unredirect_for_screen while unredirection is enabled.
Window manager warning: Log level 8: meta_end_modal_for_plugin: assertion 'compositor->modal_plugin == plugin' failed
Window manager warning: Log level 16: Called enable_unredirect_for_screen while unredirection is enabled.
Window manager warning: Log level 8: meta_end_modal_for_plugin: assertion 'compositor->modal_plugin == plugin' failed
Window manager warning: Log level 16: Called enable_unredirect_for_screen while unredirection is enabled.

[mtwebster]

Hmm sounds like you have some entirely unrelated issue - what are you doing when this happens? You mentioned gaming - steam?

[mtwebster]

I think you may somehow be confusing cinnamon, does it only happen like escaping full screen maybe? Any steps you might be able to reliably reproduce with..

[ghost]

well certainly due to the steam client being as it is a big annoying block of non GTK it does seem to eventually incur it. However right now, i have left my machine (up until this point ) from a fresh boot and it didn’t do it.
Im not sure if this is breaks stuff but I have put some # < spaces in my xsession errors so i know when new ones started just incase I report the wrong ones.

The next time it happens il log on here and report what is in that file and what i was doing + anything else you need.

thanks

[nkoropoff]

I installed the above test packages and have noticed smoother and faster graphic performance, particularly with drag/drop. No stuttering and gratefully no freezes for the past two days now. This is what I experienced with the 2.4.7 update. The information you asked for:

System: Kernel: 3.13.0-45-generic x86_64 (64 bit)
CPU type (Dual core Intel Core i7-4600U CPU (-HT-MCP-) cache: 4096 KB flags: (lm nx sse sse2 sse3 sse4_1 sse4_2 ssse3 vmx) Clock Speeds: 1: 756.00 MHz 2: 756.00 MHz 3: 756.00 MHz 4: 1300.00 MHz)
Graphics (Intel Haswell-ULT Integrated Graphics Controller HD4400 X.Org: 1.15.1 drivers: intel (unloaded: fbdev,vesa) Resolution: 2560x1440@60.0hz, 3840x2160@30.0hz GLX Renderer: Mesa DRI Intel Haswell Mobile GLX Version: 3.0 Mesa 10.1.3)

The output of ~/.xsession-errors, just one line: St-Message: cogl npot texture sizes SUPPORTED

Thanks.

[jonorthwash]

Using the git version from 2015-04-13, I continue to get the ~30ms freezes every few seconds. In addition, a new freeze was introduced: if I hover in the wrong way in the Window List With App Grouping applet, I get a ~30-second freeze. I do not have any St-Message or similar text showing up in my ~/.xsession-errors file.

[mtwebster]

@jonorthwash Please update again to try this new patch. If you're getting no St-Message, you don't have it yet.

[ghost]

I have the exact same stuttering, it took a while and didn’t appear yesterday. fresh boot 2.24hrs up time.

Window manager warning: CurrentTime used to choose focus window; focus window may not be correct.
Window manager warning: Got a request to focus 0x1800010 (audiotrack) with a timestamp of 0. This shouldn't happen!
Window manager warning: Buggy client sent a _NET_ACTIVE_WINDOW message with a timestamp of 0 for 0x220001b (Steam)

(cinnamon:2572): Clutter-CRITICAL **: clutter_actor_remove_child: assertion 'child->priv->parent != NULL' failed

That is all that is in the log since i booted so far..

also is it normal to get this output with uptime: 16:11:01 up 2:25, 2 users, load average: 0.18, 0.27, 0.20

I am the only user ? Is the other one root, sorry im a noob.

*edit, been on suspend for 1hr, i return and the problem has gone

[jonorthwash]

I pulled the latest Cinnamon and muffin today, built, installed, and restarted. I still see no St-Mesage in ~/.xsession-errors, and I still get the momentary freezes. I feel I'm missing something obvious, such as needing to update another package, or needing to apply the patch.

Also, after ending dpms standby (or whatever it's called when the screen blanks and the backlight is turned off), I'm getting lots of tearing and momentary artifacts until I switch to a vt and back to X. This is an old long-standing annoyance that had seemed to be gone in the update from a few days ago, but is now back.

[mtwebster]

Not sure what's going on - you will see that message if you're running nightlies. Were they existing git repos on your machine? Try just deleting and recloning? Should only need to update cinnamon, not muffin or anything else.

[jonorthwash]

I just moved the old Cinnamon clone, cloned a fresh copy, compiled, installed, and restarted. I get the same: no St-Message in ~/.xsession-errors, and continued momentary freezes. I think I'd prefer to run nightlies than straight from git anyway, actually, but I've been having trouble finding out how I might do that on debian unstable.

[mtwebster]

If you're in Debian Jessie, the message may be in /var/log/user.log (I think) - you can try just running cinnamon --replace and the log output will go to that terminal instead.

[jonorthwash]

Yes, I get the following in the terminal on cinnamon --replace, though /var/log/user.log is empty (though it exists).

Cjs-Message: JS LOG: About to start Cinnamon
St-Message: cogl npot texture sizes SUPPORTED
Window manager warning: Trying to remove non-existent custom keybinding "magnifier-zoom-in".
Window manager warning: Trying to remove non-existent custom keybinding "magnifier-zoom-out".
Cjs-Message: JS LOG: Cinnamon started at Thu Apr 16 2015 15:58:23 GMT-0400 (EDT)
glibtop: Non-standard uts for running kernel:
release 3.14-1-amd64=3.14.0 gives version code 200192

Cjs-Message: JS LOG: network applet: Cannot find connection for active (or connection cannot be read)
Cjs-Message: JS WARNING: [/home/jonathan/.local/share/cinnamon/applets/workspace-grid@hernejj/BarIndicatorStyle.js 96]: anonymous function does not always return a value
Cjs-Message: JS WARNING: [/home/jonathan/.local/share/cinnamon/applets/workspace-grid@hernejj/GridStyle.js 81]: anonymous function does not always return a value
Cjs-Message: JS WARNING: [/home/jonathan/.local/share/cinnamon/applets/workspace-grid@hernejj/GridStyle.js 96]: assignment to undeclared variable btn_height
Window manager warning: Log level 16: Symbolic icon /usr/share/cinnamon/applets/windows-quick-list@cinnamon.org/icons/windows-quick-list-symbolic.svg is not in an icon theme directory

However, I still get the momentary freezes.

[savuori]

I have two computers, both still get freezes.
Both are 64 bit.
One is Intel i7, another one is AMD FX 8350
One is using intel iris graphics and other one radeon (tried both open source / fglrx drivers)
Both have cogl npot texture sizes SUPPORTED.

[jonorthwash]

While I still get momentary freezes with the recent update, the general bogging down of the system and increased system load the longer I run cinnamon seem to be gone.

[mtwebster]

Please see #4072 to continue freeze-related troubleshooting