Random Screen Locks

[timmocahill]

Expected behaviour

Mate is invisible when launching apps or enabling screensaver.

Actual behaviour

Mate is preventing apps from launching from Mate components like Marco or from the 'start' menu. Sometimes the screen locks are random.

Steps to reproduce the behaviour

Ctrl-Alt-L reliably locks the screen, but it still allows a keyboard shortcut to tty1 to be visible. When screensaver comes on, it blanks the screen and Ctrl-Alt-F1doesn't display. Other similar events are random. Can reliably prevent the issue from arising by launching terminal from desktop (instead of via Mate panel) and launching all apps from the terminal via command line. Ctrl-Alt-L can be recovered by switching to tty1, finding pidof mate-screensaver, and killing it.

MATE general version

1.24.0

Package version

mate-panel/focal,now 1.24.0-2 amd64
crash.tar.gz

Linux Distribution

Linux Mint 5.11.0-25-generic #27~20.04.1-Ubuntu SMP Tue Jul 13 17:41:23 UTC 2021 x86_64 x86_64 x86_64 GNU/Linux

Link to bugreport of your Distribution (requirement)

https://termbin.com/8zi0

[timmocahill]

Had another hang with the same configuration as a youtube video played via a USB headphone (Jabra40). I was able to recover by killing Firefox, in which the video was playing. The video became choppy and garbled and then stopped. The stderr is below:

ALSA lib conf.c:5187:(snd_config_expand) Unknown parameters 1
ALSA lib control.c:1379:(snd_ctl_open_noupdate) Invalid CTL sysdefault:1
ALSA lib conf.c:5187:(snd_config_expand) Unknown parameters 2
ALSA lib control.c:1379:(snd_ctl_open_noupdate) Invalid CTL sysdefault:2
ALSA lib pcm_dmix.c:1089:(snd_pcm_dmix_open) unable to open slave
ALSA lib pcm_dmix.c:1089:(snd_pcm_dmix_open) unable to open slave
ALSA lib pcm_dmix.c:1089:(snd_pcm_dmix_open) unable to open slave
ALSA lib pcm_dmix.c:1089:(snd_pcm_dmix_open) unable to open slave
ALSA lib pcm_dmix.c:1089:(snd_pcm_dmix_open) unable to open slave

On re-launch of Firefox from terminal window, the following appeared:
[GFX1-]: More than 1 GPU from same vendor detected via PCI, cannot deduce device

[lukefromdc]

I'm wondering if this is related to random lockups I've had that appear to be GPU hangs (thus not related to MATE) where sound still plays but the keyboard and mouse appear to be unresponsive and nothing on the screen that would be expected to update (e.g a progress dialog) ever changes. Sometimes getting to another v-t used to be able to get out of this, but not recently.

Had this with two different motherboards, two different CPU and RAM setups, but the same 1st-gen AMD GCN card and AMDGPU driver. Your stderr also looks like a GPU hang, in this case followed by a reset. Mine apparently won't reset and hangs for good, forcing a reboot

[timmocahill]

While it'd be foolish to rule out hardware, the 2nd event was just another example of a crash. I've had Marco crash me (here's report from one that happened today - https://termbin.com/m5bv). I've done nothing more than hover over the LM menu and lock up.

For what it's worth, my hardware is:
Topology: 8-Core model: AMD Ryzen 7 4800H with Radeon Graphics bits: 64 type: MT MCP
arch: Zen rev: 1 L2 cache: 4096 KiB
Device-1: AMD Navi 10 [Radeon RX 5600 OEM/5600 XT / 5700/5700 XT]
vendor: Micro-Star MSI driver: amdgpu v: kernel bus ID: 03:00.0 chip ID: 1002:731f
Device-2: AMD Renoir vendor: Micro-Star MSI driver: amdgpu v: kernel bus ID: 08:00.0
chip ID: 1002:1636

It came with windows 10 which i went ahead and registered so I can run it again as a VM in the future, but it didn't give me any troubles for the brief time it ran. I also sent my issue to the AMD drivers team.

[timmocahill]

@lukefromdc I just noticed something when I plugged in a USB headphone (a small pair of Samsung headphones that came with the cell phone I bought a year ago).

ALSA lib conf.c:5187:(snd_config_expand) Unknown parameters 1
ALSA lib control.c:1379:(snd_ctl_open_noupdate) Invalid CTL sysdefault:1
ALSA lib conf.c:5187:(snd_config_expand) Unknown parameters 2
ALSA lib control.c:1379:(snd_ctl_open_noupdate) Invalid CTL sysdefault:2

So same error output. This may help convict software or hardware, but idk which.

[timmocahill]

New one from Marco today.

https://termbin.com/xy80

[timmocahill]

New one from mate-volume-control today!

https://termbin.com/xsha

[lukefromdc]

Note that some of us do not follow links to 3ed party file hosting sites for security reasons. Please post that report here directly

[timmocahill]

Fair enough, Luke! I wrongly assumed that the button in Mintreport would have some inherent security involved. I attached tars of the docs containing the marco and volume crashes.

mate-volume-control.crash.tar.gz
marco.crash.tar.gz

[lukefromdc]

You need to install debug symbols (or install the programs by make install) for these backtraces to be useful. The volume control crash was a sigabrt, but marco got a segfault. These look like two different crashes. With the debug symbols at least the segfault in marco would probably give us a bad line in the source code to look at. This part of the marco stacktrace is where we might have gotten a line number in the source code:

       Session: c2
     Owner UID: 1000 (timmo)
       Boot ID: 7a1274d5c72a4b01b535b48750a966ff
    Machine ID: 7a0781c5129d4f848f901e53f906eaf1
      Hostname: Alpha-17
       Storage: /var/lib/systemd/coredump/core.marco.1000.7a1274d5c72a4b01b535b48750a966ff.1681.1628080547000000000000.lz4
       Message: Process 1681 (marco) of user 1000 dumped core.
                
                Stack trace of thread 1681:
                #0  0x00007fc557a1c36e meta_frame_calc_borders (libmarco-private.so.2 + 0x3a36e)
                #1  0x00007fc557a0adce n/a (libmarco-private.so.2 + 0x28dce)
                #2  0x00007fc557a0cd71 n/a (libmarco-private.so.2 + 0x2ad71)
                #3  0x00007fc557a0dce1 n/a (libmarco-private.so.2 + 0x2bce1)
                #4  0x00007fc557a16b06 n/a (libmarco-private.so.2 + 0x34b06)
                #5  0x00007fc557a60bb9 n/a (libmarco-private.so.2 + 0x7ebb9)
                #6  0x00007fc556b9505f n/a (libgdk-3.so.0 + 0x7005f)
                #7  0x00007fc556b9538b n/a (libgdk-3.so.0 + 0x7038b)
                #8  0x00007fc556b5d0f4 gdk_display_get_event (libgdk-3.so.0 + 0x380f4)
                #9  0x00007fc556b950f6 n/a (libgdk-3.so.0 + 0x700f6)
                #10 0x00007fc557adf17d g_main_context_dispatch (libglib-2.0.so.0 + 0x5217d)
                #11 0x00007fc557adf400 n/a (libglib-2.0.so.0 + 0x52400)
                #12 0x00007fc557adf6f3 g_main_loop_run (libglib-2.0.so.0 + 0x526f3)
                #13 0x000055df6e380233 main (marco + 0x3233)
                #14 0x00007fc5578170b3 __libc_start_main (libc.so.6 + 0x270b3)
                #15 0x000055df6e38064e n/a (marco + 0x364e)

[timmocahill]

i'm using Linux Mint 20.04 and have tried doing the simple things to get the debug symbols (apt-get install marco-dbg) and remedial web searches are only giving me very broad-based answers (e.g. how to install xorg debug symbols, but nothing specific to Mate software). Do you have a link to instructions and/or the debug symbols that I should install?

[lukefromdc]

No I do not, as I have never interacted with Mint's bug reporting system. Basically though in Debian/Ubuntu based distros (mint incluyded) you should be looking for a package named -dbg , maybe install Synaptic package manager (if Mint still offers it) and search for all packages with marco in the name.

Others here might be able to help.

[timmocahill]

After some flailing, I have the following installed. I've been pointing find-dbgsym-packages at PIDs for various Mate services and looking for the results that listed packages that had Mate or Marco in the name. Please advise if others are recommended/needed:

mate-session-manager-dbgsym
marco-dbgsym
libmarco-private2-dbgsym
mate-panel-dbgsym
libmate-desktop-2-17-dbgsym
lightdm-dbgsym

[timmocahill]

Ok, here's what I have for tonight's crash of the Mate-settings-daemon. Of course, I didn't have the debug symbols loaded for it, so feel free to ignore the dump. Just adding for tracking purposes.

crash.tar.gz

[timmocahill]

So tonight we crashed when trying to print a simple pdf file. Looks like the notification-area-applet died first. Then mate-panel went a few seconds after that, followed by mate-terminal.... which pretty much killed everything else since I launched all my apps from the terminal. I don't have debug symbols installed for the terminal. When I try, I get this error and I'm just too worn out to figure out how to force the installation right now.

mate-terminal-dbgsym : Depends: mate-terminal (= 1.24.0-2ubuntu1) but 1.24.0-linuxmint1 is to be installed

mate-terminal.crash.tar.gz
notification-area-crash.tar.gz
mate-panel.crash.tar.gz

[lukefromdc]

Someone else here will have to work on this, more than I can figure out. It does look however like you have a problem system. No idea what the underlying problem is (bad RAM, a corrupted file somewhere, something else?) but I cannot duplicate this sort of crashiness at all. All I've had are GPU lockups, where you do not SEE anything crash, the screen just stops responding.

…

On 8/10/2021 at 12:21 AM, "timmocahill" ***@***.***> wrote: So tonight we crashed when trying to print a simple pdf file. Looks like the notification-area-applet died first. Then mate- panel went a few seconds after that, followed by mate-terminal.... which pretty much killed everything else since I launched all my apps from the terminal. [mate-terminal.crash.tar.gz](https://github.com/mate-desktop/mate- panel/files/6958113/mate-terminal.crash.tar.gz) [notification-area-crash.tar.gz](https://github.com/mate- desktop/mate-panel/files/6958099/notification-area-crash.tar.gz) [mate-panel.crash.tar.gz](https://github.com/mate-desktop/mate- panel/files/6958106/mate-panel.crash.tar.gz) -- You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub: https://github.com/mate-desktop/mate- panel/issues/1242#issuecomment-895637555