xfce4-session and marco crashes/bugs with dual GPUs

Bug #1942314 reported by JXT
8
This bug affects 1 person
Affects Status Importance Assigned to Milestone
marco (Ubuntu)
Invalid
Undecided
Unassigned
xfce4-session (Ubuntu)
New
Undecided
Unassigned

Bug Description

Install 20.04 or 21.04, install the proprietary drivers. Add a second XScreen so you can enable the primary GPU and its screens and applications start to crash. System is unusable.

As a long after update the driver isn't the issue, this will affect Dual GPU - Multi XScreen of ANY GPU combination because Gnome/GTK has broken XScreen enumeration. This is also not an exclusive XFCE issue, anything built on GTK will be affected.

ProblemType: Bug
DistroRelease: Ubuntu 20.04
Package: xorg 1:7.7+19ubuntu14
ProcVersionSignature: Ubuntu 5.11.0-27.29~20.04.1-generic 5.11.22
Uname: Linux 5.11.0-27-generic x86_64
NonfreeKernelModules: nvidia_modeset nvidia
.proc.driver.nvidia.capabilities.gpu0: Error: path was not a regular file.
.proc.driver.nvidia.capabilities.gpu1: Error: path was not a regular file.
.proc.driver.nvidia.capabilities.mig: Error: path was not a regular file.
.proc.driver.nvidia.gpus.0000.25.00.0: Error: path was not a regular file.
.proc.driver.nvidia.gpus.0000.26.00.0: Error: path was not a regular file.
.proc.driver.nvidia.registry: Binary: ""
.proc.driver.nvidia.suspend: suspend hibernate resume
.proc.driver.nvidia.suspend_depth: default modeset uvm
.proc.driver.nvidia.version:
 NVRM version: NVIDIA UNIX x86_64 Kernel Module 470.57.02 Tue Jul 13 16:14:05 UTC 2021
 GCC version: gcc version 9.3.0 (Ubuntu 9.3.0-17ubuntu1~20.04)
ApportVersion: 2.20.11-0ubuntu27.18
Architecture: amd64
BootLog: Error: [Errno 13] Permission denied: '/var/log/boot.log'
CasperMD5CheckResult: skip
CompositorRunning: None
CurrentDesktop: XFCE
Date: Wed Sep 1 03:57:12 2021
DistUpgraded: Fresh install
DistroCodename: focal
DistroVariant: ubuntu
DkmsStatus: nvidia, 470.57.02, 5.11.0-27-generic, x86_64: installed
ExtraDebuggingInterest: Yes
GraphicsCard:
 NVIDIA Corporation GP106 [GeForce GTX 1060 6GB] [10de:1c03] (rev a1) (prog-if 00 [VGA controller])
   Subsystem: ZOTAC International (MCO) Ltd. GP106 [GeForce GTX 1060 6GB] [19da:1438]
 NVIDIA Corporation TU116 [GeForce GTX 1660] [10de:2184] (rev a1) (prog-if 00 [VGA controller])
   Subsystem: ZOTAC International (MCO) Ltd. TU116 [GeForce GTX 1660] [19da:5543]
InstallationDate: Installed on 2021-09-01 (0 days ago)
InstallationMedia: Xubuntu 20.04.3 LTS "Focal Fossa" - Release amd64 (20210819.1)
MachineType: Micro-Star International Co., Ltd MS-7C02
ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-5.11.0-27-generic root=UUID=087d1eaa-f486-4779-ac4a-d57e4f8dc551 ro quiet splash vt.handoff=7
SourcePackage: xorg
Symptom: display
UpgradeStatus: No upgrade log present (probably fresh install)
dmi.bios.date: 04/22/2021
dmi.bios.release: 5.17
dmi.bios.vendor: American Megatrends International, LLC.
dmi.bios.version: 1.H5
dmi.board.asset.tag: To be filled by O.E.M.
dmi.board.name: B450 TOMAHAWK (MS-7C02)
dmi.board.vendor: Micro-Star International Co., Ltd
dmi.board.version: 1.0
dmi.chassis.asset.tag: To be filled by O.E.M.
dmi.chassis.type: 3
dmi.chassis.vendor: Micro-Star International Co., Ltd
dmi.chassis.version: 1.0
dmi.modalias: dmi:bvnAmericanMegatrendsInternational,LLC.:bvr1.H5:bd04/22/2021:br5.17:svnMicro-StarInternationalCo.,Ltd:pnMS-7C02:pvr1.0:rvnMicro-StarInternationalCo.,Ltd:rnB450TOMAHAWK(MS-7C02):rvr1.0:cvnMicro-StarInternationalCo.,Ltd:ct3:cvr1.0:
dmi.product.family: To be filled by O.E.M.
dmi.product.name: MS-7C02
dmi.product.sku: To be filled by O.E.M.
dmi.product.version: 1.0
dmi.sys.vendor: Micro-Star International Co., Ltd
version.compiz: compiz N/A
version.libdrm2: libdrm2 2.4.105-3~20.04.1
version.libgl1-mesa-dri: libgl1-mesa-dri 21.0.3-0ubuntu0.3~20.04.1
version.libgl1-mesa-glx: libgl1-mesa-glx N/A
version.nvidia-graphics-drivers: nvidia-graphics-drivers-* N/A
version.xserver-xorg-core: xserver-xorg-core 2:1.20.11-1ubuntu1~20.04.2
version.xserver-xorg-input-evdev: xserver-xorg-input-evdev N/A
version.xserver-xorg-video-ati: xserver-xorg-video-ati 1:19.1.0-1
version.xserver-xorg-video-intel: xserver-xorg-video-intel 2:2.99.917+git20200226-1
version.xserver-xorg-video-nouveau: xserver-xorg-video-nouveau 1:1.0.16-1

Revision history for this message
JXT (jtipton-x) wrote :
description: updated
description: updated
JXT (jtipton-x)
description: updated
JXT (jtipton-x)
description: updated
JXT (jtipton-x)
description: updated
Revision history for this message
JXT (jtipton-x) wrote (last edit ):

[ 46.504672] xfce4-session[1220]: segfault at 0 ip 00007f52d3e7fb7e sp 00007ffd390d35f8 error 4 in libc-2.31.so[7f52d3d1e000+178000]

firefox
[GFX1-]: More than 1 GPU from same vendor detected via PCI, cannot deduce device

X doesn't crash explicitly but it seems most things that rely on X do or have issues so I will be curious to see if what is deemed to actually be the problem if not X.

Tried Mate and Gnome to see if perhaps this was a DE issue rather than X causing issues for software under it.

Launching anything on the primary GPU in Mate causes all screens to strobe / flash.
[ 1121.457709] marco[10087]: segfault at f8 ip 00007f8075b57b66 sp 00007ffe8fd1e0f8 error 4 in libgdk-3.so.0.2404.16[7f8075b34000+81000]
[ 1121.457717] Code: 25 fe ff e9 57 fd ff ff 66 66 2e 0f 1f 84 00 00 00 00 00 90 f3 0f 1e fa 48 8b 87 f8 00 00 00 c3 0f 1f 40 00 f3 0f 1e fa 31 c0 <48> 39 bf f8 00 00 00 0f 94 c0 c3 66 66 2e 0f 1f 84 00 00 00 00 00

Gnome if I launch anything on the primary GPU/Displays windows ignore system theme/decoration and if they are moved, resized or manipulated in any way I get smearing/trails ala Windows 95 style screen not being updated/redrawn correctly.

Revision history for this message
Daniel van Vugt (vanvugt) wrote : Re: Crashes/bugs with dual GPUs when using Nvidia-470

Thank you for taking the time to report this bug and helping to make Ubuntu better. It sounds like some part of the system has crashed. To help us find the cause of the crash please follow these steps:

1. Look in /var/crash for crash files and if found run:
    ubuntu-bug YOURFILE.crash
Then tell us the ID of the newly-created bug.

2. If step 1 failed then look at https://errors.ubuntu.com/user/ID where ID is the content of file /var/lib/whoopsie/whoopsie-id on the machine. Do you find any links to recent problems on that page? If so then please send the links to us.

3. If step 2 also failed then apply the workaround from bug 994921, reboot, reproduce the crash, and retry step 1.

Please take care to avoid attaching .crash files to bugs as we are unable to process them as file attachments. It would also be a security risk for yourself.

summary: - XOrg crashes/bugs on Dual GPU 20.04 / 21.04
+ Crashes/bugs on Dual GPU 20.04 / 21.04
affects: xorg (Ubuntu) → xfce4-session (Ubuntu)
summary: - Crashes/bugs on Dual GPU 20.04 / 21.04
+ Crashes/bugs with dual GPUs when using Nvidia-470
Revision history for this message
JXT (jtipton-x) wrote (last edit ):

No thank you for taking the time to check into this and give me some extra place to peek for errors outside dmesg/syslog. I'm stressing hard with no way to upgrade outside losing a massive chunk of my hardware/workflow.

I see nothing labeled "ID" and the file name doesn't seem overly "IDish" but upon login I get _usr_bin_xfwm4.1000.crash

As a note when I triggered this I added a second XScreen on the SAME GPU that was already working. Once I restart XFWM4 dies and the system becomes laggy, unresponsive and input is seriously flaky. Windows on second or third XScreens "smear."

Is there anything from the crash that you need? I got the Whoopsie dialog when I logged back in but clicking on send did nothing as input becomes mostly miss, sometimes hit.

This issue persists on several distros and every DE I've tried. Also I'm using Nvidia-470 on my functioning config under 18.04.5 as well so in theory this isn't drivers and I'd not point the finger at XFWM despite it being the reported crasher. As I said, I can get any DE I try to crap itself on Debian / Arch distros by adding a second XScreen now and I've found others with this issue on here. Bug 1859526 for example seems to be a duplicate.

tags: added: multi-gpu nvidia
Revision history for this message
Sanjo (pabloguinness) wrote (last edit ):

Hi everyone, same error here. It truly renders the system unusable. I followed @vanvugt steps:
step 1: didn't work, generated again the same error
step 2: did work: https://errors.ubuntu.com/user/0e7a803ee9c8fa8cd6b5848d40922b7805dc68c991e42a6f9e7f769484ea2b1578393a81fb72c496bb012c3654c24327e6979a379014d7479f8283d00a4d88d9

Machine: HP Omen 16
AMD® Ryzen 7 5800h with radeon graphics × 16
NVIDIA GeForce RTX 3050 Ti Laptop GPU
Ubuntu 20.04.3 (fresh install)
Current desktop: Gnome 3.36.8 (however same error if Gnome 40 installed
Error message:
Xorg crashed with SIGABRT in _fbGetWindowPixmap()
Error: path was not a regular file
... (full report in link)

Revision history for this message
JXT (jtipton-x) wrote (last edit ):

I've come to find out this is 100% Gnomes fault. Shortly after GTK3 was finalized they ignorantly nuked the enumeration of XScreens. Because Ubuntu/Debian is slow to update this problem went semi-undetected for years until we had to move to a system where those changes/packages had been (forced) implemented. So now anything that uses GTK is basically broken.

However there are a few workarounds. I left Ubuntu fed up with this and other things. I'm on Garuda (Arch) but because of the downgrade feature I was able to reinstall XFCE4.12 before everything was broken and hold the packages. KDE also seems to be working though it was also broken for a while.

You will have to vett your applications for things that don't lock to a GPU process. So for example I use SpaceFM for file management and Sakura for a term. Sadly this will get worse as even XScreen saver is now broken as of 6.02. I reported the bug but the dev is clueless, doesn't run *nix at all.

It's a grim time to have multi GPU systems.

Revision history for this message
Daniel van Vugt (vanvugt) wrote :

Sanjo,

The Xorg crashes in comment #5 are in amdgpu_drv.so - please open a new bug for that. This bug should remain about the Nvidia issue experienced by JXT.

summary: - Crashes/bugs with dual GPUs when using Nvidia-470
+ xfce4-session and marco crashes/bugs with dual GPUs when using
+ Nvidia-470
Revision history for this message
JXT (jtipton-x) wrote (last edit ): Re: xfce4-session and marco crashes/bugs with dual GPUs when using Nvidia-470

No it's not, he HAS a nVidia card, that is the point, DUAL GPU, not single brand of GPU. Originally the bug seemed to be the driver but it was coincidence that the GTK updates arrived with the same update making it look like the driver was to blame. His issue is 100% THE issue. Dual GPU of any brand/model on a GTK based DE will be affected.

The fix is to hold XFCE4.12 ( mid release cycle builds) or move to a non-GTK based desktop. KDE was broken for a while as well but seems safe for multi GPU...for now. I myself ditched Ubuntu because you can't hold packages from previous releases once you upgrade to a newer release thus Ubuntu forces you into this bug.

description: updated
summary: - xfce4-session and marco crashes/bugs with dual GPUs when using
- Nvidia-470
+ xfce4-session and marco crashes/bugs with dual GPUs
description: updated
Changed in marco (Ubuntu):
status: New → Invalid
tags: added: multigpu
removed: multi-gpu
Revision history for this message
Hamed Karbasi (hamed.karbasi.93) wrote :

Hi,

I've faced the same issue with the same crash report. The problem occurred when I used a second screen connected to my HDMI port. My laptop has two GPUs (Intel Iris and Nvidia MX 330).
My applications keep crashing, including firefox and skype.

Nvidia driver version: 515.65.01
Ubuntu version: 22.04
Kernel: 5.15.0-53-generic

Revision history for this message
JXT (jtipton-x) wrote (last edit ):

If you're doing it with an xorg.conf for separate XScreens there is no fix because xfce uses GTK and GTK was hard coded to 1 XScreen around gtk3. Since it doesn't enumerate XScreens (by design) any longer it's completely usuable. You can try prime or xrandr style set ups and the performance will be terrible. The real fix is move to a DE/WM that does not use GTK. Sadly there aren't many. KDE (bloated, slow, has some real obnoxious behaviors with some screen setups) or roll your own "DE" with self vetted programs. I moved to CTWM, works out of the box and is more to my liking. However it's also old/obscure and learning the config to get it usable takes a lot of time. There isn't much else I can think of. Outside KDE I only found 3 or 4 extremely old WM that would work (in theory) but couldn't build most. Elementary is in the process of ditching GTK but that doesn't help now.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.