Driver crash/desktop hang when connecting 2 Displayport VRR displays at once with 570.124.04

[algosystem]

I’ve reproduced this on two different Fedora 41 computers I have

• both with 570.124.04
• Wayland
• One computer with a 4060, one with a 2060.
• both functioning fine with 1 VRR DP display and 1 VRR HDMI display connected
• Monitors in question: AOC Q27G3XMN and LG 27GL83A-B

As long as I have each monitor connected with HDMI to the same computer, or one connected to DP and one connected to HDMI, it’s fine (VRR works, both running at 144hz)

As soon as I try to connect both to the same card with DP, the Nvidia driver appears to crash (second display doesn’t initialize, primary display freezes/stops updating, forcing a hard reboot (which causes the problem to occur again as soon as GDM comes up).

I’ve tried different cables and ports, and it makes no difference on either computer - I see the same symptom.

I’ve also tried booting into Windows on one of the PCs - there, dual-Displayport out works for both displays with the same cables that fail in Linux, so I know it’s not a monitor/card/cable fault.

nvidia-bug-report.log.gz (385.3 KB)

The only dmesg errors I see from the boots that freeze up due to having 2 DP displays plugged in at once are:

Mar 20 21:51:36.037558 StolidWingnut kernel: nvidia-uvm: Loaded the UVM driver, major device number 510.
Mar 20 21:51:36.132561 StolidWingnut kernel: RPC: Registered named UNIX socket transport module.
Mar 20 21:51:36.132574 StolidWingnut kernel: RPC: Registered udp transport module.
Mar 20 21:51:36.132587 StolidWingnut kernel: RPC: Registered tcp transport module.
Mar 20 21:51:36.132591 StolidWingnut kernel: RPC: Registered tcp-with-tls transport module.
Mar 20 21:51:36.132595 StolidWingnut kernel: RPC: Registered tcp NFSv4.1 backchannel transport module.
Mar 20 21:51:39.331566 StolidWingnut kernel: usb 3-1.4: reset high-speed USB device number 7 using xhci_hcd
Mar 20 21:51:39.863948 StolidWingnut kernel: nvidia-modeset: WARNING: GPU:0: Lost display notification (0:0x00000000); continuing.
Mar 20 21:51:39.863959 StolidWingnut kernel: usbcore: registered new interface driver snd-usb-audio
Mar 20 21:51:39.962559 StolidWingnut kernel: wlp12s0: authenticate with 80:da:13:63:e6:a7 (local address=2a:30:1d:ba:99:87)
Mar 20 21:51:40.447560 StolidWingnut kernel: wlp12s0: send auth to 80:da:13:63:e6:a7 (try 1/3)
Mar 20 21:51:40.461567 StolidWingnut kernel: wlp12s0: authenticate with 80:da:13:63:e6:a7 (local address=2a:30:1d:ba:99:87)
Mar 20 21:51:40.472566 StolidWingnut kernel: wlp12s0: send auth to 80:da:13:63:e6:a7 (try 1/3)
Mar 20 21:51:40.480566 StolidWingnut kernel: wlp12s0: authenticated
Mar 20 21:51:40.480576 StolidWingnut kernel: wlp12s0: associate with 80:da:13:63:e6:a7 (try 1/3)
Mar 20 21:51:40.502569 StolidWingnut kernel: wlp12s0: RX AssocResp from 80:da:13:63:e6:a7 (capab=0x1111 status=30 aid=1)
Mar 20 21:51:40.502578 StolidWingnut kernel: wlp12s0: 80:da:13:63:e6:a7 rejected association temporarily; comeback duration 994 TU (1017 ms)
Mar 20 21:51:41.539566 StolidWingnut kernel: wlp12s0: associate with 80:da:13:63:e6:a7 (try 2/3)
Mar 20 21:51:41.569570 StolidWingnut kernel: wlp12s0: RX AssocResp from 80:da:13:63:e6:a7 (capab=0x1111 status=0 aid=1)
Mar 20 21:51:41.595569 StolidWingnut kernel: wlp12s0: associated
Mar 20 21:51:41.595584 StolidWingnut kernel: wlp12s0: Limiting TX power to 30 (30 - 0) dBm as advertised by 80:da:13:63:e6:a7
Mar 20 21:51:55.393567 StolidWingnut kernel: Bluetooth: hci0: Device setup in 20500359 usecs
Mar 20 21:51:55.393592 StolidWingnut kernel: Bluetooth: hci0: HCI Enhanced Setup Synchronous Connection command is advertised, but not supported.
Mar 20 21:51:55.668560 StolidWingnut kernel: Bluetooth: hci0: AOSP extensions version v1.00
Mar 20 21:51:55.668604 StolidWingnut kernel: Bluetooth: hci0: AOSP quality report is supported
Mar 20 21:51:55.669561 StolidWingnut kernel: Bluetooth: MGMT ver 1.23
Mar 20 21:52:07.649568 StolidWingnut kernel: snd_hda_codec_hdmi hdaudioC1D0: HDMI: invalid ELD data byte 12
Mar 20 21:52:53.817587 StolidWingnut kernel: show_signal: 94 callbacks suppressed
Mar 20 21:52:53.817631 StolidWingnut kernel: traps: gdm[1665] trap int3 ip:7f29851e7068 sp:7ffef6bc7520 error:0 in libglib-2.0.so.0.8200.2[46068,7f29851a1000+a6000]
Mar 20 21:53:01.096562 StolidWingnut kernel: usb 3-1: USB disconnect, device number 2
Mar 20 21:53:01.560302 StolidWingnut kernel: usb 3-1.1: USB disconnect, device number 4
Mar 20 21:53:01.560338 StolidWingnut kernel: usb 3-1.2: USB disconnect, device number 5
Mar 20 21:53:01.560362 StolidWingnut kernel: usb 3-1.3: USB disconnect, device number 6
Mar 20 21:53:01.560398 StolidWingnut kernel: usb 3-1.4: USB disconnect, device number 7
Mar 20 22:00:30.755578 StolidWingnut kernel: [drm:nv_drm_atomic_commit [nvidia_drm]] ERROR [nvidia-drm] [GPU ID 0x00000100] Flip event timeout on head 1
Mar 20 22:00:30.795547 StolidWingnut kernel: traps: gdm[1847] trap int3 ip:7fa063dbd068 sp:7ffeba2aa3f0 error:0 in libglib-2.0.so.0.8200.2[46068,7fa063d77000+a6000]
Mar 20 22:00:49.827580 StolidWingnut kernel: [drm:nv_drm_atomic_commit [nvidia_drm]] ERROR [nvidia-drm] [GPU ID 0x00000100] Flip event timeout on head 0
Mar 20 22:01:15.733553 StolidWingnut kernel: traps: gdm[2026] trap int3 ip:7f9c3c833068 sp:7ffe050460a0 error:0 in libglib-2.0.so.0.8200.2[46068,7f9c3c7ed000+a6000]
Mar 20 22:02:29.847560 StolidWingnut kernel: [drm:nv_drm_atomic_commit [nvidia_drm]] ERROR [nvidia-drm] [GPU ID 0x00000100] Failed to apply atomic modeset. Error code: -22
Mar 20 22:02:32.867600 StolidWingnut kernel: [drm:nv_drm_atomic_commit [nvidia_drm]] ERROR [nvidia-drm] [GPU ID 0x00000100] Flip event timeout on head 0
Mar 20 22:02:35.875683 StolidWingnut kernel: [drm:nv_drm_atomic_commit [nvidia_drm]] ERROR [nvidia-drm] [GPU ID 0x00000100] Flip event timeout on head 1
Mar 20 22:02:38.738550 StolidWingnut kernel: traps: gdm[2157] trap int3 ip:7f0f57cc7068 sp:7ffdd727e250 error:0 in libglib-2.0.so.0.8200.2[46068,7f0f57c81000+a6000]
Mar 20 22:02:38.883558 StolidWingnut kernel: [drm:nv_drm_atomic_commit [nvidia_drm]] ERROR [nvidia-drm] [GPU ID 0x00000100] Flip event timeout on head 0
Mar 20 22:02:41.891564 StolidWingnut kernel: [drm:nv_drm_atomic_commit [nvidia_drm]] ERROR [nvidia-drm] [GPU ID 0x00000100] Flip event timeout on head 1
Mar 20 22:02:59.069552 StolidWingnut kernel: Bluetooth: RFCOMM TTY layer initialized
Mar 20 22:02:59.069618 StolidWingnut kernel: Bluetooth: RFCOMM socket layer initialized
Mar 20 22:02:59.069625 StolidWingnut kernel: Bluetooth: RFCOMM ver 1.11

For now, I’m working around this by just having the secondary display run on HDMI instead of both using DisplayPort, but this seems like an Nvidia Linux driver bug.

[finster]

Sounds like:

Bug: 570.124.04 - freeze on monitor wakeup - "Flip event timeout"

Started right after I upgraded to 570.124.04. Was not present in 570.86.16. To replicate - Lock the system with two monitors. Wait until monitors are switched off. Press any key. Symptoms: Primary monitor wakes up but system is frozen. The other monitor does not wake up. I tried Ctrl+Alt+F1 to reboot, unfortunately that did not work either; it froze the system completely (Capslock no longer lit up on keyboard). System configuration - GPU: RTX 3090 Monitors: Two Operating System: Arch Lin…

[algosystem]

Yep that was it, upgrading to 570.133.07 fixed this issue for me.