RHEL 7.5 workstation - runaway X process - issue with NVIDIA card (M2000) or driver

RedHat support helped me locate the following. I have a single workstation that has twice now seen its X process runaway, at nearly 100% CPU.

Running the NVIDIA-Linux-x86_64-396.54.run driver on RHEL 7.5.

The issue is occurring in the Nvidia Modesetting driver:

Apr 5 10:18:54 slws10 kernel: INFO: task nvidia-modeset:709 blocked for more than 120 seconds.
Apr 5 10:18:54 slws10 kernel: “echo 0 > /proc/sys/kernel/hung_task_timeout_secs” disables this message.
Apr 5 10:18:54 slws10 kernel: nvidia-modeset D ffff89f9d8325ee0 0 709 2 0x00000000
Apr 5 10:18:54 slws10 kernel: Call Trace:
Apr 5 10:18:54 slws10 kernel: [] ? _nv000777kms+0x67/0xa0 [nvidia_modeset]
Apr 5 10:18:54 slws10 kernel: [] schedule+0x29/0x70
Apr 5 10:18:54 slws10 kernel: [] schedule_timeout+0x239/0x2c0
Apr 5 10:18:54 slws10 kernel: [] ? schedule_timeout+0x239/0x2c0
Apr 5 10:18:54 slws10 kernel: [] ? _nv000774kms+0xd/0x90 [nvidia_modeset]
Apr 5 10:18:54 slws10 kernel: [] ? __slab_free+0x81/0x2f0
Apr 5 10:18:54 slws10 kernel: [] __down_common+0xaa/0x104
Apr 5 10:18:54 slws10 kernel: [] ? cascade+0x20/0xc0
Apr 5 10:18:54 slws10 kernel: [] __down+0x1d/0x1f
Apr 5 10:18:54 slws10 kernel: [] down+0x41/0x50
Apr 5 10:18:54 slws10 kernel: [] nvkms_kthread_q_callback+0x4b/0xe0 [nvidia_modeset]
Apr 5 10:18:54 slws10 kernel: [] _main_loop+0x91/0x190 [nvidia]
Apr 5 10:18:54 slws10 kernel: [] ? nv_kthread_q_init+0x120/0x120 [nvidia]
Apr 5 10:18:54 slws10 kernel: [] kthread+0xd1/0xe0
Apr 5 10:18:54 slws10 kernel: [] ? insert_kthread_work+0x40/0x40
Apr 5 10:18:54 slws10 kernel: [] ret_from_fork_nospec_begin+0x21/0x21
Apr 5 10:18:54 slws10 kernel: [] ? insert_kthread_work+0x40/0x40

The above is one of several backtraces out of the Nvidia-modesetting module. You are going to need to contact Nvidia directly for further assistance with this issue as we are unable to debug their code.

[08:14:23 AM]Sangeeta: Hello Mike
[08:15:06 AM]Mike Hackelton: I’m running the NVIDIA driver 396.54. This is the only system this is occurring, and has happended twice now in 10 days.
[08:15:14 AM]Mike Hackelton: hello
[08:16:03 AM]Sangeeta: Can you help me with the system information?
[08:16:37 AM]Mike Hackelton: what do you need?
[08:17:08 AM]Mike Hackelton: HP Z640 workstation with a Quadro M2000
[08:17:15 AM]Mike Hackelton: NVIDIA driver 396.54
[08:18:59 AM]Sangeeta: Okay thank you
[08:19:01 AM]Mike Hackelton: identical to 9 other systems (OS, hardware, patch levels, etc). This workstation and one other are driving 4- 32"4K displays, all others are running 2- 24" 1920x1080 displays
[08:19:15 AM]Sangeeta: I will have this checked with the concerned team and get back to you via e-mail.
[08:19:25 AM]Mike Hackelton: thanks for your help!!
[08:19:38 AM]Sangeeta: Sure, you are welcome.
[08:20:22 AM]Sangeeta: Thank you for contacting Nvidia customer care. Have a great day ahead.
[08:20:36 AM]Mike Hackelton: You as well
[08:20:47 AM]Mike Hackelton: Standing by for more information in email.
[08:20:55 AM]Sangeeta: Sure
[08:20:59 AM]‘Mike Hackelton’ disconnected (‘Concluded by Agent’).

The v396 driver is outdated and out of support. Please check with a current v418.56 driver. If the issue persists, please run nvidia-bug-report.sh as root and attach the resulting .gz file to your post. Hovering the mouse over an existing post of yours will reveal a paperclip icon.
[url]https://devtalk.nvidia.com/default/topic/1043347/announcements/attaching-files-to-forum-topics-posts/[/url]

Thanks! Downloaded and upgraded to driver 418.56. Time will tell if that ultimately fixes the problem.