i915 boot-time hard crash with screen corruption (kernel NULL pointer deref) in 5.8 when WD19TB hub plugged in, in RIP intel_psr_atomic_check

This bug is nearly always reproducible, i.e. in order to boot successfully the Dell 5540 (00:02.0 VGA compatible controller [0300]: Intel Corporation UHD Graphics 630 (Mobile) [8086:3e9b] (rev 02)) needs to be unplugged from the Dell WD19TB hub until kernel passed i915 initialization. The crash happens during i915 initialization at boot time.

In versions prior to 5.8 (5.7 and 5.6) the boot screen would scroll past the crash stack trace and the system would be at least somewhat functional with systemd being able to do at least some process management. In version 5.8 the driver crash happens immediately, the crash is complete and even causes screen corruption. System is completely unresponsive and requires a hard power-off.

The screenshot of the crashed laptop is attached and some transcription of the stack follows:

BUG: kernel NULL pointer dereference, address: 0000000000078
RIP: 010:intel_psr_atomic_check+0x33/0xa0 [i915]

intel_digital_connector_atomic_check+0x4e/0x130 [i915]
intel_dp_mst_atomic_check+0x49/0x1e0 [i915]
? handle_conflicting_encoders+0x1ce/0x260 [drm_kms_helper]
intel_atomic_check+0x90/0x2360 [i915]
...
drm_fb_helper_hotplug_event.part.0+0xa3/0xc0 [drm_kms_helper]