[6.10-rc4] [AMD-IOMMU] Wake up from suspend: "amdgpu: PSP load kdb failed!"

Brief summary of the problem:

Suspend PC on kernel 6.10-rc1 to 6.10-rc4. Wake up from suspend. Observe black screen. Reset PC (can use SYSRQ keys). See logs from previous boot.

Working correctly in 6.9.3.

Hardware description:

• CPU: Ryzen 7950X
• GPU: Radeon 6900 XT
• System Memory: 32GB
• Display: 3840x2160@60
• Type of Display Connection: DP

System information:

• Distro name and Version: Manjaro Linux
• Kernel version: 6.10-rc4

How to reproduce the issue:

• suspend
• wake up

Logs (6.10-rc4):

amdgpu 0000:03:00.0: amdgpu: PSP load kdb failed!
amdgpu 0000:03:00.0: amdgpu: PSP resume failed
[drm:amdgpu_device_fw_loading [amdgpu]] *ERROR* resume of IP block <psp> failed -62
amdgpu 0000:03:00.0: amdgpu: amdgpu_device_ip_resume failed (-62).
amdgpu 0000:03:00.0: PM: dpm_run_callback(): pci_pm_resume returns -62
amdgpu 0000:03:00.0: PM: failed to resume async: error -62

amdgpu: Move buffer fallback to memcpy unavailable

amdgpu 0000:03:00.0: [drm] REG_WAIT timeout 1us * 1000 tries - dcn20_dpp_pg_control line:560
amdgpu 0000:03:00.0: [drm] REG_WAIT timeout 1us * 1000 tries - dcn20_hubp_pg_control line:634
amdgpu 0000:03:00.0: [drm] REG_WAIT timeout 1us * 1000 tries - dcn20_dpp_pg_control line:568
amdgpu 0000:03:00.0: [drm] REG_WAIT timeout 1us * 1000 tries - dcn20_hubp_pg_control line:642
amdgpu 0000:03:00.0: [drm] REG_WAIT timeout 1us * 1000 tries - dcn20_dpp_pg_control line:576
amdgpu 0000:03:00.0: [drm] REG_WAIT timeout 1us * 1000 tries - dcn20_hubp_pg_control line:650
amdgpu 0000:03:00.0: [drm] REG_WAIT timeout 1us * 1000 tries - dcn20_dpp_pg_control line:584
amdgpu 0000:03:00.0: [drm] REG_WAIT timeout 1us * 1000 tries - dcn20_hubp_pg_control line:658
amdgpu 0000:03:00.0: [drm] REG_WAIT timeout 1us * 1000 tries - dcn20_dpp_pg_control line:592
amdgpu 0000:03:00.0: [drm] REG_WAIT timeout 1us * 1000 tries - dcn20_hubp_pg_control line:666
amdgpu 0000:03:00.0: [drm] REG_WAIT timeout 1us * 1000 tries - dcn20_dsc_pg_control line:489
amdgpu 0000:03:00.0: [drm] REG_WAIT timeout 1us * 1000 tries - dcn20_dsc_pg_control line:497
amdgpu 0000:03:00.0: [drm] REG_WAIT timeout 1us * 1000 tries - dcn20_dsc_pg_control line:505
amdgpu 0000:03:00.0: [drm] REG_WAIT timeout 1us * 1000 tries - dcn20_dsc_pg_control line:513
amdgpu 0000:03:00.0: [drm] REG_WAIT timeout 1us * 1000 tries - dcn20_dsc_pg_control line:521
amdgpu 0000:03:00.0: [drm] REG_WAIT timeout 1us * 1000 tries - dcn20_dsc_pg_control line:529

[drm:amdgpu_job_timedout [amdgpu]] *ERROR* ring sdma0 timeout, signaled seq=1220, emitted seq=1222
amdgpu 0000:03:00.0: amdgpu: GPU reset begin!