Could you try Proton-GE 10-12?

I played Starfield back in August for about a month and had no page fault crashes. I started playing Hogwarts Legacy a week ago and have had a few page fault crashes (seems random, so far). Pulled vkd3d version from 10-12 build and currently testing that...

Edit:
Bazzite 121025
Mesa 25.3.0
RX 6700 XT

could you try RADV_DEBUG=nonggc

@Etaash-mathamsetty No dice, same issue occurs.

@Sumwunn This title has some quirks that make earlier Proton versions dangerous to use so I cannot use that version without increasing my risk of an account ban.

@Etaash-mathamsetty No dice, same issue occurs.

@Sumwunn This title has some quirks that make earlier Proton versions dangerous to use so I cannot use that version without increasing my risk of an account ban.

could you try mesa 25.2.x then ?

I can confirm the page faults, just got one within the first 10-15 minutes of the single player campaign:

MY System:

GPU: 9070XT
Mesa: Git(7ed6679361f02ea48562652d11281e927ee06fed)
Kernel: 6.19rc3 (can not move back to stable as of now)
Distri: Arch
DE: KDE 6.5.4
VKD3D-Proton version: Master (https://github.com/HansKristian-Work/vkd3d-proton/commit/4d4abaa230fed54c2bf12b56cf4e4cd2286bd034)

Settings: Autodetected by the game itself, RT active with medium settings.

VRAM and RAM usage have been fine at the time of the pagefault.

amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:7 pasid:32807)
[Sa,  3. Jan 2026, 08:18:20] amdgpu 0000:03:00.0: amdgpu:  Process BlackOpsColdWar pid 11624 thread vkd3d_queue pid 11812
[Sa,  3. Jan 2026, 08:18:20] amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x00008000ffff8000 from client 10
[Sa,  3. Jan 2026, 08:18:20] amdgpu 0000:03:00.0: amdgpu: GCVM_L2_PROTECTION_FAULT_STATUS:0x00701431
[Sa,  3. Jan 2026, 08:18:20] amdgpu 0000:03:00.0: amdgpu:        Faulty UTCL2 client ID: SQC (data) (0xa)
[Sa,  3. Jan 2026, 08:18:20] amdgpu 0000:03:00.0: amdgpu:        MORE_FAULTS: 0x1
[Sa,  3. Jan 2026, 08:18:20] amdgpu 0000:03:00.0: amdgpu:        WALKER_ERROR: 0x0
[Sa,  3. Jan 2026, 08:18:20] amdgpu 0000:03:00.0: amdgpu:        PERMISSION_FAULTS: 0x3
[Sa,  3. Jan 2026, 08:18:20] amdgpu 0000:03:00.0: amdgpu:        MAPPING_ERROR: 0x0
[Sa,  3. Jan 2026, 08:18:20] amdgpu 0000:03:00.0: amdgpu:        RW: 0x0
[Sa,  3. Jan 2026, 08:18:20] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:7 pasid:32807)
[Sa,  3. Jan 2026, 08:18:20] amdgpu 0000:03:00.0: amdgpu:  Process BlackOpsColdWar pid 11624 thread vkd3d_queue pid 11812
[Sa,  3. Jan 2026, 08:18:20] amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x00008000ffff8000 from client 10
[Sa,  3. Jan 2026, 08:18:22] amdgpu 0000:03:00.0: amdgpu: Dumping IP State
[Sa,  3. Jan 2026, 08:18:22] amdgpu 0000:03:00.0: amdgpu: Dumping IP State Completed
[Sa,  3. Jan 2026, 08:18:22] amdgpu 0000:03:00.0: amdgpu: [drm] AMDGPU device coredump file has been created
[Sa,  3. Jan 2026, 08:18:22] amdgpu 0000:03:00.0: amdgpu: [drm] Check your /sys/class/drm/card0/device/devcoredump/data
[Sa,  3. Jan 2026, 08:18:22] amdgpu 0000:03:00.0: amdgpu: ring gfx_0.0.0 timeout, signaled seq=2418799, emitted seq=2418801
[Sa,  3. Jan 2026, 08:18:22] amdgpu 0000:03:00.0: amdgpu:  Process BlackOpsColdWar pid 11624 thread vkd3d_queue pid 11812
[Sa,  3. Jan 2026, 08:18:22] amdgpu 0000:03:00.0: amdgpu: Starting gfx_0.0.0 ring reset
[Sa,  3. Jan 2026, 08:18:22] amdgpu 0000:03:00.0: amdgpu: Ring gfx_0.0.0 reset succeeded

Did some additional tests:

For me using RT in game triggers the page faults. I can reproduce that easily in mission 2 by just enabling RT in the options. This seems to be an issue with RADV as with amdvlk the issue does not show up. In addition the game has rendering issues which do not show up with amdvlk but with radv. See the following issues:

https://gitlab.freedesktop.org/mesa/mesa/-/issues/14594
https://gitlab.freedesktop.org/mesa/mesa/-/issues/14598

Confirmed that #2781 will resolve this issue for my system (RX 7800 XT).