Hi, I have a completely different issue as @matte-schwartz, using the same form to provide feedback too!

Compatibility Report

• Name of the game with compatibility issues: Microsoft Flight Simulator 2024
• Steam AppID of the game: 2537590

System Information

• GPU: RTX 3090
• Video driver version: nvidia 565.57.01
• Kernel version: 6.12.0
• Link to full system information report as Gist: https://gist.github.com/trivarialthea/bfd3b9c03551cd8da7f1ac77e967515c
• Proton version: 1731365841 experimental-9.0-20241111

I confirm:

• [ ] that I haven't found an existing compatibility report for this game.
• [x] that I have checked whether there are updates for my system available.

https://github.com/trivarialthea/syrup-verdict-operator-morphine-zoology/blob/main/steam-2537590.tar.lzma (file is too big in .gz export, so uploaded it to a repo I created as a .lzma export, to extract use tar --lzma -xvf)

Symptoms

Game crashes freezes after reaching the last simobject, otherwise works normally to that point. (Except for cursor issues, but those are also present on windows)

Reproduction

1. Login to Xbox
2. Wait for the game to load to that point (around 49%)

I get the "!status && vkCreateGraphicsPipeline" assert 100% of the time, twice in fact. The first time early in the loading process, the second time it coincides with the game freezing. But if I turn Proton logging on then I don't see the assert, though the overall behavior is the same. Grepped the log for that assert but didn't find any mentions of it.

In any case the game stops loading for me at "54% - Loading world data .."

Kubuntu 24.10, kernel 6.11, Mesa 24.2.7-kisak1-o, 7950X3D, 7900 XTX, amdgpu open source driver, Wayland, today's Proton Experimental, Steam official .deb package

Any progress in Proton development regarding this so far?

And if anyone is willing to explain, what's going on in Proton that is causing the game to not get past the loading screen?

Having the same problem here as @matte-schwartz on Manjaro Unstable branch fully up to date and using proton experimental.

tried winedlloverrides for things like d3d12... still getting the same error

if you add the PROTON_LOG=1 launch option, it doesn't immediately throw the vkcreategraphicspipeline error pop-up... but if you attempt to load into any flight it crashes (CTD)... and the log file is multiple GB in size

log shows a bunch of nvngx.dll errors, even though running the latest proprietary drivers from nvidia

I am also experiencing the same problem as @matte-schwartz on Fedora 41 Workstation. I have tried multiple proton versions, (Proton-GE 18 and 20, experimental, default proton,) but all of them result in the same error. Since listing the below specs and reporting on ProtonDB, I have updated all system packages but am still getting the same error.

Specs:

• Distro: Fedora Linux 41 (Workstation Edition)
• Kernel: 6.11.7-300.fc41.x86_64
• RAM: 32 GB
• GPU Driver: 4.6 Mesa 24.2.7
• GPU: AMD Radeon RX 6700 XT (radeonsi, navi22, LLVM 19.1.0, DRM 3.59, 6.11.7-300.fc41.x86_64)
• CPU: AMD Ryzen 5 3600 6-Core
• Steam is the native RPM, not Flatpak

Edit:
I can confirm that this bug is not happening on Windows. I installed the game on my windows install on the same hardware and did not have this issue.

The Vulkan vkCreateGraphicsPipelines error in the initial post also seems to occur with 100% frequency on the Steam Deck from what I can tell. Tested with Proton, Proton Experimental, and Proton GE 9-20.

Patched the game to 1.1.9.0 and tried again, same result as before.

Actually I take that back. Before it stopped at 54%. Now it stops at 55%! :+1:

I'm on EndeavourOS, it's stopping at 55% with this in my proton logs every time:

SPIR-V parsing FAILED:
    Unrecognized branch from block pos 113 (id=6205) to block pos 121 (id=6206)
    Inner construct 'selection': 109 -> 114  (merge=114 then=110 else=114)
    Outer construct 'selection': 108 -> 120  (merge=120 then=109 else=117)
SPIR-V parsing FAILED:
    In file ../src/compiler/spirv/vtn_structured_cfg.c:965
    Unable to identify branch type
    0 bytes into the SPIR-V binary

I'm running:
MESA_LOG_LEVEL=debug WINEDEBUG="+timestamp,+pid,+tid,-all,+err" PROTON_LOG=1 MANGOHUD_CONFIG=full gamescope --mangoapp -e -f -w 3440 -h 1440 -W 3440 -H 1440 -- %command%

Is this a driver bug by any chance? If so, I'd be quite shocked considering this is happening on both RADV and Nvidia.

Perhaps we could fill a bug report for vkd3d-proton to investigate this.

some recent reports on protondb are saying that the game is somewhat working on intel gpus... may be RADV related... did you guys try on amdvlk?

Here's a similar issue from 2023 on vkd3d-proton with the same symptoms. Looks like there was an update 2 days ago for this submodule, maybe we need to wait for it to get pulled into proton proper...?

nixos, kernel 6.12, rx6800xt, today's proton experimental [bleeding-edge] and GE-proton9.20, wayland

did you guys try on amdvlk?

On my end it looks like this with both mentioned proton versions

followed by

Removing admvlk I experience the exact same thing as OP and @StagnationPoint.

Looks like there was an update 2 days ago for this submodule, maybe we need to wait for it to get pulled into proton proper...?

:crossed_fingers:

This mornings proton-experimental update did not change anything

Edit: updating drivers fixed my issues.

Compatibility Report

• Name of the game with compatibility issues: Microsoft Flight Simulator 2024
• Steam AppID of the game: 2537590

System Information

• GPU: Intel Mesa Intel Arc A770 (DG2)
• Video driver version: 4.6 Mesa 24.0.9-0ubuntu0.2
• Kernel version: 6.8.0-49-generic
• Link to full system information report as Gist: https://gist.github.com/sivertac/4405725c23ebb8e8ebecd34a6fe5df9b
• Proton version: proton-9.0-3e (3c3d468971353a5302f401a1e017730970ab8921)

I confirm:

• [ ] that I haven't found an existing compatibility report for this game.
• [x] that I have checked whether there are updates for my system available.

steam-2537590.log.tar.gz

Symptoms

Missing textures, game seem to run okayish, some stuttering/missing frames, however these issues are gone when using bleeding-edge. A lot of textures are not present making the game unplayable.

Not getting stuck at the loading screen:

Past the loading screen, maybe some missing textures here:

Main menu, UI seem to work:

Graphics settings:

Create free flight, missing textures:

Start flight, in cockpit, missing textures, the mechanics seems to be functional:

Exterior view, missing textures:

Content manager storage, seems like it is not capable of downloading any content:

It's not playable due to the missing content, however the mechanics seems to work fine. Wonder if Microsoft is blocking non-windows machines unintentionally/intentionally from downloading content

Reproduction

Use Intel GPU.
Add VKD3D_CONFIG="enable_experimental_features" to launch options.

Update

Updating mesa driver and not using VKD3D_CONFIG="enable_experimental_features" fixed the issues.

Ubuntu 24.10, standard 6.11 kernel, RX6700XT, radv driver - The same happens (Impossible to create DirectX12 device... etc) if you use RADV_DEBUG=llvm to use the alternative compiler. Intel driver (iGPU) doesn't get the errors, but is very slow... gets to asking about narration but not responding to clicks (AlderLake-S GT1)

game is starting with latest proton experimental :party:

Navi 23 RX 6600, 6.12.1-arch1-1, Mesa 24.3.0, radv

Tried latest proton experimental build on Steam Deck, I don't get the vkcreategraphicspreset error anymore but I freeze when loading objects, specifically at 1025/1026.

Will continue to try and report back with any changes.

Edit #1: got to the main menu in Desktop Mode!!!! Making progress! Still getting stuck at the same object level in gamemode.

after i fixed my cpu fan and reinstalled the game on my linux nvme drive, it's running quiet well on medium,fsr2 balanced, on my mediocre machine : )

My experience is the same as @rariema. Today's proton experimental [bleeding-edge] does indeed fix the previous issues with !status && vkCreateGraphicsPipeline - the game launches no issues! Flight works fine, map tiles and content imagery seems to work fine.

Edit: opentrack with "neuralnet tracker", set to "freetrack 2.0 enhanced" with "use trackir, disable freetrack" works like a charm using opentrack-launcher

Any idea why the game runs in Desktop mode for the steam deck but in Game Mode it freezes loading SimObjects?

My experience is the same as @rariema. Today's proton experimental [bleeding-edge] does indeed fix the previous issues with !status && vkCreateGraphicsPipeline - the game launches no issues! Flight works fine, map tiles and content imagery seems to work fine.

I can also confirm that the latest proton experimental fixed that issue. The game loads to menu, and I can start a flight. Satellite imagery seems to be working fine, and the flight mechanics work.

I am experiencing a few other issues that aren't occurring on my Windows install, though. (Unfortunately, I don't have time at the moment to document/test these fully, but I'll put them here for future reference.)

• Loading video doesn't loop
• Clouds are pixelated/static
• The streaming scenery isn't fully working (Scenery is a mix of hi def trees and weird green polygons extending from the terrain, for ex.)
• Does not detect my Saitek x45 HOTAS (though that may be related to missing some drivers)

Fedora Linux 41 - 32 GB RAM - 4.6 Mesa 24.2.7 - AMD Radeon RX 6700 XT - AMD Ryzen 5 3600 6-Core

• Loading video doesn't loop

confirmed

• The streaming scenery isn't fully working (Scenery is a mix of hi def trees and weird green polygons extending from the terrain, for ex.)

for me it worked better than the last days in windows, it seemed fine for the first time - streaming is still a bit hit or miss overall

I seem to get a consistent hang in gamescope-session (or gamescope -e -f --hdr-enabled -h 2160 -w 7680 -r 240 -- steam -steamdeck -steamos3 -steampal -gamepadui from TTY) when simobjects finish loading, only if HDR is enabled. This happens on both OLED Deck running SteamOS and a 7900XTX rig running Arch Linux.

steam-2537590-gamescope-session-hdr.zip

The hang can be avoided if HDR is disabled before launching gamescope, or if ENABLE_GAMESCOPE_WSI=0 is set in the game's launch options within gamescope-session.

I also seem to get a spurious crash on my desktop 7900XTX rig when the game finishes loading and brings up the main menu, or when I enter into a flight. Proton log from one such crash while loading a flight:

steam-2537590-crash-before-flight.zip

For steam deck users

Run MSFS24 with Proton Experimental - Bleeding Edge beta. I have it working on that proton now.

For steam deck users

Run MSFS24 with Proton Experimental - Bleeding Edge beta. I have it working on that proton now.

Hi, with all videos running and not having any video decoding problem?

Replying to https://github.com/ValveSoftware/Proton/issues/8255#issuecomment-2499094228

Videos in the initial load still break but after that textures and everything work fine for me.

Replying to https://github.com/ValveSoftware/Proton/issues/8255#issuecomment-2499121024

ok, same here... so we will have to wait for GE-Proton I think (for vid playback)

Replying to https://github.com/ValveSoftware/Proton/issues/8255#issuecomment-2499123868

I wonder if GE-Proton will work, I tried 9-20 and still got the vkcreategraphicspreset error, when I use Proton-Experimental I don't see that issue.

You can fix the video issue while using proton experimental, using ffmpeg and convert the mp4 files using the VP9 codec.

Konsole:

ffmpeg -i Logo_Asobo.mp4 -c:v libvpx-vp9 -b:v 2M -c:a libopus output.mp4

Then rename the original e.g. Logo_Asobo.mp4 to something else and put the output.mp4 with the name of Logo_Asobo.mp4 in the videos folder of MSFS2024.

This should be: SteamLibrary/steamapps/common/Limitless/Packages/fs-base-videos/Videos

Just tried it: worked.

Replying to https://github.com/ValveSoftware/Proton/issues/8255#issuecomment-2499199265

Excellent discovery!!! Thanks for the heads up, will try it later tonight.

Replying to [#8255 (comment)](https://github.com/ValveSoftware/Proton/issues/8255#issuecomment-2499199265)

Excellent discovery!!! Thanks for the heads up, will try it later tonight.

if you are having problems with it, I may prepare a private download (not public cause of copyright things)

Replying to https://github.com/ValveSoftware/Proton/issues/8255#issuecomment-2498757785

same, it may get pass the loading objects part with HDR on at random, but it crashes most of the time.

I was also able to run with proton experimental bleeding edge, game mostly works fine.

7950X3D
7900XTX
Fedora 41 + KDE Plasma 6

$ uname -r
6.12.1-400.vanilla.fc41.x86_64

The exception is that after a few flights, there is a hangup which I suspect to be related to amd gpu drivers that causes a crash when clicking on the next mission/flight in career. Entire desktop freezes for a time, then recovers and game is solid locked, but can be killed and restarted.

There is infinite spam in journalctl during the game running (hundreds to thousands of this per second):
Nov 27 06:57:30 <hostname> kwin_wayland[2468]: kwin_wayland_drm: atomic commit failed: Invalid argument

And here is the amd (timeout) crash of the game (only included relevant, non-repeated lines):

Nov 27 07:19:40 <hostname> kwin_wayland[2468]: kwin_wayland_drm: atomic commit failed: Invalid argument
Nov 27 07:19:45 <hostname> kwin_wayland[2468]: kwin_wayland_drm: Pageflip timed out! This is a kernel bug
Nov 27 07:19:49 <hostname> kernel: amdgpu 0000:03:00.0: amdgpu: Dumping IP State
Nov 27 07:19:49 <hostname> kernel: amdgpu 0000:03:00.0: amdgpu: Dumping IP State Completed
Nov 27 07:19:49 <hostname> kernel: amdgpu 0000:03:00.0: amdgpu: ring gfx_0.0.0 timeout, signaled seq=64395085, emitted seq=64395087
Nov 27 07:19:49 <hostname> kernel: amdgpu 0000:03:00.0: amdgpu: Process information: process ACE: WinMain_Z  pid 145033 thread vkd3d_queue pid 145196
Nov 27 07:19:49 <hostname> kernel: gmc_v11_0_process_interrupt: 4 callbacks suppressed
Nov 27 07:19:49 <hostname> kernel: amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:5 pasid:32822)
Nov 27 07:19:49 <hostname> kernel: amdgpu 0000:03:00.0: amdgpu:  in process ACE: WinMain_Z  pid 145033 thread vkd3d_queue pid 145196)
Nov 27 07:19:49 <hostname> kernel: amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x0000000000000000 from client 10
Nov 27 07:19:49 <hostname> kernel: amdgpu 0000:03:00.0: amdgpu: GCVM_L2_PROTECTION_FAULT_STATUS:0x00501431
Nov 27 07:19:49 <hostname> kernel: amdgpu 0000:03:00.0: amdgpu:          Faulty UTCL2 client ID: SQC (data) (0xa)
Nov 27 07:19:49 <hostname> kernel: amdgpu 0000:03:00.0: amdgpu:          MORE_FAULTS: 0x1
Nov 27 07:19:49 <hostname> kernel: amdgpu 0000:03:00.0: amdgpu:          WALKER_ERROR: 0x0
Nov 27 07:19:49 <hostname> kernel: amdgpu 0000:03:00.0: amdgpu:          PERMISSION_FAULTS: 0x3
Nov 27 07:19:49 <hostname> kernel: amdgpu 0000:03:00.0: amdgpu:          MAPPING_ERROR: 0x0
Nov 27 07:19:49 <hostname> kernel: amdgpu 0000:03:00.0: amdgpu:          RW: 0x0
Nov 27 07:19:49 <hostname> steam[4207]: radv/amdgpu: The CS has been cancelled because the context is lost. This context is innocent.
Nov 27 07:19:49 <hostname> steam[4207]: radv: GPUVM fault detected at address 0x00000000.
Nov 27 07:19:49 <hostname> steam[4207]: GCVM_L2_PROTECTION_FAULT_STATUS: 0x501431
Nov 27 07:19:49 <hostname> steam[4207]:          CLIENT_ID: (SQC (data)) 0xa
Nov 27 07:19:49 <hostname> steam[4207]:          MORE_FAULTS: 1
Nov 27 07:19:49 <hostname> steam[4207]:          WALKER_ERROR: 0
Nov 27 07:19:49 <hostname> steam[4207]:          PERMISSION_FAULTS: 3
Nov 27 07:19:49 <hostname> steam[4207]:          MAPPING_ERROR: 0
Nov 27 07:19:49 <hostname> steam[4207]:          RW: 0

After this happens, the game can be killed by clicking stop in steam, fired back up and go again for some undetermined amount of time.

Anyone got it working on Nvidia Hardware?? I tried a clean install, removing all the compatdata for the Game. With GE-Proton9-20 the Game Launches (with the "vkCreateGraphicsPipeline" Error). But with Experimental, it crashes right after launch. The X-Box Login didn't even load fully. steam-2537590.log

AMD Ryzen 7 5800 8-Core Processor
20.04.6 LTS (64 bit) Kernel Version: 5.15.0-126-generic
NVIDIA GeForce RTX 3070 nvida-driver-560

The NVIDIA driver is broken in general when it comes to DX12 (usually games run but with a massive performance penalty) and completely broken when it comes to specific titles like MSFS (2020 & 2024).

See for example here:
https://forums.developer.nvidia.com/t/microsoft-flight-simulator-crashes-with-status-vkacquirenextimagekhr-while-loading-on-wayland-since-555/305443

HDR works under wayland by building proton-tkg bleeding edge with wayland support, unfortunately the mouse pointer is not working.

Beyond of scope so no help with this, but if anyone wants to try it, you will need

1. install vk_hdr_layer (google it)
2. clone wine-tkg-git, set _wayland_driver="false" to "true" on proton-tkg.cfg before building proton-tkg bleeding edge
3. add key HKCU\Software\Wine\Drivers\Graphics = x11,wayland (tip: protontricks -c "wine regedit" 2537590)
4. launch game with DXVK_HDR=1 ENABLE_HDR_WSI=1 DISPLAY= %command% -FastLaunch

was: "I get really bad performance with rising LOD levels. Gpu load drops, MSFS dev mode FPS meter says: "limited by RdrThread""

never mind, seems like the reason where my experiments with optiscaler

I'm interested in getting the game soon, so could anyone tell me how the game runs under Proton now? I've checked ProtonDB and there's only one recent report that seems promising.

I'm interested in getting the game soon, so could anyone tell me how the game runs under Proton now? I've checked ProtonDB and there's only one recent report that seems promising.

I think depends on if you have nvidia or amd gpu. I have nvidia and as soon as I go nearby the terrain it freezes and crash.
And at least for me carrier does not work at all (but I am reading is a common issue).
At the moment using proton bleed-edge with these options:
PROTON_ENABLE_NVAPI=1 PROTON_LOG=1 %command% -FastLaunch

Strange thing is that if I remove PROTON_LOG=1 it crash on loading with "!status &&vkCreateGraphicsPipelines" error message.
Someone suggested to use VKD3D_CONFIG="enable_experimental_features" but in my system does not seem to do anything different.

I removed a lot of stuttering installing on top of the generated prefix a compiled VKD3D and DXVK with gplasync (but I guess if it is using VKD3D then the DXVK part would not do anything).

I hope it get fixed and be able to run like 2020.

On AMD it runs quiet well, no crashes or serious issues on my end.
That said, the sim needs a strong system to play nicely and the hole experience is still a bit early access like.

One Bug i have since ms2020, (maybe kde related, dunno), game starts in a slightly wrong borderless fs resolution, i have to exit end enter fullscreen mode to get the right res.

One Bug i have since ms2020, (maybe kde related, dunno), game starts in a slightly wrong borderless fs resolution, i have to exit end enter fullscreen mode to get the right res.

I get this bug on Fedora 41 Gnome as well, so probably not a KDE issue.

It seems like ProtonGE 21 works now for MSFS 2024. I spent about 15 minutes in a flight on my Steam Deck, performance wasn't great (6 to 30 fps), and frequent audio stutters and glitches, but it seems to work for the most part.

The first load takes a long time, (specially the VFS (?) packages part). I also had to go into desktop mode to get it to finish the first load.

I'll try to make a ProtonDB report when I can.

I have nvidia and as soon as I go nearby the terrain it freezes and crash.

This is the same issue I am having, the game loads fine and I can fly decently. But as soon as I get about 100 ft from the ground the game freezes with the famous "vkCreateGraphicsPipeline" error. @RW74 did you get any closer to a solution?

I have nvidia and as soon as I go nearby the terrain it freezes and crash.

This is the same issue I am having, the game loads fine and I can fly decently. But as soon as I get about 100 ft from the ground the game freezes with the famous "vkCreateGraphicsPipeline" error. @RW74 did you get any closer to a solution?

I am not and kind of lost... it seems affecting only nvidia cards and I used to have issues with dx12 on msfs2020 too (I was using dx11), maybe it is correlated, something that microsoft use with dx12 that is not working on nvidia but no idea where to look.

Replying to https://github.com/ValveSoftware/Proton/issues/8255#issuecomment-2566458729

It's a general Nvidia driver problem in DX12 games under Proton

Testing the latest beta branch for the upcoming Sim Update 1 release and the simulator crashes at startup. What launch option should I use to collect logs to narrow down the reason?

Current Hardware Info:
OS: Fedora Linux, Silverblue 41
Kernel: Linux 6.12.10-200
Desktop Environment: GNOME 47
Compositor: Wayland / Mutter
CPU: AMD Ryzen 9 9950X
GPU: AMD Radeon RX 7900 XTX (RADV)
Steam Installation Method: Flatpak

Same for me.
You can get the log with "PROTON_LOG=1 %command%"

I found:

Unhandled exception: unimplemented function KERNEL32.dll.WerRegisterCustomMetadata called in 64-bit code (0x006ffffffa3a58).
417.006:0294:0298:fixme:dbghelp:elf_search_auxv can't find symbol in module
Register dump:
..

steam-2537590.log

Still not fixed on NVIDIA driver version 570.86.10

Also crashes entirely, pretty much immediately on the Sim Update 1 beta:
steam-2537590_sim_update_1_beta.log

As a reminder to everyone:
internal NVIDIA driver bugs can only be fixed by NVIDIA. Directly report issues to them as much as possible.

In my case I do get the same assertion error, but game continues to load, gets to the start menu and lets me start a free flight. However when the airplane gets close to ground, the game throws another of the exact same assertion error and freezes (this is with Proton GE 9-23, but I believe it's an issue with Proton itself)

My NVIDIA driver version is 565.77-11, kernel version 6.12.9-arch1-1

I was able to solve the issue of crashing when going near the ground by setting "Grass" to OFF in the advanced graphics settings. I still get a supposedly harmless assertion error when starting the game but haven't noticed any other issues.

I'm using the 550.144.03 NVIDIA driver with Proton GE 9-23.

Turning off grass worked, tysm! Btw am I the only one experiencing instability with KDE Plasma when playing the game? It crashes and switches to software rendering or becomes unresponsive. Only happens when playing MSFS 2024

I've tested changing in-game visual settings and that issue appears to be Nvidia specific. While attempting to launch the beta for Sim Update 1 with my AMD graphics card using RADV, the game still crashes at start regardless of in-game settings. I've included a log after using PROTON_LOG=1.

steam-2537590.log

Seems to be a dxil-spirv bug that was not caught by spirv-val. I'll look into it.

Seems to be a dxil-spirv bug that was not caught by spirv-val. I'll look into it.

I just tested the latest bleeding_edge version with your changes on my 3090 and the issue seems to be fixed indeed!

Thank you so much! :)

I can also confirm, both the error at launch and error when grass is enabled are fixed, thank you!

Replying to https://github.com/ValveSoftware/Proton/issues/8255#issuecomment-2612190638

I wrote a patch for wine and patched proton-ge with it to get the su1-beta running.
It adds WerRegisterCustomMetadata to wine werhapi. Maybe someone can review it and add it to wine?

from the log:
111.399:012c:0130:fixme:seh:WerRegisterCustomMetadata (L"PlayerSessionID", L"89635958-BEFA-4C57-B1FE-653EB820779B") stub

werapi-custom-metadata.patch.txt

I've found some small issues with the latest Proton:

• Game freezes when going into fullscreen and then back to windowed using Alt+Enter during a flight
• Take off and landing messages in mission objectives window are empty in career mode

And also an issue I haven't been able to reproduce consistently, the game sometimes gets limited to 60W power usage in the GPU when launching it as if on battery power, and stays that way until I close the game, even though the power limit is way above that. I have an NVIDIA GPU with driver version 570.86.16-2, I have to restart the game every time this happens

I'm using Proton GE 9-24

@rariema any news on your patch? It would be great if it could be merged

No news on this - i was short on time and the policy of wine states, that only code used by an app will get merged, so i waited for SU1. Now that SU1 is out and is still using this i will look into it asap.

edit: just filed a merge request on winehq 🤞

Congratulations @rariema!

thx : )

@kisak-valve
now that the patch has been merged and landed in wine 10.3, how can i request a backport to (experimental) proton?
https://gitlab.winehq.org/wine/wine/-/commit/9329843456750372e771d4967190374554d20523

edit:
https://github.com/ValveSoftware/wine/pull/273

In case anyone is reading this and just wants to play Microsoft Flight Simulator 2024 on the SU1 update right now:

Here's the fork that I'm using to play:
https://github.com/fxtentacle/proton-ge-custom/tree/proton-ge-hajo1

And here's a precompiled .tar.gz ready to be extracted into ~/.steam/root/compatibilitytools.d:
https://github.com/fxtentacle/proton-ge-custom/releases/tag/msfs-su1-hotfix

The patch has landed in proton-ge-26.
But something seems to be wrong, because MSFS 2024 crashes with exactly the same error as if the patch had not been applied. Unfortunately, it is not possible to create a ticket with proton-ge. Can anyone confirm that MSFS2024 does not work with Proton-GE-26?
With @fxtentacle fork it`s working flawlessly.

confirmed. so something went wrong in patching Proton-GE-26

@rariema Thank you so much! I apologize for the delay - it has been busy over here and it took me a few days to respond to your request. I have just cherry-picked the fix to experimental-9, so it should be showing up in bleeding-edge in a few minutes. Please let me know if this doesn't fix the crashing for folks - it helped in our initial testing :)

tada!
Glad i could contribute a little. Linux gaming has become a bliss. : )

Compatibility Report

• Name of the game with compatibility issues: Microsoft Flight Simulator 2024
• Steam AppID of the game: 2537590

System Information

• GPU: AMD Radeon RX 7800 XT
• Video driver version: Mesa 25.0.2
• Kernel version: 6.14.0
• Link to full system information report as Gist: https://gist.github.com/SoberSquid/d041ed2c7ca091f2e14f1561374b23e0
• Proton version: Proton Experimental

I confirm:

• [ ] that I haven't found an existing compatibility report for this game.
• [ x ] that I have checked whether there are updates for my system available.

Log uploaded as compressed tar.gz

Symptoms

Rolling Cache seems broken; No ROLLINGCACHE.CCC file is generated by default settings; areas previously flown in need to re-download each time, resulting in slower than expected startup in these areas.

Reproduction

Attempting to change the rolling Cache (Settings>General>Online) results in no changes; it reflects the path as the /AppData/Roaming/Microsoft Flight Simulator 2024/ folder, changing the Rolling Cache Limit states "Rolling cache path is invalid", changing the Rolling Cache Path (To lets say, C:\users) shows a message stating "Selected size for the rolling cache is not a valid numeric value". Going in and setting the Rolling Cache Limit again after this appears to function (No error is given), but does not pop up the expected window that would show that it is generating the rolling cache, and no file is generated.

steam-2537590.tar.gz

All right, i'll have a look at that after my virtual flight.

Hello @Ienien77, please add PROTON_LOG=1 %command% to the game's launch options and attach the generated $HOME/steam-$APPID.log to this issue report as a file. (Proton logs compress well if needed.)

Hello @Ienien77, please add PROTON_LOG=1 %command% to the game's launch options and attach the generated $HOME/steam-$APPID.log to this issue report as a file. (Proton logs compress well if needed.)

Here you go :
steam-2537590.zip
I start MSFS2024 and start a flight with the addon

When i run the required programm, I got this error : Failed to run module constructor due to (null) assembly:Z:\mnt\HDD_4TO\SteamLibrary\steamapps\common\Proton - Experimental\files\share\wine\mono\wine-mono-10.0.0\lib\mono\4.5\mscorlib.dll type:TypeInitializationException member:(null) And since it looks like gibberish to my eyes, i hope someone here will be able to guide me into understanding that stuff. Tell me if i should add infos to this post (or even if it's in the right section)

I think this is probably caused by the earlier NullReferenceException. Running with WINE_MONO_TRACE=E:System.NullReferenceException may provide more information.

I think this is probably caused by the earlier NullReferenceException. Running with WINE_MONO_TRACE=E:System.NullReferenceException may provide more information.

As a side note, the NullReferenceException doesn't happen when trying to run the addon on MSFS2020 (Which it is also compatible with).

However what is the same across both 2020 and 2024 are attempts to access a DLL called FenixBootstrapper.resources.dll which doesn't exist on neither my Proton install nor on a Windows install I have.

My guess is that the DLL is supposed to be extracted at runtime from the FenixBootstrapper.exe but for some reason this isn't happening under proton.

FWIW here's my log from MSFS2020:

steam-1250410.zip

Checkes for .resources.dll files are done optimistically and handle failure. They're usually benign.

When i run the required programm, I got this error : Failed to run module constructor due to (null) assembly:Z:\mnt\HDD_4TO\SteamLibrary\steamapps\common\Proton - Experimental\files\share\wine\mono\wine-mono-10.0.0\lib\mono\4.5\mscorlib.dll type:TypeInitializationException member:(null) And since it looks like gibberish to my eyes, i hope someone here will be able to guide me into understanding that stuff. Tell me if i should add infos to this post (or even if it's in the right section)

I think this is probably caused by the earlier NullReferenceException. Running with WINE_MONO_TRACE=E:System.NullReferenceException may provide more information.

Hey,
Finally had time to go back to it and run what you suggested.
With MSFS2024 and a flight loaded with Fenix A320 i run this command :

STEAM_COMPAT_DATA_PATH="/home/aure/.local/share/Steam/steamapps/compatdata/2537590/" STEAM_COMPAT_CLIENT_INSTALL_PATH="/home/aure/.local/share/Steam/steamapps/common/Proton 10.0/" WINE_MONO_TRACE=E:System.NullReferenceException WINEPREFIX=$PWD "/home/aure/.local/share/Steam/steamapps/common/Proton - Experimental/proton" run drive_c/Program\ Files/FenixSim\ A320/Fenix.exe

[0000000000000988:] EXCEPTION handling: System.NullReferenceException: Object reference not set to an instance of an object

"<unnamed thread>" tid=0000000000000988 this=0000000004a00130 , thread handle : 00000000020cb0f0, state : not waiting
  at _pmGTaW9nUOjSxRqkbPV3aw5SqFf._k7iWq7JuqXFAC1IDQvwFGfvHnEo._aF1LyXQMWFy66yf2JebAMK4DZJk () [0x00000] in <f7ab7bd52c514bc88db790e398e47ae1>:0
  at _pmGTaW9nUOjSxRqkbPV3aw5SqFf._k7iWq7JuqXFAC1IDQvwFGfvHnEo._HD9qu3uWFYivE5nsGFNTEmTqQQe () [0x00000] in <f7ab7bd52c514bc88db790e398e47ae1>:0
  at <Module>..cctor () [0x0005f] in <f7ab7bd52c514bc88db790e398e47ae1>:0
  at (wrapper runtime-invoke) object.runtime_invoke_void (object,intptr,intptr,intptr) [0x0002a] in <2e6290bc35d64d219e4c90c0d93f3982>:0
[0000000000000988:] EXCEPTION handling: System.NullReferenceException: Object reference not set to an instance of an object

"<unnamed thread>" tid=0000000000000988 this=0000000004a00130 , thread handle : 00000000020cb0f0, state : not waiting
  at _pmGTaW9nUOjSxRqkbPV3aw5SqFf._k7iWq7JuqXFAC1IDQvwFGfvHnEo._aF1LyXQMWFy66yf2JebAMK4DZJk () [0x0001c] in <f7ab7bd52c514bc88db790e398e47ae1>:0
  at _pmGTaW9nUOjSxRqkbPV3aw5SqFf._k7iWq7JuqXFAC1IDQvwFGfvHnEo._HD9qu3uWFYivE5nsGFNTEmTqQQe () [0x00000] in <f7ab7bd52c514bc88db790e398e47ae1>:0
  at <Module>..cctor () [0x0005f] in <f7ab7bd52c514bc88db790e398e47ae1>:0
  at (wrapper runtime-invoke) object.runtime_invoke_void (object,intptr,intptr,intptr) [0x0002a] in <2e6290bc35d64d219e4c90c0d93f3982>:0
Failed to run module constructor due to (null) assembly:Z:\home\aure\.local\share\Steam\steamapps\common\Proton - Experimental\files\share\wine\mono\wine-mono-10.2.0\lib\mono\4.5\mscorlib.dll type:TypeInitializationException member:(null)

No clue if this was what you where expecting as output when adding your command.
Let me know if you have any idea.
Can´t wait to even understand what it is telling me.

PS : I did not run this in the same computer as before but it is setup the same way.

Unfortunately, I think this is as far as I can go remotely.

I found this thread a while ago when I was having the same issue on FS2020 when using the DX12 renderer (the original issue with the vkCreateGraphicsPipeline).

I submitted an issue to the VKD3D-Proton github here and they made some fixes. I don't own 2024 but it might be the same issue if someone wants to check if the newest git build fixes this.

Hans also suggested using the launch option VKD3D_CONFIG=host_import_fallback %command% which doubles my FPS in 2020. But, again, unsure if that issue persists in 2024.

Hope this helps!

Base game MSFS 2024 runs fine with reasonable performance on my machine (rx 6600, r5 5600g, 32gb ddr4 3800, proton experimental, mesa-git 26.0.0-devel)

Don't own the Fenix A320, so can't comment on that.

Has anyone managed to get VR working on Linux? It works in every single other game I tried (all via Proton on different versions) but MSFS24 just refuses to work.

Has anyone managed to get VR working on Linux? It works in every single other game I tried (all via Proton on different versions) but MSFS24 just refuses to work.

Sadly not. It seems very picky on the implementation of VR, and I have also not been able to figure that out. I've been able to get to the point where it switches to VR, but instantly crashes before it renders anything. I have been unable to gather logs that show why it crashes, though.

I changed to 9060xt 16GB lately and revisited the sim after few months.

Game stopped working with experimental - 10.0.4 still works
Similar issue: https://github.com/GloriousEggroll/proton-ge-custom/issues/389

Crash happens at around "8% - Activating Packages" at startup

steam-2537590.log

I changed to 9060xt 16GB lately and revisited the sim after few months.

Game stopped working with experimental - 10.0.4 still works
...

Crash happens at around "8% - Activating Packages" at startup

This should hopefully be fixed in the just updated Proton Experimental bleeding edge branch (see https://github.com/ValveSoftware/Proton/wiki/Proton-Versions#proton-bleeding-edge on how to opt into bleeding edge).

Its working, thx!

• MSFS repeatedly queries rpcss.exe via QueryFullProcessImageNameW in the seconds before the crash, suggesting it is polling for the RPC service which has become unavailable.

My impression is that this is anti-cheat or telemetry. I see it repeatedly query all of those:
steam.exe, services.exe, winedevice.exe, svchost.exe, rpcss.exe, plugplay.exe, xalia.exe, winedevice.exe, FlightSimulator2024.exe (itself).

And when I deliberately split the log files into 2, so that I have one log file for startup (and before I trigger the 2nd window that triggers the crash) and then another logfile for the time during which it crashes, then I can sometimes time it so that the crash log does not contain any QueryFullProcessImageNameW lines, while the running log did.

BTW, I also see

015c:trace:seh:dispatch_exception code=80000004 (EXCEPTION_SINGLE_STEP) flags=0 addr=000000014359D830

about 20k times in my log, which looks like DRM to me.

but the RPC connection to Xbox Live/gaming services is dropped during or after window creation

When I tried to investigate the crash order, rpcss.exe outlived MSFS by ~924 ms.

To me, it also looks like all used RPC endpoints start with "nca" so I think it's all wine internal.

Crash when popping out instrument panels

Here's my analysis, based on creating 5 crash logs and then analyzing similarities:

After MSFS2024 creates the first swap chain, 1.47s to 1.52s later a new Linux PID is created. My logs didn't capture any interaction of it with Proton or vkd3d. But then after MSFS2024 creates the 2nd swap chain, the crash happens within 4.340s to 4.396s. It's not really a crash of MSFS2024 or of any Proton component, because I can see those components shut down normally. That's why rpcss.exe and services.exe outlive the main MSFS exe.

Using bpftrace, I tried to look at the signals that might trigger this shutdown. MSFS2024 will very regularly send itself SIGTRAP, which I believe is part of their DRM:

495184.167526: sig=5 target_pid=3914935 (sender=3914935 ACE: WinMain_Z )
495184.168797: sig=5 target_pid=3914935 (sender=3914935 ACE: WinMain_Z )
495184.399872: sig=5 target_pid=3914935 (sender=3914935 ACE: WinMain_Z )
495184.401111: sig=5 target_pid=3914935 (sender=3914935 ACE: WinMain_Z )
495185.171867: sig=5 target_pid=3914935 (sender=3914935 ACE: WinMain_Z )
495191.276962: sig=5 target_pid=3914935 (sender=3914935 ACE: WinMain_Z )
495191.288462: sig=5 target_pid=3914935 (sender=3914935 ACE: WinMain_Z )
495192.216674: sig=5 target_pid=3914935 (sender=3914935 ACE: WinMain_Z )
495192.253914: sig=5 target_pid=3914935 (sender=3914935 ACE: WinMain_Z )
495192.326554: sig=5 target_pid=3914935 (sender=3914935 ACE: WinMain_Z )

But shortly before the MSFS stops, something interesting happens:

495220.066744: KILL_USTACK: sig=11 target_pid=3915259 target_group=0 sender_pid=3914935
                             sender_tid=3915259 target_comm=ACE: WinMain_Z 

    7bbb928fb411 0x7bbb928fb411 ([unknown])    ← libnvidia-glcore.so.580.126.18 + 0xCFB411
    7bbb93f6d640 0x7bbb93f6d640 ([unknown])    ← anonymous rw (thread stack / false frame)

To me, it looks like 1.47s to 1.52s after the 1st swap chain is setup, the Linux-side NVIDIA driver will set up some sort of watchdog/presentation thread. And 4.340s to 4.396s after the 2nd swap chain is created, that thread crashes. That's the reason why this MSFS2024 exit leaves so few traces in WINEDBG logs: Because it's the Linux-side GPU driver that crashes. In line with that theory, I do see some rendering-related warnings in the logs:

495215.770556:02bc:trace:x11drv:x11drv_client_surface_update 0x100c6/0x55559aa53e60
495215.770601:02bc:trace:x11drv:client_surface_update_offscreen 0x100c6/0x55559aa53e60 offscreen 0
495215.770716:0270:warn:debugstr:OutputDebugStringA "[09-55-45][streamline][warn][tid:348][63s:303ms:739us]dxgiSwapchain.cpp:152[notifyAfterPresent] Streamline supports only one DXGISwapChain (currently it is 0x000000000765DBD0). Skipping some Present() hooks for DXGISwapChain 0x000000002D0E9C80\n"

debug: DXGI: For mode 3840x2160@144 found closest mode 3840x2160@144
495215.745:0158:015c:warn:vkd3d-proton:d3d12_command_queue_QueryInterface: {db6f6ddb-ac77-4e88-8253-819df9bbf140} not implemented, returning E_NOINTERFACE.
495215.745:0158:015c:info:vkd3d-proton:dxgi_vk_swap_chain_init: Creating swapchain (3736 x 2126), BufferCount = 2.
495215.746:0158:015c:info:vkd3d-proton:dxgi_vk_swap_chain_init_sync_objects: Ensure maximum latency of 3 frames with KHR_present_wait.
495215.755:0158:015c:info:vkd3d-proton:dxgi_vk_swap_chain_init_waiter_thread: Enabling present wait path for frame latency.
495215.755:0158:015c:info:vkd3d-proton:dxgi_vk_swap_chain_init_sleep_state: Timer interval is 1.0 ms.
warn:  DXGI: MakeWindowAssociation: Ignoring flags
495215.830:0158:02bc:info:vkd3d-proton:dxgi_vk_swap_chain_recreate_swapchain_in_present_task: Got 4 swapchain images.
495215.830:0158:02bc:warn:vkd3d-proton:dxgi_vk_swap_chain_record_render_pass: Application is presenting user index 0, but it has never been rendered to.

but, oddly enough, after those warnings the game will still continue to work for 4.340s to 4.396s (a bit more than 260 frames @ 60 FPS) and during this time it renders error-free, responds to mouse input and updates the rendered images accordingly.

=> My working hypothesis is that the vkd3d setup of the 2nd swap chain (inside Proton) accidentally triggers a watchdog in the GPU driver of the Linux host, even though everything appears to be working correctly.

🎉 In line with my crash cause hypothesis, PROTON_DISABLE_NVAPI=1 PROTON_HIDE_NVIDIA_GPU=1 DXVK_ASYNC=0 fixes both multi-camera and instrument pop-out rendering for my MSFS2024 Sim Update 5 with Proton 11.0.1-b2 commit 55ac62a.

Replying to https://github.com/ValveSoftware/Proton/issues/8255#issuecomment-4352410256

So it's the same solution as for crashing when starting the VR mode.