The Outer Worlds (578650)

[dennis1248]

Compatibility Report

• Name of the game with compatibility issues: The Outer Worlds
• Steam AppID of the game: 578650

System Information

• GPU: AMD RX 5700
• Driver/LLVM version: Mesa 20.2.1 / LLVM 10.0.0
• Kernel version: 5.8.16
• Link to full system information report as Gist: https://gist.github.com/dennis1248/289a286f6f5ce11751bdca2c1996a49c
• Proton version: 5.13-1 and 5.0-9

I confirm:

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

proton-5.0-9_steam-578650.log
proton-5.13-1_steam-578650.log

Symptoms

The game tries to launch but crashes shortly after a full-screen black game window appears.

Reproduction

1. Launch the game via the Steam client
2. Wait until the full-screen game window appears
3. Wait another 5 seconds before it crashes and puts you back on to the desktop

[kisak-valve]

Hello @dennis1248, looks like this game would benefit from support for Media Foundation maturing.

[dennis1248]

Hey @kisak-valve thank you for the response. Since you mentioned Media Foundation I tried out GloriousEggroll's Proton-5.9-GE-8-ST and the game launches fine using it, at least I can get in to the main menu just fine.It gets stuck and eventually crashes during the loading screen after starting a new game.

Edit: After trying a few more times it launches fine now and I managed to get in game using Proton-5.9-GE-8-ST. So the stability might not be great.

Edit 2: The game is unplayable with Proton-5.9-GE-8-ST, it suffers from various bugs such as crashes on load screens and some videos do not display and instead show a black screen or freeze half way through. So Proton GE is not a suitable workaround.

[McMarius11]

the game runs okay with Proton-5.9-GE-7-ST
i just had to turn down the graphics
i run the game in borderless windowed mode

OS: Manjaro Linux
KERNEL: 5.8.16-2-MANJARO
CPU: Intel Core i5-7600K @ 3.80GHz
GPU: NVIDIA GeForce GTX 1070
GPU DRIVER: NVIDIA 455.28
RAM: 24 GB

[zaggynl]

Proton-5.9-GE-2-MF worked for me.

My Steam sysinfo: https://gist.githubusercontent.com/zaggynl/aaab9a69b87425c7bb04badefd9a931e/raw/8989c3a30ced2f422842b01c61fdaa09c94fde7d/gistfile1.txt

Edit: sometimes the game starts and only shows cursor, other time it shows all studio logos and continues, what gives?

[leinardi]

For me the game freezes with Proton 5.9-GE-8-ST few seconds after the character creation, when you are on the pod on the way to the planet:

I tried to lower the graphic settings from Ultra to Very High but didn't make any change.
Anyone else with this issue? I'm on Nvidia 450.66 and 2080Ti.

[McMarius11]

For me the game freezes with Proton 5.9-GE-8-ST few seconds after the character creation, when you are on the pod on the way to the planet.
I tried to lower the graphic settings from Ultra to Very High but didn't make any change.
Anyone else with this issue? I'm on Nvidia 450.66 and 2080Ti.

Maybe a graphics driver update would help. Just give the game some time to create the shaders if it stutters at the beginning

[leinardi]

Thanks, after updating the drivers to 455.23.04 I was able to get past the point where it was freezing. Tomorrow I'll try to play a little more to see if I find other issues.

[hevanaa]

Black screen with Proton 5.13-1. Running fine with Proton 5.9-GE-8-ST. The first time I had to start it twice to fully start the game. One crash so far after playing maybe two hours. Nvidia driver 455.28-1 (from RPM Fusion repository) on Fedora 32. GTX 1070 and G-sync on Freesync display 34" 3440x1440.

[pradeepbyju5]

I am using 5.9 GE and experiencing perf issue whenever inventory menu is opened. The frame rate drop to 10 fps and does not recover, game needs to be restarted.

Elementary OS 5.1.7
i5 4690k
GTX 960

[zaggynl]

There's a memory leak or thread issue in there somewhere, after a couple hours I seem to run out of something, game freezes and becomes unresponsive.
If I try to alt tab and start htop in a terminal I get back: bash: fork: retry: Resource temporarily unavailable
Shortly after that the game crashes and exits.
dmesg shows:

[ 9596.667687] pipewire-media-[6325]: segfault at 50 ip 00007fad0851de70 sp 00007fff090bee68 error 4 in libpipewire-0.3.so.0.313.0[7fad084eb000+40000]
[ 9596.667693] Code: 66 2e 0f 1f 84 00 00 00 00 00 48 8b 47 20 48 85 c0 74 10 48 8b 80 58 01 00 00 48 85 c0 74 04 48 8b 40 10 c3 66 0f 1f 44 00 00 <48> 8b 47 50 48 89 56 10 48 89 4e 18 48 8b 10 48 89 46 08 48 89 16
[15840.839358] cgroup: fork rejected by pids controller in /user.slice/user-1000.slice/[email protected]
[15841.126704] cgroup: fork rejected by pids controller in /user.slice/user-1000.slice/session-2.scope

[GordianDziwis]

I am using 5.9 GE and experiencing perf issue whenever inventory menu is opened. The frame rate drop to 10 fps and does not recover, game needs to be restarted.

Elementary OS 5.1.7
i5 4690k
GTX 960

This happens with all the proton versions also with nvidia 450 and 455.

[sandsmark]

From the log it seems like the last output before things start going haywire is this:

warn:  DXGI: VK_FORMAT_D24_UNORM_S8_UINT -> VK_FORMAT_D32_SFLOAT_S8_UINT
warn:  D3D11DXGIDevice::QueryInterface: Unknown interface query
warn:  a7aa6116-9c8d-4bba-9083-b4d816b71b78

Which according to this seems like related to renderdoc detection, which was buggy in 2017:
baldurk/renderdoc#718 (comment)

edit: nvm, there's just a bunch of exceptions reporting "not implemented" before the actual crash. the actual crash seems like a segfault in mf.dll.

[folknor]

It might be worth noting that with https://github.com/GloriousEggroll/proton-ge-custom/releases/tag/5.21-GE-1 I've just done a 5 hour session with dozens of loading screens and no crashes. Perf is still the same as 5.9, but it's more than good enough for me.

[intelligentgaming]

I can confirm that with Proton-GE 5.21-GE-1 the game loads and plays correctly.

[ghost]

I'm unable to get the game to run with Proton-GE 5.21-GE-1. No window appears or anything, just crashes after no display output. I am able to get it to run with Proton 5.13-4 by disabling the startup videos in the Game.ini but no videos will play when starting a new game. It's like I'm outside the game map and roam around while the audio plays but no cutscene video. Tried doing <Workaround removed by moderator> with no luck.

OS: Manjaro Linux
KERNEL: 5.4.80-2-MANJARO
CPU: Ryzen 3 2200 @ 3.50GHz
GPU: NVIDIA GeForce GTX 1050 Ti
GPU DRIVER: NVIDIA 455.45
RAM: 32 GB

[kisak-valve]

Hello @dalzot, the workaround you mentioned is legally problematic and has been removed.

[kattjevfel]

Been playing for over 10 hours straight with Proton Experimental, only mild stuttering once in a while (compiling shaders I'd assume) and ofc no intro video other than a test pattern.

[kattjevfel]

Well it seems I jinxed it, went to launch it the morning after and now all I get is a black screen, after a minute or so the steam overlay shows up but then freeze again.

Just as I wrote this, I got an update for steam beta, only including Steam Input changes, nonetheless the game now launches again and works fine. I'll provide the log from when it wasn't working for science anyway: steam-578650.log

[intelligentgaming]

Not sure what has happened, but the game seems to have stopped working.

In summary it does not load using Proton 5.13-6, all builds of Proton-GE 5.21 and 6.1 load initially but any subsequent loads crash to a black screen and you have to kill the process.

I know it is not recommended, but I installed the Media Foundation patches with 5.13-6 just to rule that out causing the crash, but that still does not work.

Removing the compact folder does initially allow you to load the game but only with Proton-GE builds.

Can't understand why it would initially work, and then decide to not after subsequent loads?

Kubuntu 20.10
AMD Ryzen 5 3600
nVidia GTX 1080
Driver version 460.67
16GB DDR4
1 x TB SSD

Currently running the latest version of Steam beta client.

This is the only game in my library that has this problem.

[ktaherig]

I've tried it Proton-GE 5.21-GE-1 and Proton-6.8-GE-1, along with the 5 most recent official Proton releases, plus Proton experimental, and I can't get it to run.

Xubuntu 20.04
AMD Ryzen 9 4000 Series
nVidia GTX 2060
Driver version 460.73.01
16GB DDR4
1 x TB SSD

As @intelligentgaming has experienced, this is the one-and-only game in my library that has this problem.

[folknor]

It still runs perfectly for me using Proton-6.5-GE-1. No custom launch options or prefix tweaks. nvidia 1070.
EDIT: Please remember that with 6.5+ you need to delete old wine prefixes before launching. So if you try the game first on 5.21 or 6.1, then you need to delete the prefix before trying 6.6-6.8.

[intelligentgaming]

It still runs perfectly for me using Proton-6.5-GE-1. No custom launch options or prefix tweaks. nvidia 1070.
EDIT: Please remember that with 6.5+ you need to delete old wine prefixes before launching. So if you try the game first on 5.21 or 6.1, then you need to delete the prefix before trying 6.6-6.8.

Try closing the game, and reopening it, that is the problem I have.

The game initially launches and then does not on the second and further attempts.

[folknor]

Works fine, I just restarted it and played 5 minutes. Steam had been running in the meantime and no other games had been launched. I do need to forcefully kill the game after exit every time. It still shows as "Running" in Steam.

kill -9 $(ps -ef | grep -E -i '(wine|processid|\.exe)' | awk '{print $2}')
killall -9 pressure-vessel-adverb

https://www.youtube.com/watch?v=uxWJ1xvowMk

[intelligentgaming]

Works fine, I just restarted it and played 5 minutes. Steam had been running in the meantime and no other games had been launched. I do need to forcefully kill the game after exit every time. It still shows as "Running" in Steam.

kill -9 $(ps -ef | grep -E -i '(wine|processid|\.exe)' | awk '{print $2}')
killall -9 pressure-vessel-adverb

https://www.youtube.com/watch?v=uxWJ1xvowMk

I can confirm the game does now work with Proton 6.5-GE-1.

Played and recorded about 15 minutes of footage.

As you said the game does not close down properly put pressing the Stop button in Steam closes it.

[EmptyVisual]

The game gets stuck on a black screen after launching. Intro videos do not play and I'm unable to reach the main menu. This happens after every subsequent launch of the game in a single boot. Probably the most bizarre bug I've encountered.

If I somehow manage to get the game to run, sometimes on first try after reboot. If I then quit the game, Steam still thinks the game is Running. I have to then force quit the game by pressing 'Stop' in Steam. The commands mentioned above my comment also work.

From my short amount of research, this appears to have happened after the recent game DLC update, can anyone confirm?

Tested/ Affected Versions:
Proton Experimental
Proton 6.3-4
Proton 5.13-6
Additionally - Proton 6.8-GE-1, 6.5-GE-1 and 5.21-GE-1 were tested and the issue is also present with these builds.

Environment
Linux 5.12.4-156-tkg-bmq on Arch
Nvidia 465.27-6
Tried with FUTEX2 on and off
Tried with Steam Runtime and Native

[intelligentgaming]

Replying to #4309 (comment)

Delete the prefix compat folder and launch with 6.5-GE-1.

That fixed it for me.

[EmptyVisual]

Replying to #4309 (comment)

Delete the prefix compat folder and launch with 6.5-GE-1.

That fixed it for me.

Edit: Thank you.

Somehow it magically started working. I'm using 6.8-GE-1

[ktaherig]

Works fine, I just restarted it and played 5 minutes. Steam had been running in the meantime and no other games had been launched. I do need to forcefully kill the game after exit every time. It still shows as "Running" in Steam.

kill -9 $(ps -ef | grep -E -i '(wine|processid|\.exe)' | awk '{print $2}')
killall -9 pressure-vessel-adverb

https://www.youtube.com/watch?v=uxWJ1xvowMk

Can confirm that this does in fact work. Can finally play The Outer Worlds as normal, and it runs like a dream. Thank you all!

[smcv]

I do need to forcefully kill the game after exit every time. It still shows as "Running" in Steam.

If the game still has processes running, then it is intentional that pressure-vessel-adverb is also still running. It should exit when all the game processes have exited, and it should not exit until that has happened. Every direct or indirect child of pressure-vessel-adverb counts as part of the game (use pstree or ps f or similar to see).

You shouldn't need to use kill -9 or killall -9 (SIGKILL), and in fact it might be harmful, because it will prevent the container runtime from cleaning up all the processes that are part of the container (it can tell which processes are involved more easily than you can). The Stop button in Steam should be enough. If you prefer the command-line, sending SIGTERM (kill or killall with no special parameter, or kill -TERM, killall -TERM) should also be enough.

It would probably be useful if someone with this game can try this:

• run the game
• exit from the game
• Steam still thinks it's "Running"
• pstree -alp
• Copy the part of the tree that starts at bwrap or pressure-vessel-adverb
• Paste it here if small, or in a gist if large

pressure-vessel-adverb is designed to catch the TERM signal, and pass it on to all its child processes - so if you send SIGTERM to pressure-vessel-adverb (and nothing else), it should automatically clean up all the child processes within a few seconds. It will also send the KILL signal (the equivalent of kill -9) if there are child processes that still haven't exited after a few seconds. If that isn't working, please report a bug to https://github.com/ValveSoftware/steam-runtime and include:

• pstree -alp output for the part of the process tree that contains pressure-vessel-adverb, after exiting from the game
• pstree -alp output for the part of the process tree that contains pressure-vessel-adverb, after sending SIGTERM to pressure-vessel-adverb and waiting a few seconds

[Lightsockie]

Currently trying with Proton-6.9-GE-1. Current release Proton (6.3) won't even load.
With 6.9 GE, game loaded fine. I get passed character creation, but after a couple cut scenes to the next loading screen, it's a CTD

At launch I see this, I don't think it's relevant tho:

[0527/235557.420356:INFO:crash_reporting.cc(270)] Crash reporting enabled for process: renderer
Installing breakpad exception handler for appid(gameoverlayui)/version(1.0)
[0527/235557.518867:ERROR:frame_sink_video_capturer_impl.cc(206)] Invalid resolutions constraints: 0x0 must not be greater than 0x0; and also within media::limits.

Right before the crash (during that loading screen), eventfd: Too many open files gets dumped to the terminal

And then as it crashes:

Protocol error: process 0128: sendmsg: Bad file descriptor
pid 23610 != 23609, skipping destruction (fork without exec?)
Game removed: AppID 578650 "", ProcID 23695 

ulimit -n says 1024 (hard is 4096). Seems the file descriptor and esync are related, so tried PROTON_NO_ESYNC=1, but doesn't resolve the issue (or cause the eventfd warnings to stop).

Just bumped up my fd limits to 10000 100000. Gonna try that and see what happens. About 60min into my 120min refund window

---

And here I am on the other side of that limit change (and a reboot for good measure) and it seems to be working! So basically with the fd increased, this woulda worked out of the box with the GE release. Game seems to just be cooking my GPU, irregardless of graphics settings, and despite not seeming to stress resources. So that's rather strange, and I'll keep looking into why that may be

specs incase it's useful for someone down the line:

Kernel Version:  5.12.6-artix1-1
Driver:  AMD AMD Radeon RX 5700 XT (NAVI10, DRM 3.40.0, 5.12.6-artix1-1, LLVM 11.1.0)
Driver Version:  4.6 (Compatibility Profile) Mesa 21.1.1

[smcv]

ulimit -n says 1024 (hard is 4096). Seems the file descriptor and esync are related

Yes, "esync" uses relatively large numbers of eventfd file descriptors for synchronization (the "e" stands for eventfd).

Just bumped up my fd limits to 100000

As recommended in https://github.com/ValveSoftware/wine/blob/proton_6.3/README.esync you should leave the soft limit at 1024, and only increase the hard limit (a high soft limit system-wide can cause compatibility issues, here's a good explanation). Proton will use the higher hard limit to increase its own soft limit, but without breaking older programs that assume all their fds will be below 1024. Other large-fd-limit-aware programs like dbus-daemon do the same thing.

Recent Debian-derived distros (Debian, Ubuntu, Mint) set the soft limit to 1024 and the hard limit to about a million, and recent versions of systemd use 1024/512K, either of which should work well. If your OS doesn't raise these limits itself, it would make sense to follow one of those.

[Lightsockie]

here's a good explanation

Oooh, after getting over the initial impression hurdle that this is more garbage from Poettering, and tuning out the systemd wankery, it actually wasn't that bad a read and pretty informative to see how file descriptors have evolved. Totally makes sense, too. Thanks brother for the link, and genuinely appreciate the warning/save (despite my inability to miss a chance to spew vitriol at the blog's author)

the "e" stands for eventfd
As recommended in https://github.com/ValveSoftware/wine/blob/proton_6.3/README.esync

Very first line: "This is eventfd-based synchronization, or 'esync' for short"

Serves me right for skimming that last night when I came across it, haha. I've corrected my limits to 1024/1048576

[hevanaa]

Oooh, after getting over the initial impression hurdle that this is more garbage from Poettering, and tuning out the systemd wankery,

Hysterical nonsense keeping on bashing Poettering. I get it, all people can't like each other, but please keep it on a sane level. Thanks.

[folknor]

I just played an hour using https://github.com/GloriousEggroll/proton-ge-custom/releases/tag/6.10-GE-1 and it ran flawless. Previously I used 6.5-GE, and I've not tried any of the releases inbetween 6.5 and 6.10.

I did not use a fresh wine prefix. I ran the game on 6.5 since I bought it, and today changed to 6.10 using the same prefix.

[Optimus22Prime]

Game not running at all completely, tested with proton-6.3-5, proton-experimental, proton-6.12-GE-1. Here is the log from proton-experimental, maybe someone can explain to me whats wrong, because some ppls wrote before that game are working for them.
steam-578650.log

[folknor]

Game not running at all completely, tested with proton-6.3-5, proton-experimental, proton-6.12-GE-1. Here is the log from proton-experimental, maybe someone can explain to me whats wrong, because some ppls wrote before that game are working for them.

Did you read the comment just before yours? In any case https://github.com/GloriousEggroll/proton-ge-custom/releases/tag/6.12-GE-1 has been hotfixed.

HOTFIX: 7/6/2021:
-fix mfplat issue where some patches were not correctly applied, fixes The Outer Worlds, Power Rangers Battle for the Grid, and a few others that broke.

[Optimus22Prime]

10x man, I see hotfix now, but anyway I'm already got refund.

[kisak-valve]

Wrong colors on videos

Issue transferred from #5370.
@esistgut posted on 2021-12-01T15:00:11:

https://imgur.com/a/qCF7BZ2 this is the intro video of The Outer Worlds (the background should be black) but I get this kind of colors on a lot of videos. I can workaround this using <Workaround removed by moderator> but I would like to avoid it.

I'm on Arch, this is the proton log: https://gist.github.com/esistgut/791e7c7f83e9c2ce646d495ac01b666c

[howdev]

stuttering way too much as long as you move, stuttering starts in the Monarch map

[TC01]

The game currently appears to run more or less okay for me on Fedora with a Nvidia GPU, though there is a little stutter at times.

However, the Murder on Eridanos DLC appears to not be playable at the moment. Starting it seems to requires playing a video, during which I just see the placeholder video / color bars screen, and shortly afterwards the game becomes unresponsive-- there's an option to skip over the video but it doesn't do anything. I have to alt-tab back to steam and manually stop the game to exit.

My understanding is that this is due to the video playback issues described in #1464; from reading that discussion it sounds like the issue might get resolved automatically once the video is uploaded to the rendering queue? I thought I'd comment here though, since it's not totally clear to me how the upload queue works and since this specific playback issue seems somewhat serious.