After some debugging, I discovered the online stuttering seems to be caused by having CPU governor set to balanced rather than performance. Running the game with Feral Gamemode solves it.

I see you also have a 12th gen Intel. One thing you could try is manually setting the processor affinity for the game to only run on the P-cores (should be the first 12 cores) rather than the E-cores.

New Intel CPUs have "performance" and "efficiency" cores, with the 12600k having 6 P-cores (which have hyperthreading, so there are 12 logical P-cores in total) and 4 E-cores (which do not have hyperthreading). I have heard that some games can get confused by them, though I haven't noticed such issues myself. Denuvo also didn't work on 12th gen Intel when the processors first came out but that should be fixed by now.

Tried setting the core affinity to only the p-cores (0-11 in my case), and still the same performance during online matches unfortunately.

I have a friend that has the game on steam deck and for some reason when he is offline (no internet) the game constantly crashes while playing, will be good to check that out.

@J-Lowrance do you have the input lag reduction option turned on in the game settings? That also causes performance issues for me, regardless of CPU governor.

Same issue with me on a i5-13600K. Manually setting the CPU governor to performance doesn't fix it for me.

Same problem here with the same graphics card. The problem seems to happen regardless of in-game settings for the most part. In addition to the framerate loss, it also causes inputs to be dropped in online matches, which is a severe problem (tested by cross-referencing a local recording with an external input viewer, and the in-game replay system which records the inputs received.)

This was with the schedutil cpu governor set system wide, and running with gamemode which "should" have set the cpu governor for the game to performance. It seems like that didn't have a complete effect, though. Setting the system-wide default cpu governor to performance in the kernel made the game fps higher, which helped to some extent, however it did not solve the problem. It brings the fps up from the 30s to the high 40s according to steam's in-game fps viewer. Still not as high as it should be, but noticeably less obstructive. I'm not sure to what extent the inputs still drop.

In case its relevant, the specs for my system which suffers from this issue:

CPU: Intel i5-13600K
GPU: RTX 3070
Driver: Nvidia-driver-525
Kernel version: 6.2.6
Proton version: Proton Experimental & Proton GE 8-4
Distro: Pop!_OS 22.04

Okay, solved the issue on my machine. The problem was split lock protections (I had the same problem with God of War, although that was MUCH worse). Still not sure why this only shows up in online tho...

To diagnose the split lock issue, run the following command while the game is running and you should see a bunch of warning messages, like so:

$  sudo dmesg | grep split                 
[    0.000000] x86/split lock detection: #AC: crashing the kernel on kernel split_locks and warning on user-space split_locks
[41786.661399] x86/split lock detection: #AC: StreetFighter6./47861 took a split_lock trap at address: 0x16886c89a
[41788.750104] x86/split lock detection: #AC: StreetFighter6./47930 took a split_lock trap at address: 0x15f853006
[41789.268254] x86/split lock detection: #AC: StreetFighter6./47947 took a split_lock trap at address: 0x1684dde0a

The fix is simple, if you're running a Kernel >= 6.2, you can disable the split_lock protection during runtime with the following command. You can do this while the game is running, and you should see the problem immediately dissapear:

$ sudo sysctl kernel.split_lock_mitigate=0

You can also change a boot parameter to permanently disable split lock protections (see linked github issue below).

Heres a github issue and an article where I initially learned about this fix for god of war.

Thanks @RTbecard, that seems to fix it on my end.

@RTbecard you saved me from so much frustration, you have no idea. I've now a solid 60FPS online, you get a kiss from Chun Li for that

My theory why it's only online: rollback netcode ask for much more from the CPU because it has to calculate the potential moves in advance, so higher CPU consumption (at least it's what I read, and it makes sense to me). Maybe it's when they abuse this split stuff...

A few minutes into World Tour starts to cause vertices to explode until the game crashes. Using Nvidia GTX 1660 Ti with driver version 525.116.04. Seems like many other Nvidia users are having the same issue on ProtonDB.
I've tried both Proton experimental and GE 8.4

Looks like this is also affecting other RE Engine games like Monster Hunter Rise, albeit to a much lesser extent

For me, when entering busy areas in the game (particularly in Metro City, another large spoiler-y World Tour location, and in populated Battle Hub lobbies), the game will often hit 100% GPU usage and the FPS will tank, after which there is a chance that the game will freeze. The frame rate also tends to tank when using the "Status" app on the in-game phone. I haven't noticed this happen in the smaller World Tour locations, in less-populated Battle Hub lobbies, or in Fighting Grounds.

I can confirm that I've also had vertex glitches, which have almost always coincided with the frame rate drops mentioned before, but for me this particular bug was rare during my play-through of World Tour.

This is with an RTX 2070 Super, running the 530.41.03 driver on Fedora 38. I have to use Proton Experimental in order to get past the opening splash screens.

The game doesn't launch on an Nvidia2060 mobile. It gets stuck after the "Street Figther 6" splash screen and won't launch regardless of which proton version is used (tried Experimental, Hotfix and 8.4)

Also, i'm now getting a persistent error at launch:

The game doesn't launch on an Nvidia2060 mobile. It gets stuck after the "Street Figther 6" splash screen and won't launch regardless of which proton version is used (tried Experimental, Hotfix and 8.4)

This should be fixed now on Proton Experimental with the [bleedige-edge] beta branch.

Please give it a try.

Also, i'm now getting a persistent error at launch:

The message is from Denuvo DRM (Incorporates 3rd-party DRM: Denuvo Anti-tamper
5 different PC within a day machine activation limit).

After 24 hours you should be able to launch the game again.

Replying to https://github.com/ValveSoftware/Proton/issues/6817#issuecomment-1611903327

You where right about the message coming from denuvo. I was able to launch the game again (finally) using Proton Experimental bleeding edge, but it remains stuck on this screen (same as yesterday), I left it for a while but it won't go pas this screen:

I have updated Proton Experimental before attempting to launch. Here is my driver information:

Replying to https://github.com/ValveSoftware/Proton/issues/6817#issuecomment-1612778095

When the game gets stuck: run sudo dmesg | tail and post the output. I wonder if you are getting a Xid error.

The fix in Proton Experimental [bleedige-edge] solved it on 3070 and 3080 at least...

Here is the output of sudo dmesg | tail

[11290.053551] wlp0s20f3: Limiting TX power to 30 (30 - 0) dBm as advertised by c8:99:b2:dc:d7:88
[11323.575633] wlp0s20f3: disconnect from AP c8:99:b2:dc:d7:88 for new auth to c8:99:b2:dc:d7:87
[11323.695305] wlp0s20f3: authenticate with c8:99:b2:dc:d7:87
[11323.695336] wlp0s20f3: 80 MHz not supported, disabling VHT
[11323.700846] wlp0s20f3: send auth to c8:99:b2:dc:d7:87 (try 1/3)
[11323.758219] wlp0s20f3: authenticated
[11323.761474] wlp0s20f3: associate with c8:99:b2:dc:d7:87 (try 1/3)
[11323.782355] wlp0s20f3: RX ReassocResp from c8:99:b2:dc:d7:87 (capab=0x1c31 status=0 aid=10)
[11323.793282] wlp0s20f3: associated
[11323.924169] wlp0s20f3: Limiting TX power to 30 (30 - 0) dBm as advertised by c8:99:b2:dc:d7:87

Same output in better organized form:

Also, judging by the Protondb page (https://www.protondb.com/app/1364780) there are numerous reports of the game freezing and graphics degrading with time even in single player mode. Perhaps the issue that affects me, is the one that caused these freezes in the first place.

System Information

• GPU: AMD Radeon RX 6650 XT
• Driver/LLVM version: Mesa 23.1.3
• Kernel version: 6.4.1-zen1-1-zen
• https://gist.github.com/SuperBunChan/da544ec3fb55cad65bd06a4c1660ab58
• Proton version: 1687882519 experimental-8.0-20230627

I confirm:

• [✓] that I have checked whether there are updates for my system available.

steam-1364780.log

Symptoms

Extreme lag/fps drop whenever 3D models are on the screen, 2D menus work fine.

Reproduction

Easy way to reproduce is to go into training mode.

System Information

• GPU: NVIDIA GeForce GTX 1080 Ti
• Driver/LLVM version: NVIDIA 535.54.03
• Kernel version: 6.3.9-arch1-1
• Link to full system information report
• Proton version: 1687882519 experimental-8.0-20230627

I confirm:

• [✅] that I have checked whether there are updates for my system available.

sf6-shader-lag-steam-1364780.log

Symptoms

Temporary game slowdown (drop to 0 framerate) occurs the first time a special move / animation with special effects is performed in game. The main game menu is also initially slow during its initial animation.

Reproduction

1. Launch the game and enable shader cache warming in the options menu.
2. Restart game and wait for shader cache warming to complete.
3. Start a practice match as Ryu, try to throw a Hadouken. It works fine (no lag / stuttering).
4. Exit the game and restart it.
5. Start a practice match as Ryu, try to throw a Hadouken. At this point, there is significant lag and a frame rate drop to 0 during the first animation of the special move. Subsequent animations of the same special move work smoothly with no lag / stuttering.

Workaround

Using the launch command VKD3D_CONFIG=pipeline_library_app_cache %command% worked around this issue for me. After launching the game and allowing the shader warming to complete, the performance was consistently good even after relaunching the game multiple times.

Replying to https://github.com/ValveSoftware/Proton/issues/6817#issuecomment-1595495268

$ sudo sysctl kernel.split_lock_mitigate=0
sysctl: cannot stat /proc/sys/kernel/split_lock_mitigate: No such file or directory

This unfortunately doesn't run on my Ubuntu Machine. The game runs 60FPS perfectly everywhere, except when an online match starts, even the win/lose animation and the intro anitmations are at 60FPS, but as soon as gameplay is starts in an online match, it locks to ~30 FPS.

I ended up having to add split_lock_detect=off to my GRUB config in order to resolve the split lock issue. Here are instructions for anyone else having this issue.

1. Open the GRUB configuration file. You'll need to use sudo to get the necessary permissions:

   sudo nano /etc/default/grub
   

2. Find the line that begins with GRUB_CMDLINE_LINUX_DEFAULT. This line contains the boot parameters that are applied each time your system boots.

   It might look something like this:

   GRUB_CMDLINE_LINUX_DEFAULT="quiet splash"
   

3. Add split_lock_detect=off to the end of the parameters, inside the quotes:

   GRUB_CMDLINE_LINUX_DEFAULT="quiet splash split_lock_detect=off"
   

4. Save your changes and exit the text editor. If you're using nano, you can do this by pressing Ctrl+O followed by Enter to save, and then Ctrl+X to exit.

5. After saving the changes to the GRUB configuration, you must update GRUB to apply these changes:

   sudo update-grub
   

6. Reboot your system for the changes to take effect.

   sudo reboot
   

After your system reboots, the new kernel parameter should be in effect. After boot steam and try an online match. Should run at 60FPS now.

Replying to [#6817 (comment)](https://github.com/ValveSoftware/Proton/issues/6817#issuecomment-1595495268)

$ sudo sysctl kernel.split_lock_mitigate=0
sysctl: cannot stat /proc/sys/kernel/split_lock_mitigate: No such file or directory

This unfortunately doesn't run on my Ubuntu Machine. The game runs 60FPS perfectly everywhere, except when an online match starts, even the win/lose animation and the intro anitmations are at 60FPS, but as soon as gameplay is starts in an online match, it locks to ~30 FPS.

This sysctl option is added in kernel 6.2, so if you are on an LTS distro you probably don't have it yet.

Game is freezing for me when loading new matches (Not always). All I see on the logs:

jul 24 12:54:16 artorias kernel: umip: StreetFighter6.[3140] ip:165d43681 sp:571f7238: SGDT instruction cannot be used by applications.
jul 24 12:54:16 artorias kernel: umip: StreetFighter6.[3140] ip:165d43681 sp:571f7238: For now, expensive software emulation returns the result.
jul 24 12:54:16 artorias kernel: umip: StreetFighter6.[3140] ip:1606fab6c sp:571f7228: SGDT instruction cannot be used by applications.
jul 24 12:54:16 artorias kernel: umip: StreetFighter6.[3140] ip:1606fab6c sp:571f7228: For now, expensive software emulation returns the result.
jul 24 12:54:16 artorias kernel: umip: StreetFighter6.[3140] ip:162c1b4fa sp:571ff718: SGDT instruction cannot be used by applications.

Having the same issues as @alosarjos

6900 XT
Output when crashed:

[24413.657790] amdgpu 0000:2f:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:3 pasid:32769, for process StreetFighter6. pid 1172524 
thread vkd3d_queue pid 1172562)
[24413.657796] amdgpu 0000:2f:00.0: amdgpu:   in page starting at address 0x000080009f54c000 from client 0x1b (UTCL2)
[24413.657799] amdgpu 0000:2f:00.0: amdgpu: GCVM_L2_PROTECTION_FAULT_STATUS:0x00301431
[24413.657800] amdgpu 0000:2f:00.0: amdgpu:      Faulty UTCL2 client ID: SQC (data) (0xa)
[24413.657801] amdgpu 0000:2f:00.0: amdgpu:      MORE_FAULTS: 0x1
[24413.657802] amdgpu 0000:2f:00.0: amdgpu:      WALKER_ERROR: 0x0
[24413.657802] amdgpu 0000:2f:00.0: amdgpu:      PERMISSION_FAULTS: 0x3
[24413.657803] amdgpu 0000:2f:00.0: amdgpu:      MAPPING_ERROR: 0x0
[24413.657804] amdgpu 0000:2f:00.0: amdgpu:      RW: 0x0
[24413.657808] amdgpu 0000:2f:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:3 pasid:32769, for process StreetFighter6. pid 1172524 
thread vkd3d_queue pid 1172562)

I have been logged out and after logging in and trying to launch the game:

[24474.911149] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!

@kisak-valve

• gist
• Proton log file steam-1364780.log

Maybe this extra info will help with the issue - I've tried playing with newest Mesa RC drivers to resolve this issue but after having no luck I've restored them to previous version (22.3.5).

Same match-load issues on Steam Deck. It seems to have coincided with the latest Street Fighter 6 update.

I experience the crash most often when loading into a ranked multiplayer match. The pre-match faces appear, then the screen fades to black, the match background music plays partially and then stalls (the introductory musical phrase loops), and the multiplayer match never loads. I've also experienced the crash in causal matchmaking, and when switching combo trial scenarios.

• gist
• Proton log file steam-1364780.log

Thanks for reviewing!

To everyone having the black screen issue, there is a method to recover control of the system,
https://wiki.archlinux.org/title/Keyboard_shortcuts
using Alt+SysRq+r I recover my keyboard and i can go to another tty an kill wine streetfighter and gdm

I am also experiencing this issue where the game crashes when loading to or from matches.

Computer Information:
Manufacturer: ASUSTeK COMPUTER INC.
Model: PRIME B550-PLUS AC-HES
Form Factor: Desktop
No Touch Input Detected
Processor Information:
CPU Vendor: AuthenticAMD
CPU Brand: AMD Ryzen 5 5600G with Radeon Graphics
CPU Family: 0x19
CPU Model: 0x50
CPU Stepping: 0x0
CPU Type: 0x0
Speed: 4463 MHz
12 logical processors
6 physical processors
Hyper-threading: Supported
FCMOV: Supported
SSE2: Supported
SSE3: Supported
SSSE3: Supported
SSE4a: Supported
SSE41: Supported
SSE42: Supported
AES: Supported
AVX: Supported
AVX2: Supported
AVX512F: Unsupported
AVX512PF: Unsupported
AVX512ER: Unsupported
AVX512CD: Unsupported
AVX512VNNI: Unsupported
SHA: Supported
CMPXCHG16B: Supported
LAHF/SAHF: Supported
PrefetchW: Unsupported
Operating System Version:
"Arch Linux" (64 bit)
Kernel Name: Linux
Kernel Version: 6.4.6-arch1-1
X Server Vendor: The X.Org Foundation
X Server Release: 12101008
X Window Manager: KWin
Steam Runtime Version: steam-runtime_0.20230509.49499
Video Card:
Driver: AMD AMD Radeon RX 6900 XT (navi21, LLVM 15.0.7, DRM 3.52, 6.4.6-arch1-1)
Driver Version: 4.6 (Compatibility Profile) Mesa 23.1.4
OpenGL Version: 4.6
Desktop Color Depth: 24 bits per pixel
Monitor Refresh Rate: 59 Hz
VendorID: 0x1002
DeviceID: 0x73bf
Revision Not Detected
Number of Monitors: 2
Number of Logical Video Cards: 1
Primary Display Resolution: 2560 x 1440
Desktop Resolution: 5120 x 1440
Primary Display Size: 27.44" x 15.43" (31.46" diag), 69.7cm x 39.2cm (79.9cm diag)
Primary VRAM: 16384 MB
Sound card:
Audio device: ATI R6xx HDMI
Memory:
RAM: 31889 Mb
VR Hardware:
VR Headset: None detected
Miscellaneous:
UI Language: English
LANG: en_US.UTF-8
Total Hard Disk Space Available: 455768 MB
Largest Free Hard Disk Block: 32816 MB
Storage:
Number of SSDs: 3
SSD sizes: 6001G,2000G,500G
Number of HDDs: 0
Number of removable drives: 0

steam-1364780.log

kisak-valve

Unsure if it matters, but when my game crashes to black screen it is in match making or training and happens in the middle of a fight, not loading to or from.

@kisak-valve Apologies if I'm supposed to fill out a formal report. I have the same bug described recently: crashing when loading/exiting a match. I was able to reproduce by repeatedly starting a session in training mode. Gist of my Steam system information here. Log file attached. I was tailing the log file while trying to get the bug to happen, and my system froze between lines 23298 and 23307 of the log, somewhere in there. Here's an example of what is being repeated on those lines:
820.741:0140:018c:err:vkd3d-proton:vkd3d_wait_for_gpu_timeline_semaphore: Failed to wait for Vulkan timeline semaphore, vr -4.
Immediately before those lines and also unique in the log:
811.618:0140:01fc:err:sync:RtlpWaitForCriticalSection section 0000020004380888 (null) wait timed out in thread 01fc, blocked by 01f8, retrying (60 sec)
steam-1364780.log

@kisak-valve Just an note, seems like this is probably an issue for all steam deck users after the rashid update a few days ago. but for me it seems only battle grounds is crashing with loading screens. World tour is fine and I have not noticed any issues.

Everything seemed fine for me, and after playing a random amount of ranked or casual matches, exiting a match with someone gives me black screen, which subsequently crashes and also completely hoses my x11, and crashes x too.

Easily repro'd just by playing the game online, give it 3-4 matches and it will crash.

Gentoo, with kernel 6.4.7, xorg
radeon 6900xt, 7950x, proton 8.0-3

Game was flawless until Rashid update on Monday, I have like 130 hours in so something has changed in this update.

Can confirm, game worked perfectly fine until Rashid patch and now crashes after a couple of online matches after the character loading screen mostly.

Similar issue. Can consistently replicate by going into Combo Trial and manually switching the trial, crashes every time.
steam-1364780.log
gist

On steam deck and linux (ubuntu) I get black screen after some gameplay.
The number of games may vary:
Yesterday I played arcade mode and it crashed:

• after 5 matches
  I rebooted
• after 8 matches
  I rebooted
• after I completed the 12 arcade games and entered the practice mode
  I rebooted

Looks like it does not happen during the game, but it happens randomly during the loading screens

I did multiple tests with:
Tested with proton 7, proton 8, Proton experimental, proton GE
Also kernel.split_lock_mitigate=0 does not make any difference

• Hardware 1:
  steam deck

• Hardware 2:
  Ryzen 5 7600x
  RX6800xt
  x760E motherboard

error from my dmesg in ubuntu:

[drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!

[  558.150065] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:2 pasid:32770, for process StreetFighter6. pid 6539 thread vkd3d_queue pid 6589)
[  558.150071] amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x00008000b053e000 from client 0x1b (UTCL2)
[  558.150072] amdgpu 0000:03:00.0: amdgpu: GCVM_L2_PROTECTION_FAULT_STATUS:0x00201431
[  558.150073] amdgpu 0000:03:00.0: amdgpu: 	 Faulty UTCL2 client ID: SQC (data) (0xa)
[  558.150074] amdgpu 0000:03:00.0: amdgpu: 	 MORE_FAULTS: 0x1
[  558.150074] amdgpu 0000:03:00.0: amdgpu: 	 WALKER_ERROR: 0x0
[  558.150075] amdgpu 0000:03:00.0: amdgpu: 	 PERMISSION_FAULTS: 0x3
[  558.150076] amdgpu 0000:03:00.0: amdgpu: 	 MAPPING_ERROR: 0x0
[  558.150076] amdgpu 0000:03:00.0: amdgpu: 	 RW: 0x0
[  558.150081] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:2 pasid:32770, for process StreetFighter6. pid 6539 thread vkd3d_queue pid 6589)
[  558.150082] amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x00008000b053e000 from client 0x1b (UTCL2)
[  558.150083] amdgpu 0000:03:00.0: amdgpu: GCVM_L2_PROTECTION_FAULT_STATUS:0x00000000
[  558.150084] amdgpu 0000:03:00.0: amdgpu: 	 Faulty UTCL2 client ID: CB/DB (0x0)
[  558.150084] amdgpu 0000:03:00.0: amdgpu: 	 MORE_FAULTS: 0x0
[  558.150085] amdgpu 0000:03:00.0: amdgpu: 	 WALKER_ERROR: 0x0
[  558.150085] amdgpu 0000:03:00.0: amdgpu: 	 PERMISSION_FAULTS: 0x0
[  558.150086] amdgpu 0000:03:00.0: amdgpu: 	 MAPPING_ERROR: 0x0
[  558.150086] amdgpu 0000:03:00.0: amdgpu: 	 RW: 0x0
[  558.150091] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:2 pasid:32770, for process StreetFighter6. pid 6539 thread vkd3d_queue pid 6589)
[  558.150092] amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x00008000b053e000 from client 0x1b (UTCL2)
[  558.150093] amdgpu 0000:03:00.0: amdgpu: GCVM_L2_PROTECTION_FAULT_STATUS:0x00000000
[  558.150093] amdgpu 0000:03:00.0: amdgpu: 	 Faulty UTCL2 client ID: CB/DB (0x0)
[  558.150094] amdgpu 0000:03:00.0: amdgpu: 	 MORE_FAULTS: 0x0
[  558.150094] amdgpu 0000:03:00.0: amdgpu: 	 WALKER_ERROR: 0x0
[  558.150095] amdgpu 0000:03:00.0: amdgpu: 	 PERMISSION_FAULTS: 0x0
[  558.150095] amdgpu 0000:03:00.0: amdgpu: 	 MAPPING_ERROR: 0x0
[  558.150095] amdgpu 0000:03:00.0: amdgpu: 	 RW: 0x0
[  558.150101] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:2 pasid:32770, for process StreetFighter6. pid 6539 thread vkd3d_queue pid 6589)
[  558.150101] amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x00008000b053e000 from client 0x1b (UTCL2)
[  558.150102] amdgpu 0000:03:00.0: amdgpu: GCVM_L2_PROTECTION_FAULT_STATUS:0x00000000
[  558.150103] amdgpu 0000:03:00.0: amdgpu: 	 Faulty UTCL2 client ID: CB/DB (0x0)
[  558.150103] amdgpu 0000:03:00.0: amdgpu: 	 MORE_FAULTS: 0x0
[  558.150103] amdgpu 0000:03:00.0: amdgpu: 	 WALKER_ERROR: 0x0
[  558.150104] amdgpu 0000:03:00.0: amdgpu: 	 PERMISSION_FAULTS: 0x0
[  558.150104] amdgpu 0000:03:00.0: amdgpu: 	 MAPPING_ERROR: 0x0
[  558.150105] amdgpu 0000:03:00.0: amdgpu: 	 RW: 0x0
[  558.150110] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:2 pasid:32770, for process StreetFighter6. pid 6539 thread vkd3d_queue pid 6589)
[  558.150111] amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x00008000b053e000 from client 0x1b (UTCL2)
[  558.150112] amdgpu 0000:03:00.0: amdgpu: GCVM_L2_PROTECTION_FAULT_STATUS:0x00000000
[  558.150112] amdgpu 0000:03:00.0: amdgpu: 	 Faulty UTCL2 client ID: CB/DB (0x0)
[  558.150112] amdgpu 0000:03:00.0: amdgpu: 	 MORE_FAULTS: 0x0
[  558.150113] amdgpu 0000:03:00.0: amdgpu: 	 WALKER_ERROR: 0x0
[  558.150113] amdgpu 0000:03:00.0: amdgpu: 	 PERMISSION_FAULTS: 0x0
[  558.150114] amdgpu 0000:03:00.0: amdgpu: 	 MAPPING_ERROR: 0x0
[  558.150114] amdgpu 0000:03:00.0: amdgpu: 	 RW: 0x0
[  558.150120] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:2 pasid:32770, for process StreetFighter6. pid 6539 thread vkd3d_queue pid 6589)
[  558.150120] amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x00008000b053e000 from client 0x1b (UTCL2)
[  558.150121] amdgpu 0000:03:00.0: amdgpu: GCVM_L2_PROTECTION_FAULT_STATUS:0x00000000
[  558.150122] amdgpu 0000:03:00.0: amdgpu: 	 Faulty UTCL2 client ID: CB/DB (0x0)
[  558.150122] amdgpu 0000:03:00.0: amdgpu: 	 MORE_FAULTS: 0x0
[  558.150122] amdgpu 0000:03:00.0: amdgpu: 	 WALKER_ERROR: 0x0
[  558.150123] amdgpu 0000:03:00.0: amdgpu: 	 PERMISSION_FAULTS: 0x0
[  558.150123] amdgpu 0000:03:00.0: amdgpu: 	 MAPPING_ERROR: 0x0
[  558.150124] amdgpu 0000:03:00.0: amdgpu: 	 RW: 0x0
[  558.150129] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:2 pasid:32770, for process StreetFighter6. pid 6539 thread vkd3d_queue pid 6589)
[  558.150130] amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x00008000b053e000 from client 0x1b (UTCL2)
[  558.150131] amdgpu 0000:03:00.0: amdgpu: GCVM_L2_PROTECTION_FAULT_STATUS:0x00000000
[  558.150131] amdgpu 0000:03:00.0: amdgpu: 	 Faulty UTCL2 client ID: CB/DB (0x0)
[  558.150132] amdgpu 0000:03:00.0: amdgpu: 	 MORE_FAULTS: 0x0
[  558.150132] amdgpu 0000:03:00.0: amdgpu: 	 WALKER_ERROR: 0x0
[  558.150132] amdgpu 0000:03:00.0: amdgpu: 	 PERMISSION_FAULTS: 0x0
[  558.150133] amdgpu 0000:03:00.0: amdgpu: 	 MAPPING_ERROR: 0x0
[  558.150133] amdgpu 0000:03:00.0: amdgpu: 	 RW: 0x0
[  558.150139] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:2 pasid:32770, for process StreetFighter6. pid 6539 thread vkd3d_queue pid 6589)
[  558.150140] amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x00008000b053e000 from client 0x1b (UTCL2)
[  558.150140] amdgpu 0000:03:00.0: amdgpu: GCVM_L2_PROTECTION_FAULT_STATUS:0x00000000
[  558.150141] amdgpu 0000:03:00.0: amdgpu: 	 Faulty UTCL2 client ID: CB/DB (0x0)
[  558.150141] amdgpu 0000:03:00.0: amdgpu: 	 MORE_FAULTS: 0x0
[  558.150141] amdgpu 0000:03:00.0: amdgpu: 	 WALKER_ERROR: 0x0
[  558.150142] amdgpu 0000:03:00.0: amdgpu: 	 PERMISSION_FAULTS: 0x0
[  558.150142] amdgpu 0000:03:00.0: amdgpu: 	 MAPPING_ERROR: 0x0
[  558.150143] amdgpu 0000:03:00.0: amdgpu: 	 RW: 0x0
[  587.597621] [drm:amdgpu_job_timedout [amdgpu]] *ERROR* ring gfx_0.0.0 timeout, signaled seq=107000, emitted seq=107002
[  587.597873] [drm:amdgpu_job_timedout [amdgpu]] *ERROR* Process information: process gnome-shell pid 2440 thread gnome-shel:cs0 pid 2453

Same issue, the game freezes the wholes system, during the loading screen after an online match.

I'm experiencing the same issue with Ryzen 7600x & RX 6800. However, I'd like to add that this black screen/loading crash doesn't only happen when playing online matches. It can happen before a versus screen, after the versus screen, or after finishing a match. Meaning even arcade mode and training mode can crash.

Even switching characters in training mode can trigger it, so it is directly tied to loading and not to the versus screen.

It also happened when I was in trial mode and I decided to switch to a different combo from the menu.
More specifically:
I was using Guile, I was doing Intermediate 7, opened the menu, selected intermediate 5. Black screen. Crash

Every time there is a loading screen (black with Ryu spinning a the bottom right corner) there is a risk of game getting stuck and crashing.

To whom this may concern, maybe @kisak-valve, is this being looked into? Although this crash occurs sporadically it stops me from being able to play ranked mode.

This is due to the fact that SF6 implements a leave mechanic where once you have a history of leaves (regardless of if its a crash) your account becomes quarantined (yellow carded) and you only get matched up with other yellow cards which severely limits how many people you can queue with. I'm at the point where if it crashes even once, the next 20 games i spend in quarantine which amounts to being there all the time.

I believe this is a major compatibility issue which stops me from playing the most important mode of the game and at the very least the steam store needs to reflect this by specifying this game is NOT compatible with the steam deck.

I'm sure this isn't an easy solve and I appreciate any time spent trying to fix this.

Thank you

Operating System Version:
"Arch Linux" (64 bit)
Kernel Name: Linux
Kernel Version: 6.4.8-arch1-1
X Server Vendor: The X.Org Foundation
X Server Release: 12101008
X Window Manager: KWin
Steam Runtime Version: steam-runtime_0.20230606.51628
Driver: AMD AMD Radeon RX 7900 XTX (gfx1100, LLVM 15.0.7, DRM 3.52, 6.4.8-arch1-1)
Driver Version: 4.6 (Compatibility Profile) Mesa 23.3.0-devel (git-746b90839f)
OpenGL Version: 4.6
Proton Version: Proton Experimental

dmesg | grep amdgpu

Pastebin

dmesg | grep StreetFighter

[ 5874.587389] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:6 pasid:32784, for process StreetFighter6. pid 9229 thread vkd3d_queue pid 9371)
[ 5874.587412] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:6 pasid:32784, for process StreetFighter6. pid 9229 thread vkd3d_queue pid 9371)
[ 5874.587429] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:6 pasid:32784, for process StreetFighter6. pid 9229 thread vkd3d_queue pid 9371)
[ 5874.587446] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:6 pasid:32784, for process StreetFighter6. pid 9229 thread vkd3d_queue pid 9371)
[ 5874.587463] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:6 pasid:32784, for process StreetFighter6. pid 9229 thread vkd3d_queue pid 9371)
[ 5874.587479] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:6 pasid:32784, for process StreetFighter6. pid 9229 thread vkd3d_queue pid 9371)
[ 5874.587496] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:6 pasid:32784, for process StreetFighter6. pid 9229 thread vkd3d_queue pid 9371)
[ 5874.587512] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:6 pasid:32784, for process StreetFighter6. pid 9229 thread vkd3d_queue pid 9371)
[ 5874.587528] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:6 pasid:32784, for process StreetFighter6. pid 9229 thread vkd3d_queue pid 9371)
[ 5874.587544] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:6 pasid:32784, for process StreetFighter6. pid 9229 thread vkd3d_queue pid 9371)

I am also getting random black screen crashes in online matches, requiring either a PID kill or system restart to fully regain system control. I did not realize I needed to set up the Proton logging in advance, so I do not have that to match this, but could provide it in the future.

I'm back from vacation now and actively looking into this.

Chiming in that I also have locks after the Rashid update.
Fedora 38
Kernel: 6.4.7-200.fc38.x86_64
Proton version: I've tried GE, Proton Experimental, Proton 8, etc with all the same result

If I play a custom room it seems to happen very rarely. Once every 20 matches or so I have a lock. But in ranked it happens pretty often (probably within 1-3 matches of starting a session) requiring a full reboot. Unlike others I have not been quarantined or yellow carded. For me it seems to happen after the "Quit" prompt and returning to Practice Mode (where I started matchmaking from).

Usually I get a black screen but occasionally I see the little whirlwind kick icon spin for a moment and then freeze.

Root caused to a game bug. This PR should fix the hang: https://github.com/HansKristian-Work/vkd3d-proton/pull/1647. Should hit bleeding-edge very soon.

Root caused to a game bug. This PR should fix the hang: HansKristian-Work/vkd3d-proton#1647. Should hit bleeding-edge very soon.

Does Valve report this to Capcom directly or something (To fix it on their end)?

@HansKristian-Work Thank you so much for this!

Thanks!

I have been playing all day on Experimental with no lockups during ranked matches and none happening during custom room matches. I tried using combo trials to recreate what some of the people above mentioned and no lockups there, either.

Thanks!

I have been playing all day on Experimental with no lockups during ranked matches and none happening during custom room matches. I tried using combo trials to recreate what some of the people above mentioned and no lockups there, either.

I played for ~2 hours last night after a Proton and SF6 (TMNT cosmetics) update and the crashes also seem to have stopped. I am also running Proton experimental.

Crash happened to me last night in ranked, I'll play again tonight

edit: Using proton experimental.

Crash happened to me last night in ranked, I'll play again tonight

edit: Using proton experimental.

Proton Experimental will still have crash, fix is in Experimental's Bleeding Edge branch. Right click to Experimental in library-properties-set it to bleeding edge branch.

Crash happened to me last night in ranked, I'll play again tonight
edit: Using proton experimental.

Proton Experimental will still have crash, fix is in Experimental's Bleeding Edge branch. Right click to Experimental in library-properties-set it to bleeding edge branch.

How long does it take for changes to be pushed to the main Experimental branch? Where can I find out when that's happened?

Edit: After some more play testing, the disconnects seem unrelated to the patch and have basically gone away.

Good news, no more black screens! However, with the latest bleeding-edge experimental version I am getting a ton of "communication error" messages. This may be a new bug but figured I would bring it up here. Its happening inside matches and while in the queue.

Please let me know what additional information I can provide.

steam-1364780.log.gz

Edit: After some more play testing, the disconnects seem unrelated to the patch and have basically gone away.

If you played within the last few days I would say the errors were the game servers. I had a few but my friends on PS5 also complained about communications errors, etc so I believe it was on Capcom"s side.

Root caused to a game bug. This PR should fix the hang: HansKristian-Work/vkd3d-proton#1647. Should hit bleeding-edge very soon.

Game hangs on an nvidia 2060 mobile on archlinux when trying to boot the game. Immediately after the capcom screen and before the menu loads, the game will hang. It requires deleting config.ini multiple times for it to eventually launch. when it doesn't it launches in windowed mode and there is no fullscreen functionality.

If you like I can reply with a screenshot.

The game freezes on this screen. The screenshot is from the freeze i experienced with the latest proton experimental version:

FWIW Experimental Bleeding Edge seems to work for me. I was able to play about two hours of ranked last night with no crashes. I'm on a fully AMD system.

So I am also not getting crashes so far with experimental, but shutting the game down doesn't work? Basically "going back to desktop" screen freezes and steam app also gets unresponsible, can only kill it through tray/cli.

Basically "going back to desktop" screen freezes and steam app also gets unresponsible, can only kill it through tray/cli.

I have this problem but I just hit stop in steam and it's fine.

I see some graphical mess up during fights:

Still not sure what triggers this. Sometimes the game just works fine for hours.

Proton Logfile: https://gist.github.com/chtaube/1b497ad3c04f24c5d8238e238ef0b44d

This is using Proton Experimental, bleeding-edge. Same Problem has been seen on Proton 8, Proton-GE and Proton Experimental, default branch.

• Hardware Information
• Steam Runtime System Information

Hello @kisak-valve, I've switched to RADV and haven't seen any rendering artifact yet. I think this was the solution to my problem. Thanks for pointing out!

Issue appears to have returned. Black screen, crashed WM and I had to log back into KDE, additionally none of the windows had the title bar anymore after the login. Meaning I couldn't minimize, maximize or close them. Had to reboot to get everything back. This was after a 2-3 of hours of casual online.

steam-1364780.log.gz

Something strange I noticed about high refresh rate monitors and AMD GPU. My monitor supports 144Hz. Trying to set X11 (via xrandr --rate 144) and KDE (via standard configs) to this refresh rate works on the system overall, but causes extremely high GPU usage in game, dropping it to unplayable 30fps, even on low graphics settings. Changing the refresh rate to 120Hz in aforementioned places brings usage back to usable ~60%, similar to playing on 60Hz. Not sure what to make of that, but wanted to note it here in case somebody experiences similar problem. I can't find anything similar for Windows/Consoles, so I'm not sure if that's the problem with the game itself. 120Hz is still good for me, so I'll stick with it for now.

@SuperPrower you are likely affected by this bug: https://gitlab.freedesktop.org/drm/amd/-/issues/2657

Hello @pepi55, you're running the game with AMDGPU-Pro. Can you test how it behaves with mesa/RADV?

How could you tell? The system info gives me Mesa 23? Im confused where you found it as Im not sure what to change right now to get it working with mesa.

Edit: This was indeed the problem, thanks alot! It appears that installing the pro drivers through the package manager does not make them "optional" unlike what the Wiki would suggest at a glance. I was always under the impression that im running games with Mesa and have the option to specify the pro drivers for native games.

Running steam with:
VK_ICD_FILENAMES=/usr/share/vulkan/icd.d/radeon_icd.x86_64.json /usr/bin/steam
fixes the issue and the game loads properly.

Now I wonder how many other "broken" games suffer from the same issue...

Quick update on my case:
I let SF6 away for a couple of weeks.
Then reinstalled an clean distrib on my laptop (Ubuntu instead of Zorin btw, by I tried POP and Fedora in the meantime).
Reinstalled Steam and SF6, and... it works now flawlessly.
I don't know if it was corrupted nvidia drivers, Denuvo DRM or ZorinOS... no explanations.
But at least, case closed!

Compatibility Report

• Name of the game with compatibility issues: Street Fighter 6
• Steam AppID of the game: 1364780

System Information

• GPU: Radeon 7900 XTX
• Video driver version: Driver Version: 4.6 (Compatibility Profile) Mesa 23.2.1-1ubuntu3.1
• Kernel version: 6.5.0
• Link to full system information report as Gist: https://gist.github.com/jcgit1/39ac6f1d6624c72fddaa27df2eaee5b5
• Proton version: Proton Experimental

I confirm:

• [X] 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-1364780.log

Symptoms

While playing online or training mode the game will randomly stutter and lock out controls. No button presses will register while this happens. It continues for about 10 seconds before the controls start to respond again. This only started around 4 days ago.

Reproduction

Start the game.
Go into Fighting Ground -> Online -> Casual Match
Wait for a match to begin
Play the match - after about a minute the stutter will occur. It happens 2-3 times per match.

Tried using Proton Hotfix, Proton 8 and Proton Experimental.

I can't confirm the Shader Compilation issue (AMD RX 6600 card), works as normal for me. Maybe that's an Intel-specific issue?

When I throw tech seems like the graphics freeze. Music is still playing but I have to force quit the game and start again. Running on Ryzen 4500G and Intel Arc A770

Getting slowdowns during some special attacks (consistently it is caused by Hadoukens).
This happens irregardless of graphics settings. Game will start to run at a locked 40fps, this only resets when a Super with a cutscene (mostly level 3‘s) is used. Fps shoots back up to 60 until the next Hadouken.

Tried different Proton versions (8, 9, Experimental, GE) and this happens no matter what.

I can't confirm the Shader Compilation issue (AMD RX 6600 card), works as normal for me. Maybe that's an Intel-specific issue?

I'm having this issue on AMD RX 6800 XT as well.

Working fine on AMD 7700 XT

Latest update to the game has appeared to have broken it on Nvidia afaik. No Proton versions seem to alleviate it, not 8, 9, Experimental, or GE. Game runs at like 15 fps everywhere, even at the title screen. A few people have reported similar results on the Steam discussion page. https://steamcommunity.com/app/1364780/discussions/0/4756451924772514267/

Distro: Pop!OS 22.04 LTS
Kernel: 6.9.3-76060903-generic
CPU: Intel i9-13900K
GPU: RTX 4070
Driver version: 555.58.02
Window Manager: X11

I'm curious as to what the cause is and if it can be fixed soon. Please forgive my formatting. I'm new to the site and I simply wanted to make a report.

FINAL EDIT: It was Split Lock. Follow this link here and it will show you how to address it. I've also heard Gamemode automatically adjusts for splitlock. https://github.com/ValveSoftware/Proton/issues/6817#issuecomment-1595495268

Compatibility Report

• Name of the game with compatibility issues: Street Fighter 6
• Steam AppID of the game: 1364780

System Information

• GPU: AMD RX 6600
• Video driver version: Mesa 24.2.7
• Kernel version: 6.11.11-300.fc41
• Link to full system information report as Gist:
• Proton version: experimental-9.0-20241206

I confirm:

• [x] 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-1364780.log

Symptoms

Game freezes itself as well as the whole os while in "Battle Hub" game mode.
The picture doesn't fully freeze, but the game is totally stuck and doesn't let me do anything.
Can't exit the game as a result, having to reboot the pc.
This happens to me every time I visit Battle Hub.

Reproduction

Get into a Battle Hub lobby, interact with things for a moment, game freezes.
Freezes specifically happen when trying to interact with something like a cabinet or a shop, or while exiting a shop.
It might not happen immediately, but usually it's very quick.

Hello @eOliverr, this is a blind guess, can you try running the game with the launch options set to WINE_OLD_MEDIA_SOURCE=1 %command% and retest?

Hi, tested with the launch command, tried a few times and kept having the same freezing issue.
proton log

OS: Garuda Linux Bird of Prey x86_64
Kernel: 6.12.4-zen1-1-zen
Resolution: 2560x1440
DE: Plasma 6.2.4
CPU: AMD Ryzen 5 5600 (12) @ 4.468GHz
GPU: AMD ATI Radeon RX 7700 XT / 7800 XT
Memory: 4500MiB / 15900MiB

Choppy FPS when monitor refresh is set above 60hz. I've tried multiple refresh rates, vsync on/off, input lag reduction option, but no success. FPS counter still showing 60fps, but something is clearly not right with the game motion. Checked the split_locks issue, but not showing anything related on dmesg.

OS: Garuda Linux Bird of Prey x86_64 Kernel: 6.12.4-zen1-1-zen Resolution: 2560x1440 DE: Plasma 6.2.4 CPU: AMD Ryzen 5 5600 (12) @ 4.468GHz GPU: AMD ATI Radeon RX 7700 XT / 7800 XT Memory: 4500MiB / 15900MiB

Choppy FPS when monitor refresh is set above 60hz. I've tried multiple refresh rates, vsync on/off, input lag reduction option, but no success. FPS counter still showing 60fps, but something is clearly not right with the game motion. Checked the split_locks issue, but not showing anything related on dmesg.

@davernos you could be running into this issue: https://gitlab.freedesktop.org/drm/amd/-/issues/3618

It was giving me very similar problems. It's fixed now in kernel version 6.12.5.

@mattipulkkinen Thanks for the answer. I'll try again with the new kernel. Anyway, I don't think thats the issue. The choppiness goes away when I set the monitor resolution to 60hz.

Choppy FPS when monitor refresh is set above 60hz. I've tried multiple refresh rates, vsync on/off, input lag reduction option, but no success. FPS counter still showing 60fps, but something is clearly not right with the game motion. Checked the split_locks issue, but not showing anything related on dmesg.

@davernos Same on my end, choppiness/stuttering only goes away when I set my monitor to 60hz.

Compatibility Report

• Name of the game with compatibility issues: Street Fighter 6
• Steam AppID of the game: 1364780

System Information

• GPU: AMD AMD Radeon RX 6600
• Video driver version: 4.6 (Compatibility Profile) Mesa 24.3.2-arch1.1
• Kernel version: 6.12.7-arch1-1
• Link to full system information report as Gist: gist
• Proton version: Proton Experimental (EDIT: just noticed I was using bleeding-edge, will re-check soon with normal)

I confirm:

• [x] that I have checked whether there are updates for my system available.

Symptoms

When starting the game, sometimes there are issues with (particle) effects (like smoke, dust, Drive Impact effects, etc.), making them very colorful and losing some of transparency. Issue is inconsistent, best observed when starting the game for the first time after resuming from suspend, though it may be instead related to updates to Proton Experimental (EDIT: just tried suspending and resuming, didn't observe the issue; will wait for Proton update to verify). Tested without mods, gamemode, or mangohud.

My system uses Wayland, but I distinctly remember sometimes having this issue on X11 as well.

Reproduction

1. Suspend the system and resume it (EDIT: or maybe not, waiting for the next Proton update to verify if that's what correlating with issue appearing)
2. Start the game, go into one-on-one battle. Preferably select one of the Metro City Downtown, Suval'Hal Arena, or Macho Ring stages, as I've most often observed the issue there.
3. Observe the colorful smokes, dust, etc.

Game log:
steam-1364780.log

@SuperPrower I remember having similiar issues when my GPU is undervolted. Probably not your case.

@drnooooo Getting desperate with the monitor refresh issue. Tried different distros (Garuda, Nobara and Bazzite), all with the same issue. Maybe a KDE Plasma or Wayland problem? Any luck on your side? Tired of setting the monitor to 60hz everytime I want to play :S

Checking dmesg:

[  253.362362] umip: StreetFighter6.[8445] ip:15d747832 sp:546fd778: SGDT instruction cannot be used by applications.
[  253.362367] umip: StreetFighter6.[8445] ip:15d747832 sp:546fd778: For now, expensive software emulation returns the result.

Does anyone know what is this?

Update to my earlier issue report https://github.com/ValveSoftware/Proton/issues/6817#issuecomment-2543091226
I no longer have this issue. I started experiencing similar freezes with other applications which led me to doing a clean OS install. Now the game is running perfectly.

@drnooooo Getting desperate with the monitor refresh issue. Tried different distros (Garuda, Nobara and Bazzite), all with the same issue. Maybe a KDE Plasma or Wayland problem? Any luck on your side? Tired of setting the monitor to 60hz everytime I want to play :S

It's not a Plasma issue, nor is it a Plasma Wayland issue, it's also not your distro. Honestly the only explanation left is your GPU drivers, since you're on AMD. I've got 2,500 hours in this game, all of them on Plasma on Arch Linux, and over 100 on Wayland (since the 570 drivers released), and I've never once had a single issue with stuttering related to refresh rate. Regardless of whether I'm in X11 and so both my monitors are stuck at 165Hz, or on Wayland where G-Sync is available (but both are still configured to run at 165Hz, just with G-Sync.

I've never enabled V-Sync, this is a goddamn fighting game. That's insanity.

The only difference is I'm on Nvidia.

Checking dmesg:

[  253.362362] umip: StreetFighter6.[8445] ip:15d747832 sp:546fd778: SGDT instruction cannot be used by applications.
[  253.362367] umip: StreetFighter6.[8445] ip:15d747832 sp:546fd778: For now, expensive software emulation returns the result.

Does anyone know what is this?

Yes, and it's 100% irrelevant to your issue.

I use hitbox and arcade joystick, both use the same board, GP2040-CE on a Raspberry Pico, when I happen to change one for the other, I need to restart the game, wouldn't there be a command to solve this?

I use hitbox and arcade joystick, both use the same board, GP2040-CE on a Raspberry Pico, when I happen to change one for the other, I need to restart the game, wouldn't there be a command to solve this?

I know that on Steam Deck there is an option to change the controller order/priority. I wonder if there is something similar in the steam overlay or in the controller settings?

I use hitbox and arcade joystick, both use the same board, GP2040-CE on a Raspberry Pico, when I happen to change one for the other, I need to restart the game, wouldn't there be a command to solve this?

I main a Mayflash arcade stick BUT I also have a BITFUNX GP2040-CE-based leverless controller to start building up leverless muscle memory, and any time I want to go from the stick to the leverless, I have to restart the game. And yes I tried @SuperPrower 's idea many times. I even tried running with Steam Input disabled.

Meanwhile, I can go from the leverless to the arcade stick by doing literally nothing except unplugging the leverless and either moving the Arcade Stick input method (Dinput, Xinput, etc) before immediately moving it back to its correct position (this serves as a power cycle), or I just press the switch on the USB hub port the stick is plugged into to turn it off then back on.

Although for your specific use case I think you can just launch with both connected, the game will assign one to Player 1 and the other to Player 2. And any time you wanna switch, I believe you can just have Player 2 be the player playing in any matches, and if for some reason I'm imagining that, I know there are ways. When I sit down to play tonight I'll see what I can find out.

Would you mind explaining like... why you need to use both controllers in one session? If they're both GP2040-CE then they're gonna be the same except maybe size and where any extra macro buttons might be, but in that case it shouldn't be hard to choose which one you prefer. But if you could explain your reasoning it might help someone come up with a solution that gives you what you want via an outside the box method.

Also, if you could tell me the model for each controller, which Proton version you're using, and each controller's info line from lsusb -tvv or even just lsusb. Cause that info is necessary if the time comes where we need to do some Linux fu.

Replying to https://github.com/ValveSoftware/Proton/issues/6817#issuecomment-2833857756

Hello...

I won't be playing on both forever, I just bought my hitbox, and I'm giving up on my Full Sanwa (I never got used to it). Honestly, I prefer to learn on the hitbox and know that I'll "never again" have problems or miss a combo.

However, even if I decide to use only the Hitbox, the fact that I disconnect a controller and it stops working is still not cool. I thought this type of problem was something common, and I expected to solve it with just some line at startup. If it's something very complex, then it's not a problem, I'll adapt to having to restart, or just open the game with the controller already connected.

I was experimenting with setting my headphones to the sample rate of 96000 Hz in the Wireplumber config (because they click a bit during the normal use, but that turned out to be the common issue with the model of the headphones), and with that setting, the sound in the game was weirdly muted and hollow and noticeably delayed (I also tried using speakers and the issue was present as well). I switched from Proton Experimental to 9.0-4 thinking it was an issue with the Proton, but audio went completely bad, playing one note every second. After that I changed the Wireplumber setting back to 48000, restarted it, and both issues went away. Not sure if that's related to Proton or to the game itself, so leaving this comment here just in case someone encounters similar problem.

@procobain , I only have one GP2040 controller, but I tried disconnecting it in the game and connecting it back again and it works. It also works if I connect it after starting the game. Maybe you should try exiting to the main menu (mode select, or even the Fighting Ground screen) before you want to change your controller? Works for me both on Proton Experimental and 9.0-4, I have the Steam Input disabled and my controller is in PS4 mode. If the issue is only when you have two controllers, I could experiment a bit as I have a spare GP2040 board laying around, just need to update the firmware and connect some buttons to it.

Operating System: Fedora Linux 42
KDE Plasma Version: 6.3.5
KDE Frameworks Version: 6.14.0
Qt Version: 6.9.0
Kernel Version: 6.14.6-300.fc42.x86_64 (64-bit)
Graphics Platform: Wayland
Processors: 16 × AMD Ryzen 7 5700X3D 8-Core Processor
Memory: 31.2 GiB of RAM
Graphics Processor: AMD Radeon RX 7900 XT
Proton version: Proton Experimental

I am experiencing cases where upon the entering matches, there is massive bloom among what I think might be the particles. This has been been randomly in quite rare instances and I'm not quite sure what causes this but it has been happening for a while.
I was unable to get logs this time as it doesn't happen very often. If this occurs again, I'll try to have a log prepared.

Managed to get the issue to occur again. I've included a clip and managed to include a log as well.

steam-1364780.log

https://github.com/user-attachments/assets/8b198f86-845b-4d74-a640-a8a0f5eea4cc

@Yemster-1020 , reminds me of an issue I had where particles were very colorful, but not flashbang-white (see above in the topic). I've since switched to using stable proton rather than experimental and letting Steam download the shaders, and haven't seen the problem - maybe you could give it a shot.

@Yemster-1020 can you give more details?

1. When the issue happens, does it last for the whole match? Are the following matches guaranteed to have the same issue?
2. What graphic settings are you using?
3. When did you first notice this issue? Weeks or months ago?
4. How often do you see this issue happen? I imagine this is variable, so a rough estimation is fine. 20 matches? More?

@simifor

1. So the issue pretty much persists for the entire match and any subsequent matches until I close and restart the game.
2. I've included screenshots of the current graphic settings I am using, hopefully that will be of use.
3. I believe I started noticing issue a few months ago but it could possible be more.
4. So the weird thing is that I see this issue quite rarely but once it happens, I'm pretty certain it's guaranteed to continue until the game is reset. It pretty much starts as soon as boot into your first match in most stages, once you reset the game though the issue doesn't occur. Maybe it's coincidence but a few days ago it did trigger again after I tried messing around with my GPU settings on LACT.

EDIT: Just thought I add this reddit thread as there has been others experiencing the problem so seems to be fairly common atm.

@Yemster-1020 , reminds me of an issue I had where particles were very colorful, but not flashbang-white (see above in the topic). I've since switched to using stable proton rather than experimental and letting Steam download the shaders, and haven't seen the problem - maybe you could give it a shot.

I might try that, thanks. The issue thankfully doesn't occur often enough for it to be a pain (though I may have to test stages before I queue for matches just to be sure), just a tad annoying when it does pop up.

@Yemster-1020 so just to make sure I understand correctly, if the issue doesn't happen the first time you boot a specific map the issue won't appear on that map on subsequent matches? One comment there mentions they mostly get it in 3 specific maps, have you noticed if it happens more often in certain maps? If so, which ones?

@simifor I'm pretty sure it doesn't matter if I boot into a specific stage the first time or not. If I boot into a match, if it's one of the problematic stages, then it will appear in those stages in the first match and any subsequent matches within those stages until I restart. If I boot up the game and it doesn't appear in a match in a problematic stage, then it doesn't appear during the rest of my session regardless of stage.

Training Room is one of the stages I'm pretty certain it never happens in. I have had instances where I boot into Training Stage for say Training Mode and it's fine then get into a match in one of the problematic stages and the issue occurs.

Dhalsim's Temple, Macho Ring, Bather's Beach, Genbu Temple are stages where I have seen this occur.

I'm having the same issue of low fps just on online matches. I tried checking the logs with sudo dmesg | tail and got this output:

[19567.539289] umip: StreetFighter6.[25582] ip:15df9171e sp:53eff458: For now, expensive software emulation returns the result.
[19568.676165] [drm:nv_drm_gem_alloc_nvkms_memory_ioctl [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000600] Failed to allocate NVKMS memory for GEM object
[19568.676192] [drm:nv_drm_gem_alloc_nvkms_memory_ioctl [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000600] Failed to allocate NVKMS memory for GEM object
[19569.568732] [drm:nv_drm_gem_alloc_nvkms_memory_ioctl [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000600] Failed to allocate NVKMS memory for GEM object
[19570.358252] [drm:nv_drm_gem_alloc_nvkms_memory_ioctl [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000600] Failed to allocate NVKMS memory for GEM object
[19570.358277] [drm:nv_drm_gem_alloc_nvkms_memory_ioctl [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000600] Failed to allocate NVKMS memory for GEM object
[19570.878963] [drm:nv_drm_gem_alloc_nvkms_memory_ioctl [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000600] Failed to allocate NVKMS memory for GEM object
[19604.494403] umip: StreetFighter6.[25582] ip:15df9171e sp:53eff538: SGDT instruction cannot be used by applications.
[19611.564297] [drm:nv_drm_gem_alloc_nvkms_memory_ioctl [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000600] Failed to allocate NVKMS memory for GEM object
[19611.564321] [drm:nv_drm_gem_alloc_nvkms_memory_ioctl [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000600] Failed to allocate NVKMS memory for GEM object

It seems this is different from the split lock issue since I don't see it in my logs. I also tried the suggested workaround but still didn't work.

My system specs:
OS: Fedora Linux 42
Kernel: Linux 6.14.9-300.fc42.x86_64
CPU: AMD Ryzen 5 5600X
GPU: NVIDIA GeForce RTX 3060 Ti

I'm having the same issue with the fps , in any offline play the fps is from 55-60 never a locked 60 which is tolerable, but in any online match its always between 45-50 fps never above.

I've tried multiple proton versions and even proton-ge and its all the same.

System specs:
Distro: Ubuntu 24.04.3 LTS
Kernel: 6.14.0-28-generic
RAM: 16 GB
GPU Driver: NVIDIA 575.64.03
GPU: NVIDIA GeForce RTX 4050 Laptop GPU
CPU: 13th Gen Intel Core i7-13650HX

@zagerus @Skydler maybe its a bit more informative if you let everyone know what your GPU & CPU utilization is while being FPS limited.

But ya, i believe worse performance during online suggests your being CPU bottlenecked somehow, as the rollback netcode adds some extra CPU overhead.

@zagerus did u verify that splitlock is not causing ur fos limit? (Earlier posts in this thread by me describe how to diagnose and solve the splitlock issue).

Latest version of Proton (10.0) has playback issues with the logo intro video.
When the logo video plays, it plays in slow-mo instead of full speed. From my testing, Proton 9.0-4 doesn't have this issue.

Happens on my RX6800 (w/ 5700x3D), and 5600g iGPU systems.

Hi @kisak-valve ! Thanks for replying.

Here's the gist: https://gist.github.com/1TheBlueWii1/9c5eb4762fdf7a17db5ad5ab2e94584e

Here's the game's log file:
steam-1364780.log

And here's a recording demonstrating what the issue looks like:
https://www.youtube.com/watch?v=mKoHJM_7EJc

The "RE Engine" video plays fine, but when the logo video plays, its unusually slow.

@1TheBlueWii1 I can also see the slow playback on proton 10, but on proton experimental it seems to play correctly. Given that proton experimental is the default for this game, can you check if it plays normally for you in that version?

@1TheBlueWii1 I can also see the slow playback on proton 10, but on proton experimental it seems to play correctly. Given that proton experimental is the default for this game, can you check if it plays normally for you in that version?

Got hit with the game's DRM, since I've been swapping Proton versions and running it on different machines multiple times. I'll post about it in 24 hours!

Tested Proton Experimental. Same issue there.

@1TheBlueWii1 thank you for checking. I ran it again, and I'm also seeing the slowdowns now. I wonder if there's a rare chance for it to play normally, or if my perception was off.

I report having a similar issue on an Intel Arc A770 with Mesa drivers

I run the game with the default steam proton and it crashes after entering a match, randomly, sometimes just after choosing the characters, sometimes after one minute sometimes after a few more. But even if that doesn't cause a crash blocking a throw has a 100% chance of freezing the game and sometimes freezing the entire machine.

At the moment of the throw blocking inside the game, journalctl logged the following output from the kernel

Sep 24 11:38:31 pc kernel: i915 0000:09:00.0: [drm] GPU HANG: ecode 12:1:85def5fb, in StreetFighter6. [343912]
Sep 24 11:38:31 pc kernel: i915 0000:09:00.0: [drm] StreetFighter6.[343912] context reset due to GPU hang
Sep 24 11:38:31 pc kernel: Fence expiration time out i915-0000:09:00.0:StreetFighter6.[343912]:181ecc!
Sep 24 11:38:31 pc kernel: Fence expiration time out i915-0000:09:00.0:StreetFighter6.[343912]:181eca!
Sep 24 11:38:31 pc kernel: Fence expiration time out i915-0000:09:00.0:StreetFighter6.[343912]:181ec8!
Sep 24 11:38:31 pc kernel: Fence expiration time out i915-0000:09:00.0:StreetFighter6.[343912]:181ec6!
Sep 24 11:38:31 pc kernel: Fence expiration time out i915-0000:09:00.0:StreetFighter6.[343912]:181ec4!
Sep 24 11:38:31 pc kernel: Fence expiration time out i915-0000:09:00.0:StreetFighter6.[343912]:181ec2!

Replying to [#6817 (comment)](https://github.com/ValveSoftware/Proton/issues/6817#issuecomment-2833857756)

Hello...

I won't be playing on both forever, I just bought my hitbox, and I'm giving up on my Full Sanwa (I never got used to it). Honestly, I prefer to learn on the hitbox and know that I'll "never again" have problems or miss a combo.

However, even if I decide to use only the Hitbox, the fact that I disconnect a controller and it stops working is still not cool. I thought this type of problem was something common, and I expected to solve it with just some line at startup. If it's something very complex, then it's not a problem, I'll adapt to having to restart, or just open the game with the controller already connected.

@procobain @SuperPrower so, I completely forgot to update my comment, but literally like a week after I made my comment, I was able to not only unplug one controller and connect another while in-game (even mid-match), but also even use two controllers at the same time - No, not as Player 1 and Player 2, I mean both as Player 1.

I have a GP2040-CE-based COSMOX C16, and a Mayflash F300. Both controllers have multiple USB ports for passthrough and PS4/PS5 adapters. If I plug in the mayflash, and then connect the C16 to the USB port ON the Mayflash and NOT directly to the PC, then I can literally use both at the same time. As in, I can be using my C16, and mid match I can set it down, pick up the Mayflash, and start using it, they both act as Player 1.

Beyond that, if I'm using just one controller at a time I can freely unplug and replug it as many times as I want, and it works every time. And I don't have to disable SteamInput, And I damn sure don't have to use Dinput (PS4 mode), I use Xinput on both controllers, but every mode - Xinput, Original Xbox, XboxOne, PS4, PS5, Nintendo Switch, Xinput, Keyboard mode on the C16, and Xinput + Dinput on the Mayflash.

No special Proton is necessary. I use proton-cachyos, but Proton-GE, and Proton Experimental both work fine.

Compatibility Report

• Name of the game with compatibility issues: Street Fighter 6
• Steam AppID of the game: 1364780

System Information

• GPU: RX 9070 XT
• Video driver version: Mesa 25.2.4-cachyos1.2
• Kernel version: 6.17.1-2-cachyos (64-bit)
• Link to full system information report as Gist: https://gist.github.com/Aryetis/85fb591c497597cd2e7a133df5cce3fd
• Proton version: Proton 1760113612 experimental-10.0-20251006b

I confirm:

• [x] 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.

Symptoms

The application's window flickers A LOT in borderless fullscreen but works fine when "windowed".

Here's some footage of the issue uploaded on youtube ( epilepsy warning, flickers start at 1:21)

Problem occur in KDE Wayland, KDE x11, Hyprland and probably others. It does not appear to be related to the DE.

Reproduction

1. Launch Street Fighter 6 (AppID : 1364780)
2. Go to the option with the Tab key, put the game in fullscreen mode.

Additional information and mitigation attempts

1. Attached is the Proton log of the recorded and uploaded game session
2. I already tried to run with multiple proton versions, including GE ones, cachyOS ones
3. I already tried launching the game with various combinations of the following launch arguments : PROTON_ENABLE_WAYLAND=1 , PROTON_USE_NTSYNC=1 , PROTON_NO_ESYNC=1 , PROTON_NO_FSYNC=1
4. The game seems to detect 4 displays when I only have two (cf 1:13 in the video)
5. I also suffer from slowdowns when playing the game outside of the training stage regardless of the split_lock_mitigate settings. Disabling "Input Delay Reduction" in the game's settings does help a lilttle but not enough.

steam-1364780.log

Workaround

In KDE (wayland), I can right click the a the top part of a window and edit its settings for specific programs. Among those settings that can be tweaked is an option to "enable/disable titlebar and frame"'.... So yeah I just play in windowed mode in a frameless and titleless window.

Additional Workaround (for slowdown this time)

Turns out the game works fine if I turn off my secondary (60hz, portrait) monitor and play on my main (144hz, landscape) one. Yes I tried to play with my main monitor at 60hz and/or with my secondary monitor in landscape. No it does not fix the flicker. So to play this game I have to play in windowed mode + turn off the frame and title bar of the game's window + turn off my secondary monitor. Kinda a pain not gonna lie. But at least it works now (until I find the next problem).

Also no, my slowdowns had nothing to do with split_lock_mitigate in my case.

@sanroxcode - AFAIK the game doesn't directly support LAN play, if you've used a Custom Room for a "local match" it still has to go through the game coordinator or something like that, but of course afterwards it should be p2p. I've played in a lobby some days ago on Proton Experimental, on Arch Linux, also with AMD GPU and a Realtek ethernet chip in my motherboard, with no such ping issues. I would propose you to somehow test outside of your local network, maybe with a mobile hotspot (when phone isn't connected to your local Wi-Fi).

Did anyone tried to fix the fps performance issue that have been happening since last week?

As someone who has more hours in this game on linux than literally anyone else (by several times over), I know for a fact that PROTON_ENABLE_WAYLAND=1 has been working for the last several months with no issues, and it's very important that I use native Wayland for the game because of the reduction in input lag, as someone who plays this game competitively on semi-serious level.

But starting 2 days ago, it just flat-out stopped working. Doom Eternal still works with Wayland (xprop acts as if I haven't clicked anything when I click the Doom Eternal window, meanwhile SF6 gives me the full output and shows it's an X window. This never happened before, and more than that it was obvious when it would be working because the initial launch window had a blue titlebar that would disappear when the window went fullscreen. That's all gone).

@kisak-valve I hate to ping you, but this is extremely serious I absolutely have to have this fixed, but I have no idea where to start (I've already tried all the usual steps. Different kernels, different GPU drivers, different proton versions, the whole lot).

Oh the other huge thing is screen capture gets messed up when it's not native wayland, and more importantly, the game stops receiving inputs unless the window is in focus (I use two monitors). So if I'm mid match and accidentally switch windows, I can lose a match because my inputs stop working until I refocus the game window. This is not an issue on native wayland, I can have any window focused I like and the game still receives inputs. This is EXTREMELY frustrating.

Hey @kisak-valve ! The regression I mentioned seems to be fixed in Proton 11 Beta! No more weird slow downs with the logo screen.