-
Notifications
You must be signed in to change notification settings - Fork 921
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
"Hoshi no Otome to Rikka no Shimai" (AdvHD.exe) crashes on startup with DXVK enabled #4662
Comments
Something uncommon in your log. These errors come from methods that read monitor EDID.
dxvk/src/wsi/win32/wsi_monitor_win32.cpp Line 291 in 8860bde
Line 656 in 8860bde
The trial version of game is available on official website (NSFW warning!). Link is NSFW!!!Game information: Reproducable on Windows on both this game and LoveKami Useless Goddess (game with same engine, engine version 1.9.5.0, tested dxvk 2.5.3). Both game directly crash after launch with same exception. See my a bit complex investigation (may be helpful or not)Information from debugger:
Stack when exception: Collapsed, very long stack
The game imports d3d11.dll, dxgi.dll, d3d9.dll, d3dx9_43.dll. The game only creates D3D11Device in start option menu (and use this to list resolution and other system information), and creates both D3D11Device and D3D9Device after start pressed. But SpecialK says it's a d3d9 game. Only use dxvk d3d9.dll: menu shows, crash on start This engine's behavior is a bit confusing. Need to figure out how it interact with both d3d9 and d3d11. |
Very pro comment, thanks! So will it work if we just return a "fake" edid of a generic device instead of an empty one? |
Not sure. The engine's behavior is really confusing. I need to figure out how this engine interact with both d3d9 and d3d11 first. |
I am excited to see how this develops and looking forward to your updates. |
The monitor/display errors are harmless |
Yes I have confirmed this. The game always tries to enum monitors on system default soft render device (WARP on Windows), and of course there should be no monitors on WARP device. Made a better apitrace on Windows with Intel Xe 80EU TigerLake on driver 101.5768. (trace is SFW) https://github.com/qinlili23333/PageConnect/releases/download/TEST/AdvHD.apitrace.7z When trace with dxvk on Windows, the trace file is empty. Even if don't display this window ticked, the game still calls d3d11 and dxgi during launch. Tried to replay apitrace with dxvk, and got this... And millions of buffer overflow in apitrace output. |
I have found solution. Same commit's mingw build is still broken. So I would say, this issue is compiler toolchain issue. |
Have you verified that is the case? If it is related to the crash I have been poking at locally then it isn't fixed by that PR. |
Not checked yet. Maybe later. |
Might as well comment my own findings from poking at the game. At least here locally the game and it's launcher will crash with builds compiled with the latest mingw-w64 bump https://archlinux.org/packages/?sort=&q=mingw-w64. Current version is 14.2.0 for gcc, 12.0.0 for headers, crt and winpthreads and binutils is 2.43. Downgrading to their previous version and compiling dxvk makes it work again. Apparently with 12.0 there was a change of the default C runtime so it is UCRT now instead of MSVCRT as previous. Compiling my mingw-w64 system packages with The error shown with a bad build is
which happens on this line in the dxvk/src/util/config/config.cpp Line 1217 in 5569274
It seems like, as it goes through all the builtin configs, that it crashes when it meets a regex that contains a range. They are the square brackets that can be seen in two builtin configs.
I am unsure why this specific game is special and causes a crash as it doesn't happen with any of the other games i have installed. It isn't the path with Japanese characters as it happens even when those are removed. Something special with the exe maybe. |
(Referring to deleted comment) I presume you didn't get it working with MingW anyway since you reopened? |
Alright it seems to be crashing again on my Windows 7 VM.(Sorry for the mistake i've made with using the wrong 64bit dll) Read our privacy statement online: If the online privacy statement is not available, please read our privacy statement offline: The generated AdvHD_d3d9.log and AdvHD_CHS_d3d9.log is empty. Tried the Chinese translation patch to see if the translators have some quirks but apparently not. |
I think so, since none of these builds can run the game on my local wine installation. |
Can't confirm the regex crash, it just crashes for some random reason with a completely nonsensical stack trace that points from dxgi to mfplat (??) on my end. Not sure how to debug that though. |
Me on wine too. :( |
Can you try the PR above (#4668)? Works with that on my end now. |
In regards to my regex crash it was found that it only happens with the Windows version of #4669 tries to work around the crash. |
Did you try the game provided one in runtime/ucrt in game folder? In my experience that specific UCRT runtime can be installed from that game's installer. |
Yep, that fixed the game. But on a Windows 7 VM, the one mentioned above, still crashes with zero logs generated. I believe it's an outdated OS issue instead of ours. |
Windows 7 lacks ucrtbase.dll. |
Wilco |
Doesn't work. Maybe my VM is cursed? |
StackHash_0a9e usually occurs when display driver is not configured in the right way. This is common in VM since VM need to wrap display driver on host driver. Use GPU passthrough for VM may help. I would also suggest to get a new Windows 11 VM for test. |
I believe so. Maybe we can close this now and call it a day since dxvk is already working fine on others' systems.
…On February 11, 2025 8:51:23 PM GMT+08:00, "琴梨梨OvO" ***@***.***> wrote:
> Doesn't work. Maybe my VM is cursed?
StackHash_0a9e usually occurs when display driver is not configured in the right way. This is common in VM since VM need to wrap display driver on host driver. Use GPU passthrough for VM may help. I would also suggest to get a new Windows 11 VM for test.
--
Reply to this email directly or view it on GitHub:
#4662 (comment)
You are receiving this because you modified the open/close state.
Message ID: ***@***.***>
|
Another game from Essemble using the WillPlus (Advhd) engine. The same problem happens on many other Essemble games too. You open the game's launcher, click start, and boom. Pretty sure not a wine issue since disabling DXVK makes it work.
This problem is also comfirmed on "Otome no Ken to Himegoto Concerto", which is using the same engine.
Software information
Hoshi no Otome to Rikka no Shimai. VNDB Link: https://vndb.org/v31700
With DXVK enabled, using wine 10-rc2.
System information
Apitrace file(s)
Disabled DXVK: https://oshi.at/zQpu
Enabled DXVK (Crashes): https://oshi.at/QPmb
Log files
game.log
AdvHD_d3d11.log
The text was updated successfully, but these errors were encountered: