[d3d9] Codename: Panzers, Phase One (411310) Visuals Freeze on Cinematic Intro #3196

zzhiyi · 2023-01-18T07:10:28Z

Software information

Steps to reproduce:

Install Codename: Panzers, Phase One (SteamID 411310) and run it.
You should see an intro video that shows a countdown from 5 to 4, 3, ... click the mouse left button to skip the video

Expected result: the video is skipped and the game's main menu is shown.
Actual result: The frame when the video is skipped remains on the screen. Wined3d doesn't have this issue. The game is still running because you can move the mouse to the lower right area to exit the game.

The video being played is "intro.bik"
$ file intro.bik
intro.bik: Bink Video rev.i, 640x480, 3801 frames at rate 1/25, 1 audio track 44100Hz stereo

System information

GPU: AMD 5500XT
Driver: MESA RADV 22.2.1
Wine version: Wine-8.0-rc4
DXVK version: 1.5.1~2.0

Log files

d3d9.log:
PANZERS_d3d9.log

Blisto91 · 2023-01-18T08:02:26Z

Hello.

Would you be able to make a apitrace with wined3d?
https://github.com/doitsujin/dxvk/wiki/Using-Apitrace

zzhiyi · 2023-01-18T09:00:20Z

Hello.

Would you be able to make a apitrace with wined3d? https://github.com/doitsujin/dxvk/wiki/Using-Apitrace

Hi, I should be able to get one. But I am not near a computer right now. I will get you the file in a few days.

llde · 2023-01-19T23:55:32Z

I have a retail version of the game, I will try to reproduce myself tomorrow

llde · 2023-01-20T21:47:25Z

Hi. It seems that it seems that either DXVK or Winevulkan regressed in last version for this game (I compiled DXVK from master today and wine is the last staging ).
After the video playback (either normal completition or skipping) is causing an Assertion error:
Assertion failed: !status, file ../wine-git/dlls/winevulkan/loader_thunks.c, line 2703

line 2703 being inside vkCreateGraphicsPipelines loader thunk

Other applications doesn't seem to be affected (tested Control)

Wined3d opengl backend can play this game fully, while the vulkan backend complain about missing 1 bit texture format.

mbriar · 2023-01-20T21:56:37Z

Unlikely to be a winevulkan regression, winevulkan will just assert there if vkCreateGraphicsPipelines() failed, so it's either a driver or DXVK problem. Can you try with older DXVK and/or run this with vulkan validation layers? FWIW, wined3d vulkan doesn't support d3d9 and older at the moment because vkd3d-shader doesn't support older shader models yet.

llde · 2023-01-20T21:57:50Z

@mbriar Vulkan validation layer just prevent this game to load at all with DXVK for me. Not a problem with the layers itselves, as they do work with Control

doitsujin · 2023-01-20T22:02:18Z

Assertion failed: !status, file ../wine-git/dlls/winevulkan/loader_thunks.c, line 2703

This just means the call crashed inside the driver.

Kinda need an apitrace to debug this if you're unable to get validation to work. Would highly appreciate one since we're planning to do a release very soon and I don't want to introduce random regressions, but I also don't want this to block that release forever.

llde · 2023-01-20T22:10:24Z

It seems to definetly be a DXVK regression that happened after 2.0.
If I use DXVK 20.0 the behaviour is consistent, with the screen hang after video playback.
WIth latest master I get the first load with the previous behaviour, while all the successive ones are crashing. (First execution after it's executed with DXVK 2.0 release )
Now why the first execution have a difference here?
I will make an apitrace.

mbriar · 2023-01-20T22:12:59Z

Seems to hit this assertion with a RADV debug build

mesa/src/vulkan/runtime/vk_graphics_state.c:656: vk_multisample_state_init: Assertion `ms_info->rasterizationSamples <= MESA_VK_MAX_SAMPLES' failed.

mbriar · 2023-01-20T22:15:43Z

It segfaults with validation layers right after this

VUID-VkSubmitInfo2-semaphore-03881(ERROR / SPEC): msgNum: -1771252314 - Validation Error: [ VUID-VkSubmitInfo2-semaphore-03881 ] Object 0: handle = 0xf56c9b0000000004, type = VK_OBJECT_TYPE_UNKNOWN; Object 1: handle = 0x7d4b7dd0, type = VK_OBJECT_TYPE_QUEUE; | MessageID = 0x966cd5a6 | vkQueueSubmit2(): pSubmits[0].pSignalSemaphoreInfos[0] signal value (0x4) in VkQueue 0x7d4b7dd0[] must be greater than current timeline semaphore VkNonDispatchableHandle 0xf56c9b0000000004[] value (0x72654b68) The Vulkan spec states: If the same semaphore is used as the semaphore member of both an element of pSignalSemaphoreInfos and pWaitSemaphoreInfos, and that semaphore is a timeline semaphore, the value member of the pSignalSemaphoreInfos element must be greater than the value member of the pWaitSemaphoreInfos element (https://www.khronos.org/registry/vulkan/specs/1.3-extensions/html/vkspec.html#VUID-VkSubmitInfo2-semaphore-03881)
    Objects: 2
        [0] 0xf56c9b0000000004, type: 0, name: NULL
        [1] 0x7d4b7dd0, type: 4, name: NULL

mbriar · 2023-01-20T22:23:06Z

trace: https://mega.nz/file/KzwwjCJA#XQfK44IlufN85v-ROIempQAORee_BmT6M7i1ELZpFJg

llde · 2023-01-20T22:24:47Z

Thanks, I'm unable to generate a trace. I get the apitrace: loaded into C:\Program Files (x86)\Panzers\PANZERS.EXE
but when I look for the trace there is nothing

mbriar · 2023-01-20T22:26:08Z

You need to look for the trace in $WINEPREFIX/drive_c/users/$USER/Desktop.

llde · 2023-01-20T22:33:29Z

@mbriar They aren't there.

BTW replaying the trace, crash apitrace either with the same assertion using master or an memory violation on 2.0.
So it seems consistent.

Trace do crash even when replayed by WineD3D however

mbriar · 2023-01-20T22:40:46Z

The trace replays fine for me with wined3d, but crashes/asserts with dxvk.

llde · 2023-01-20T22:54:31Z

Ok It finally generate a trace for me too. If I use a trace generated by a good run of the game it will replay with WineD3D and DXVK 2.0, but will crash on master.
If I use a trace generated from a bad take it will crash everywhere.
(Apitrace will not show the stuck video BTW )
@mbriar sorry you were right, forgot to disable the DLL override

At this point I will do a bisect.

mbriar · 2023-01-20T23:01:39Z

I don't know how the game renders it's intro video, but doesn't look like it's using d3d9 or opengl for it. Neither the DXVK hud nor the gallium hud shows up for it. So it makes sense that apitrace doesn't capture it.

llde · 2023-01-20T23:44:31Z

Bisect result
e5157a5 is the first bad commit
commit e5157a5
Author: Philip Rebohle [email protected]
Date: Mon Jan 9 21:26:15 2023 +0100

[dxvk] Add pre-rasterization pipeline libraries to the state cache

This allows compiling tessellation or geometry shader pipelines
early while still using the pipeline library path.

Also removes compute shaders. Since API-provided compute shaders
are always compiled early, supporting them is no longer needed.

Blisto91 · 2023-01-20T23:47:27Z

Try to set dxvk.enableGraphicsPipelineLibrary = False in a dxvk.conf

doitsujin · 2023-01-20T23:47:53Z

That's... interesting since we should never hit any of those code paths in D3D9.

mbriar · 2023-01-20T23:54:12Z

Try to set dxvk.enableGraphicsPipelineLibrary = False in a dxvk.conf

I didn't enable GPL for RADV, so I don't think this would have an effect?

llde · 2023-01-20T23:55:05Z

@Blisto91 No it doesn't have effect

Blisto91 · 2023-01-20T23:55:23Z

Oh sorry i totally missed it was AMD cards involved. I probably need to to sleep soon 😅

llde · 2023-01-20T23:58:06Z

I was thinking, this is related to a crash that's happening only on non first execution. The only thing it's changing is the state cache, and the e5157a5 is state cache related.
MAybe something is written wrong to the cache (it may also explain why updating or downgrading DXVK solve the issue for the firs run after the change)

doitsujin · 2023-01-21T00:02:36Z

Yes, this is a corrupt state cache. I can reproduce this with the trace but don't know what's going on yet.

doitsujin · 2023-01-21T00:15:40Z

Should be fixed now, thanks @mbriar for providing the trace and info.

doitsujin · 2023-01-21T00:19:22Z

Note that you might need to delete your state cache before this works for now.

I'll add another commit that will basically invalidate all v16 state caches created since the bad commit, unfortunately it is not possible to recover corrupted cache files since we lose some information.

llde · 2023-01-21T00:24:41Z

@doitsujin I confirm the crash is solved.
The Video Freeze is not, so it may be worth to repen this issue

doitsujin · 2023-01-21T00:26:38Z

That is a separate issue, but from what I understand, video playback is not done through D3D anyway?

In which case DXVK can't really do much.

doitsujin · 2023-01-21T00:28:12Z

well anyway, things do get rather confusing when different bugs are discussed in one report.

llde · 2023-01-21T00:30:32Z

@doitsujin While that seems true (video not done with D3D), it's working with WineD3D. So there may still be something on D3D side at the end of the video DXVK is missing.

K0bin · 2023-01-24T19:32:12Z

Can someone make an apitrace of the intro cinematic?

The one @mbriar made works fine with DXVK.

mbriar · 2023-01-24T19:43:54Z

I don't know what else to trace, like mentioned, neither the DXVK hud nor the gallium hud shows up for the video playback (they do later in game of course) and the apitrace doesn't contain the intro.

Blisto91 · 2023-01-24T19:46:07Z

https://mega.nz/file/w4RXybDA#slgLDO3kMj25S2X7qPiGfr1YW0oE4Ypd_XShKwRgfNk

Edit: oh i see what mbriar means now so this trace doesn't help. Hadn't checked the traces before.

mbriar · 2023-01-24T19:48:38Z

Yeah, this one doesn't have the intro video in it either.

Blisto91 · 2023-01-24T20:20:13Z

Found that d3d9.enableDialogMode = True works around the issue.

Fixes doitsujin#3196, not to mention that this was god-awful code anyway.

K0bin added needs apitrace d3d9 labels Jan 18, 2023

doitsujin closed this as completed in 3d24560 Jan 21, 2023

doitsujin reopened this Jan 21, 2023

This was referenced Jan 24, 2023

[util] enableDialogMode for Codename Panzers Phase One/Two #3209

Merged

D3D9 Partial Presentation Mega-issue #2749

Closed

Blisto91 mentioned this issue Mar 5, 2023

[D3D9] Fix for back buffer data loss during D3D9 present #2967

Closed

misyltoad closed this as completed in #3209 Mar 7, 2023

Etaash-mathamsetty pushed a commit to Etaash-mathamsetty/dxvk that referenced this issue Jul 16, 2023

[dxvk] Fix reading out-of-bounds memory in state cache

c361348

Fixes doitsujin#3196, not to mention that this was god-awful code anyway.

[d3d9] Codename: Panzers, Phase One (411310) Visuals Freeze on Cinematic Intro #3196

[d3d9] Codename: Panzers, Phase One (411310) Visuals Freeze on Cinematic Intro #3196

Comments

zzhiyi commented Jan 18, 2023 • edited Loading

Software information

System information

Log files

Blisto91 commented Jan 18, 2023

zzhiyi commented Jan 18, 2023

llde commented Jan 19, 2023

llde commented Jan 20, 2023 • edited Loading

mbriar commented Jan 20, 2023 • edited Loading

llde commented Jan 20, 2023

doitsujin commented Jan 20, 2023 • edited Loading

llde commented Jan 20, 2023 • edited Loading

mbriar commented Jan 20, 2023

mbriar commented Jan 20, 2023 • edited Loading

mbriar commented Jan 20, 2023

llde commented Jan 20, 2023

mbriar commented Jan 20, 2023

llde commented Jan 20, 2023

mbriar commented Jan 20, 2023

llde commented Jan 20, 2023

mbriar commented Jan 20, 2023 • edited Loading

llde commented Jan 20, 2023

Blisto91 commented Jan 20, 2023 • edited Loading

doitsujin commented Jan 20, 2023

mbriar commented Jan 20, 2023

llde commented Jan 20, 2023

Blisto91 commented Jan 20, 2023

llde commented Jan 20, 2023

doitsujin commented Jan 21, 2023

doitsujin commented Jan 21, 2023

doitsujin commented Jan 21, 2023 • edited Loading

llde commented Jan 21, 2023

doitsujin commented Jan 21, 2023 • edited Loading

doitsujin commented Jan 21, 2023

llde commented Jan 21, 2023

K0bin commented Jan 24, 2023

mbriar commented Jan 24, 2023

Blisto91 commented Jan 24, 2023 • edited Loading

mbriar commented Jan 24, 2023

Blisto91 commented Jan 24, 2023

zzhiyi commented Jan 18, 2023 •

edited

Loading

llde commented Jan 20, 2023 •

edited

Loading

mbriar commented Jan 20, 2023 •

edited

Loading

doitsujin commented Jan 20, 2023 •

edited

Loading

llde commented Jan 20, 2023 •

edited

Loading

mbriar commented Jan 20, 2023 •

edited

Loading

mbriar commented Jan 20, 2023 •

edited

Loading

Blisto91 commented Jan 20, 2023 •

edited

Loading

doitsujin commented Jan 21, 2023 •

edited

Loading

doitsujin commented Jan 21, 2023 •

edited

Loading

Blisto91 commented Jan 24, 2023 •

edited

Loading