8000 Danball Senki BOOST crash / graphical glitches · Issue #5918 · hrydgard/ppsspp · GitHub
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content

Danball Senki BOOST crash / graphical glitches #5918

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.

Sign up for GitHub

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

Jump to bottom
Open
Salijour opened this issue Apr 19, 2014 · 90 comments
Open

Danball Senki BOOST crash / graphical glitches #5918

Salijour opened this issue Apr 19, 2014 · 90 comments
Labels
GE emulation Backend-independent GPU issues
Milestone
Future

Comments

@Salijour
Copy link

Encountered two crashes during battles. Screen captured the log.

Fast memory off and no hacks.

b30aa646ce237736713af1c428e9d7af
@sum2012
Copy link
Collaborator
sum2012 commented Apr 19, 2014

How to start the war ?
I pressed X,O
1

@Salijour
Copy link
Author

What? No. About every 1 in 25 battles it will randomly crash for me in the battle never at the sortie screen.

@unknownbrackets
Copy link
Collaborator

@Salijour I think he's asking how to get from the sortie screen into battle. Maybe it's SELECT?

-[Unknown]

@Salijour
Copy link
Author

Ah. Sorry. The game pulls the confirmation button from firmware. So it is default to X in ppsspp.

@sum2012
Copy link
Collaborator
sum2012 commented Apr 22, 2014

I don't have problem.Just picture have some problem
Log will provide later.(my github has some problem)
2

@sum2012
Copy link
Collaborator
sum2012 commented Apr 26, 2014

v0.9.8-524-gfb5eda2 log
https://gist.github.com/sum2012/910e413363cfab31c016

@unknownbrackets
Copy link
Collaborator

Is the graphical glitch at least gone? Does this error above happen only with multithreading or something?

-[Unknown]

@Salijour
Copy link
Author

Nope, but it happens a lot less though. I didn't have any hacks or unstable/experimental options on. Never used multithreading on PPSSPP, since I don't need it in the games I test, I have to worry more about ppsspp not using enough resources and tripping my CPU in and out of idle mode. I did have a single crash right before a "super melee attack" from the enemy, but couldn't replicate it again after playing for roughly 5 hours. What I did notice from the log is that FBO reusing depthbuffer appeared when it messed up. Screenshots from log: http://i.gyazo.com/b6890a485d5f1d8dd79ea8de7792230c.png http://i.gyazo.com/98c1705e480bf0b11320218fd9b1c4e8.png

@unknownbrackets
Copy link
Collaborator

You can use "force real time sync" to keep your cpu at full power.

-[Unknown]

@Salijour
Copy link
Author

Ah. I wondered what that option did exactly. Anyway, I forgot to mention that I was using 0.9.8-1241, if that matters.

@Alphz
Copy link
Alphz commented Jun 20, 2014

whenever I used machine at arcade game (masked ranger mission) suddenly crash. what should I do? im using 0.9.8-1241 with no hack turn on. please help me..

@Alphz
Copy link
Alphz commented Jun 20, 2014

whenever I used a vending machine at arcade (masked ranger mission) game suddenly crash.im using 0.9.8-1241 with no hack turn on. is anyone have same problem ? what build run danball senki boost well?? sorry for double post

@sum2012
Copy link
Collaborator
sum2012 commented Jun 21, 2014

@Alphz Can you provide save staus ? Thanks

@Alphz
Copy link
Alphz commented Jun 25, 2014

Ohh im sorry...
game crash because i used patched game, when i played non patched danball senki boost (jp) game run normally, i'm still using 0.9.8-1241 version... maybe pached game is the problem not the emu.

@sum2012
Copy link
Collaborator
sum2012 commented Jun 25, 2014

Do patched game do not crash in real psp ?

@Alphz
Copy link
Alphz commented Jun 26, 2014

I'm using patched en ver iso from nicoblog and I dont have real psp. im sorry I cant provide the link to iso. but its easy to search by google. if my post violate rules please erase or edit.

@hrydgard
Copy link
Owner

Closing until this has been repro'd with a patched iso that works on the real PSP, otherwise there's no way to be sure if there's really a bug here or the patch is broken.

@sum2012
Copy link
Collaborator
sum2012 commented Jun 26, 2014

I confirm patched iso works on the real PSP.
PPSSPP first time freeze on war,next time crash after war
But I cannot reproduce in save status\real save
PPSSPP v0.9.8-1408-g3d61028 crash log:
https://gist.github.com/sum2012/87224594d8627ec9a58c

@hrydgard hrydgard reopened this Jun 26, 2014
@sum2012
Copy link
Collaborator
sum2012 commented Jun 26, 2014

PPSSPP v0.9.8-1425-ge73fcf4 crash stace
PPSSPPDebug64.exe!ConvertBGRA8888ToRGBA5551(unsigned short * dst, const unsigned int * src, const unsigned int numPixels) Line 377 C++
PPSSPPDebug64.exe!ConvertFromRGBA8888(unsigned char * dst, const unsigned char * src, unsigned int stride, unsigned int width, unsigned int height, GEBufferFormat format) Line 1651 C++
PPSSPPDebug64.exe!FramebufferManager::PackFramebufferSync_(VirtualFramebuffer * vfb, int x, int y, int w, int h) Line 1922 C++
PPSSPPDebug64.exe!FramebufferManager::ReadFramebufferToMemory(VirtualFramebuffer * vfb, bool sync, int x, int y, int w, int h) Line 1503 C++
PPSSPPDebug64.exe!FramebufferManager::NotifyBlockTransferBefore(unsigned int dstBasePtr, int dstStride, int dstX, int dstY, unsigned int srcBasePtr, int srcStride, int srcX, int srcY, int width, int height, int bpp) Line 2356 C++
PPSSPPDebug64.exe!GLES_GPU::DoBlockTransfer() Line 1939 C++

PPSSPPDebug64.exe!GLES_GPU::Execute_BlockTransferStart(unsigned int op, unsigned int diff) Line 1328 C++
PPSSPPDebug64.exe!GLES_GPU::FastRunLoop(DisplayList & list) Line 658 C++
PPSSPPDebug64.exe!GPUCommon::InterpretList(DisplayList & list) Line 516 C++
PPSSPPDebug64.exe!GPUCommon::ProcessDLQueueInternal() Line 686 C++
PPSSPPDebug64.exe!GPUCommon::ProcessEvent(GPUEvent ev) Line 643 C++
PPSSPPDebug64.exe!GLES_GPU::ProcessEvent(GPUEvent ev) Line 695 C++
PPSSPPDebug64.exe!ThreadEventQueue<GPUInterface,GPUEvent,enum GPUEventType,0,8,7>::RunEventsUntil(unsigned __int64 globalticks) Line 96 C++
PPSSPPDebug64.exe!ThreadEventQueue<GPUInterface,GPUEvent,enum GPUEventType,0,8,7>::ScheduleEvent(GPUEvent ev) Line 48 C++
PPSSPPDebug64.exe!GPUCommon::ProcessDLQueue() Line 666 C++
PPSSPPDebug64.exe!GPUCommon::UpdateStall(int listid, unsigned int newstall) Line 341 C++
PPSSPPDebug64.exe!sceGeListUpdateStallAddr(unsigned int displayListID, unsigned int stallAddress) Line 477 C++
PPSSPPDebug64.exe!WrapI_UU<&sceGeListUpdateStallAddr>() Line 194 C++
PPSSPPDebug64.exe!CallSyscallWithoutFlags(const HLEFunction * info) Line 472 C++
[External Code]
1
2
3
4
5
6
7

@unknownbrackets
Copy link
Collaborator

Try:

        if (convert) {
            int dstByteOffset = y * vfb->fb_stride * dstBpp;
            h = std::min(h, (int)vfb->height);
            ConvertFromRGBA8888(Memory::GetPointer(fb_address + dstByteOffset), packed + byteOffset, vfb->fb_stride, vfb->width, h, vfb->format);
        }

I guess this may mean multiple framebuffers downloaded in the same block transfer.

-[Unknown]

@sum2012
Copy link
Collaborator
sum2012 commented Jun 26, 2014

Play 2 wars,no crash
but has graphical glitch
1

@unknownbrackets
Copy link
Collaborator

I wonder if it's better to skip the copy entirely in that situation. It seems like it means the framebuffer is probably old.

That would be:

        if (convert) {
            int dstByteOffset = y * vfb->fb_stride * dstBpp;
            if (h < (int)vfb->bufferHeight) {
                ConvertFromRGBA8888(Memory::GetPointer(fb_address + dstByteOffset), packed + byteOffset, vfb->fb_stride, vfb->width, h, vfb->format);
            }
        }

(this is not the best way, but should suffice as a test.)

-[Unknown]

@sum2012
Copy link
Collaborator
sum2012 commented Jun 27, 2014

crash without crash strace
1

@Bigpet
Copy link
Collaborator
Bigpet commented Jun 27, 2014

@sum2012 have you checked other threads, also is that the stack frame of the main-thread?

@sum2012
Copy link
Collaborator
sum2012 commented Jun 27, 2014

I forget diasble fast memory,I redo the test

@sum2012
Copy link
Collaborator
sum2012 commented Jun 27, 2014

This time no crash but in second stage have graphical glitch
2

@unknownbrackets unknownbrackets mentioned this issue Jun 28, 2014
Merged
@unknownbrackets
Copy link
Collaborator

Do any of the crashes here still happen? How's the graphical glitch?

-[Unknown]

@hrydgard hrydgard added the GE emulation Backend-independent GPU issues label Dec 13, 2020
@nassau-tk
Copy link
Contributor

@sum2012
Really?
I didn't check it.
But,These backends are same result.
I didn't inclued specia lmeaning.

@sum2012
Copy link
Collaborator
sum2012 commented Dec 14, 2020

I don't have the word problem in opengl and vulkan
Windows 7 64 bit NVIDIA GeForce GT 710
1

@nassau-tk
Copy link
Contributor
nassau-tk commented Dec 14, 2020 •
edited
Loading

@sum2012

Yes,It has no problem me too.

Text glitch will happen in the battle scean.
And, The glitch will appear at enemy's name position.

And, Most important things is ...
Windows will not happen.
It will happen only Android .(Maybe.)

Issue point 1's chara is meaning equal to player character graphics.Not word.

Issue point 2 is meaning word.

2020_12_15_072409

Sometime appear right.
2020_12_15_074247

@sum2012
Copy link
Collaborator
sum2012 commented Dec 15, 2020

My phone Black shark 3 also don't have problem
Re-test in v1.10.3-1330-g73da378ef
The blur screen is fixed in windows and android

Screenshot_2020-12-15-20-04-14-173_org ppsspp ppsspp

Screenshot_2020-12-15-20-04-56-661_org ppsspp ppsspp

@sum2012
Copy link
Collaborator
sum2012 commented Dec 15, 2020

I play 2 wars,no graphical glitch now.
Wait @nassau-tk final confirm

@nassau-tk
Copy link
Contributor

@sum2012

Sorry,Not "PAUSE".
Push Triangle to enter the main menu mode then blur should working.
But,not working.
Plus,player characters is lost on the screen.
If exit menu then player characters return.

And,Enemy's name is not glitch in the battle screen?
Please,show me the Screenshot.
Maybe I can reproduce the glitch in the 3 times.

@sum2012
Copy link
Collaborator
sum2012 commented Dec 15, 2020

@nassau-tk yes,Push Triangle to enter the main menu mode then blur should working.
But,not working.
Enemy's name is not glitch in the battle screen
1

@sum2012
Copy link
Collaborator
sum2012 commented Dec 15, 2020

Enemy's name is glitch in the battle screen in my phone Black Shark 3 :(

@nassau-tk
Copy link
Contributor

Um...
Why Enemy's name glitch not happen on @sum2012 phone?
鮫 F438 三 is special phone?:smile:

Screenshot_20201216-075618_PPSSPP
My Galaxy S20 above.

@nassau-tk
Copy link
Contributor

敵の名前は私の携帯電話の戦闘画面の不具合ですBlackShark 3 :(

???
The glitch happend on your phone?
Or not happen?

@sum2012
Copy link
Collaborator
sum2012 commented Dec 15, 2020

The okay picture is in my computer GeForce GT 710

@nassau-tk
Copy link
Contributor
nassau-tk commented Dec 15, 2020 •
edited
Loading

The okay picture is in my computer GeForce GT 710

"Everytime no text glitch on Black Shark 3."
You want said this?

@sum2012
Copy link
Collaborator
sum2012 commented Dec 15, 2020 via email

@nassau-tk
Copy link
Contributor

OK! I understood it.

Ah...I don't have enough knowledge.
Regrettable!:dizzy_face:

@nassau-tk
Copy link
Contributor
nassau-tk commented Feb 1, 2021 •
edited
Loading

[Issue No.1]
Target OS: Windwos & Android both
If playable screen go to background then it should work blur.
But, PPSSPP is not became it.
Plus, Graphic is missing. (example: Player chara / Save Robot etc...)

[Issue No.2]
Target OS: Only Android
The enemys and supportes Indication space of the name has glitch to in battle scene.
This issue happen only Android device.
I have never seen in Windows.
And, There is following feature.

  1. Task kill PPSSPP on Android device.
  2. Boot PPSSPP and start the DanballSenki.
  3. Start the battle of 1st time on boot.(This time will appear the name correctly.)
  4. Continue the next battle = 2nd times.(This time will broken the name appearance)
  5. Continue the next battle = 3rd times.(This time will more broken the name appearance)

Exit game and return to PPSSPP menu and re-start the game, but it's not appear the name correctly. (Not Reset.)
It's need "Task kill" for reset appear the name correctly on Android.
Is this issue will happen by "Problem about some memory control"?

Movement is different on Windows & Android.
but, I can't understand to happen at a high-end device like Galaxy S20.
Glaxy S20 Snapdragon865/RAM 12GB

@unknownbrackets
Copy link
Collaborator

Does the Android only behavior per chance replicate on Windows under Direct3D 9? That API is more limited so just trying to see if it also has the problem which might help narrow it down.

-[Unknown]

@nassau-tk
Copy link
Contributor

Wao....D3D9 has other problem of other drawing.

GEdump is here.
It will seems "a little graphic missing" on D3D9.
ULJS00361_2files.zip
But, Real view is below!!

[Vulkan / Drawing correctly]
image

[D3D9 / Not drawing correctly]
image

[Vulkan / Drawing correctly]
image

[D3D9 / Not drawing correctly]
image

[Vulkan / Drawing correctly]
image

[D3D9 / Not drawing correctly / But Name is appear correctly]
image

@Panderner
Copy link
Contributor

I don't think depth copying does not support for D3D9?

@nassau-tk
Copy link
Contributor

Maybe, The "name appearance glitch" will be appearance correctly by turn off to simulate block transfer on Android.
Probably? Maybe? Parhaps?

I will try to check my other Android device too.

NOTICE: Windows10 is not need to turn off the simulate block transfer.

@nassau-tk
Copy link
Contributor
nassau-tk commented Feb 2, 2021 •
edited
Loading

I tested for "Name appearance glitch issue" by 2 Android device.

  • Galaxy S20 SD865/RAM 12GB
  • Xperia XZ1 Compact SD835/RAM 4GB

Backend : Vulkan
Mode : Buffered Rendering
Simulate Transfer Effects : Off

[Galaxy S20]
It will not happen the glitch at continued 5 times battle.
And, If used "fast foward" but it's not happen the glitch too.

[Xperai XZ1 Compact]
It will not happen the glitch at continued 5 times battle.
But, If used "fast foward" then it's happen the glitch.

I think It will related "hardware performance" and "simulate block transfer" both.

And, This issue (#13338 (comment)) will happen to easy by "fast forward" on both device.

An amature's idea below.
I think some buffer is not enough all scene for this game.

@sum2012
Copy link
Collaborator
sum2012 commented Feb 11, 2021

GPU driver Test fail in mobile phone while desktop computer success.
That is explain why only mobile phone have graphic bug.
I will create another issue for that

@sum2012
Copy link
Collaborator
sum2012 commented Feb 11, 2021 •
edited
Loading

@nassau-tk
For this screen
Screenshot_2021-02-11-10-45-30-581_org ppsspp ppsspp

Please test your phone
Disable hardware transform
(I magicly find this fix )

edit:I didn't read above post,disable Simulate block transfer effects also fix

@nassau-tk
Copy link
Contributor

@sum2012

Yes, That's right!
"Danball Senki & Danball Senki Boost" needs turn off the Simulate block transfer effects on Android.
But, If user using "High-Spec Android device" but it will need too.

But, Windows PC is not need to turn of the Simulate block transfer effects.

Is it general difference?
(Like a "Tactics Ogre -Unmei no Wa-".)

@sum2012
Copy link
Collaborator
sum2012 commented Feb 11, 2021

@nassau-tk Sorry,I mean in OpenGL ,turn on Simulate block transfer effect and turn off hardware transform

@nassau-tk
Copy link
Contributor

@nassau-tk Sorry,I mean in OpenGL ,turn on Simulate block transfer effect and turn off hardware transform

Yes, l tested above setting on Galaxy S20.
It will happen the name appearance issue on OpenGL & Vulkan.

OpenGL
Screenshot_20210211-192904_PPSSPP

Vulkan
Screenshot_20210211-192720_PPSSPP

@sum2012
Copy link
Collaborator
sum2012 commented Feb 12, 2021

In the playsotre comment,someone said
Danball Senki W black screen at the end
Screenshot_2021-02-12-13-10-15-134_com android vending

@hrydgard hrydgard modified the milestones: v1.12.0, Future Aug 21, 2021
@ghost ghost mentioned this issue Apr 16, 2022
Open
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
GE emulation Backend-independent GPU issues
Projects
None yet
Milestone
Future
Development

No branches or pull requests
10 participants
@hrydgard @unknownbrackets @Bigpet @sum2012 @daniel229 @Salijour @Alphz @Panderner @akuda @nassau-tk
0