Metal Gear Acid [ULUS10006] crash during level exit

[i30817]

Seems to be repeatable on this computer if moving to the 'second' target circle on the save state provided. Which card or direction used doesn't seem to matter. The first circle works.

The log says:
10:19:017 idle0 I[COMMON]: /home/user/Documents/Netbeans_projects/ppsspp/Common/ChunkFile.h:600 ChunkReader: Done loading /home/user/.ppsspp/PSP/PPSSPP_STATE/ULUS10006_1.00_0.ppst
10:19:032 Main I[SCEGE]: /home/user/Documents/Netbeans_projects/ppsspp/GPU/GLES/Framebuffer.cpp:748 Creating FBO for 00044000 : 480 x 272 x 1
10:19:214 Main I[SCEGE]: /home/user/Documents/Netbeans_projects/ppsspp/GPU/GLES/Framebuffer.cpp:748 Creating FBO for 00000000 : 480 x 272 x 1
10:19:214 Main W[SCEGE]: /home/user/Documents/Netbeans_projects/ppsspp/GPU/GLES/Framebuffer.cpp:758 FBO sharing existing depthbuffer (unsupported), 00000000/00088000 and 00044000/00088000
Segmentation fault (core dumped)

The savestate:
http://speedy.sh/9q7W7/ULUS10006-1.00-0.ppst.zip

[sum2012]

Can upload the savestate to google drive ?

[i30817]

Good idea, that's much better than those other upload services.
But has the project some specialized account for it or just use my own?

[i30817]

Well anyway, here is a google drive link, copy it over to your account if you have one
https://drive.google.com/file/d/0BzxkBmaf1EiWeXFFaUljUWYzZmc/edit?usp=sharing

[sum2012]

Do you mean this circle ?

No probrem on me

[i30817]

yes that one... crashes 100% of the time for me, on linux amd64 on the Gallium 0.4 on AMD RV710 driver

[i30817]

Doesn't crash with 'fast memory' off

[sum2012]

Can close issue ?

Doesn't crash with 'fast memory' off

[sum2012]

I can reproduce with fast memory on

[sum2012]

When I do not enable JIT.It fixed.
@unknownbrackets can you do something fix ?

[i30817]

This game has a built in level save function, to make sure that it was nothing from defective savestate save i saved with it, closed ppsspp and loaded the save with 'continue' (with fast memory on) and it still crashed. I took the opportunity to turn off all the optimizations except fast memory (all threading, the atomic audio clocks - no way to turn off JIT on the sdl port) and the log before the crash was slightly different, referring to atrac

18:43:679 sd_bgm I[ME]: /home/user/Documents/Netbeans_projects/ppsspp/Core/HLE/sceAtrac.cpp:1268 0=sceAtracSetDataAndGetID(09e8f700, 0009883c)
18:43:679 sd_bgm W[ME]: /home/user/Documents/Netbeans_projects/ppsspp/Core/HLE/sceAtrac.cpp:1158 This is an atrac3+ stereo audio
Segmentation fault (core dumped)

[unknownbrackets]

Does this still happen in the latest git build? As mentioned by @i30817 it would be best to not use the savestate to verify this, it might have saved the bug's state.

-[Unknown]

[i30817]

It still crashes. I tried to make a debug build and backtrace it but i didn't get a usuable one? Am i doing something wrong here:

cd ~/Documents/Netbeans_projects/ppsspp/ && git pull && git submodule update --init && cmake -DCMAKE_BUILD_TYPE=Debug . && make && gdb ./PPSSPPSDL
[...]
GNU gdb (GDB) 7.6.50.20131218-cvs-ubuntu
Copyright (C) 2013 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later http://gnu.org/licenses/gpl.html
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law. Type "show copying"
and "show warranty" for details.
This GDB was configured as "x86_64-linux-gnu".
Type "show configuration" for configuration details.
For bug reporting instructions, please see:
http://www.gnu.org/software/gdb/bugs/.
Find the GDB manual and other documentation resources online at:
http://www.gnu.org/software/gdb/documentation/.
For help, type "help".
Type "apropos word" to search for commands related to "word"...
Reading symbols from ./PPSSPPSDL...done.
(gdb) run --nofork
Starting program: /home/i30817/Documents/Netbeans_projects/ppsspp/PPSSPPSDL --nofork
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
[New Thread 0x7fffec958700 (LWP 21402)]
[Thread 0x7fffec958700 (LWP 21402) exited]
[New Thread 0x7fffec958700 (LWP 21403)]
OpenGL 2.0 or higher.
D: /home/i30817/Documents/Netbeans_projects/ppsspp/native/file/zip_read.cpp:308: D: Registered VFS for prefix : /home/i30817/Documents/Netbeans_projects/ppsspp/assets/
D: /home/i30817/Documents/Netbeans_projects/ppsspp/native/file/zip_read.cpp:308: D: Registered VFS for prefix : /home/i30817/Documents/Netbeans_projects/ppsspp/
D: /home/i30817/Documents/Netbeans_projects/ppsspp/native/file/zip_read.cpp:308: D: Registered VFS for prefix : /home/i30817/
[New Thread 0x7fffed82b700 (LWP 21404)]
Pixels: 960 x 544
Virtual pixels: 960 x 544
I: /home/i30817/Documents/Netbeans_projects/ppsspp/native/gfx_es2/gl_state.cpp:127: I: GPU Vendor : X.Org ; GL version str: 3.0 Mesa 10.0.1
I: /home/i30817/Documents/Netbeans_projects/ppsspp/native/gfx/texture.cpp:283: I: ZIM loaded: 1 levels
I: /home/i30817/Documents/Netbeans_projects/ppsspp/native/gfx/texture.cpp:314: I: Gen-binding texture
[New Thread 0x7fffe7fff700 (LWP 21405)]
[New Thread 0x7fffe77fe700 (LWP 21406)]
[New Thread 0x7fffd8cfe700 (LWP 21407)]

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x7fffe77fe700 (LWP 21406)]
0x000000004035a34a in ?? ()
(gdb) bt
#0 0x000000004035a34a in ?? ()
#1 0x00007fffe77fe700 in ?? ()
#2 0x00007fffe77fe700 in ?? ()
#3 0x00007fffe77fe9c0 in ?? ()
#4 0x0000000000000000 in ?? ()
(gdb) quit

[i30817]

oh, the backtrace format linked to bugs 1-4, excuse me.

[unknownbrackets]

Heh. Well, if it's crashing in jit, then it will provide a pretty useless stack trace. I would recommend:

• Make sure you're compiling in debug mode, e.g. with -g. I assume DCMAKE_BUILD_TYPE=Debug does that, but I don't really know for sure, might use -v just to double check.
• Make sure to turn off fast memory, which should give you a better stack trace.

-[Unknown]

[i30817]

-v (version i assume) doesn't output anything special on PPSSPPSDL, it just starts it, so i don't have any indication this is a debug compile.
I assume that bug what sum012 meant is that with fast memory on, and jit disabled the crash doesn't occur - because it does occur with fast memory off and jit on. However, i'm been unable to make the game actually load without jit and with fast memory on so i have no idea.

edit: oh wait, it runs if i turn multithreading off. And indeed, with fast memory on and JIT off, no crash.

[unknownbrackets]

Sorry, I meant cmake -v I think.

With jit disabled, fast memory doesn't matter. Fast memory only applies with jit on.

Does this still crash?

-[Unknown]

[sum2012]

#5365 appear to fixed in v0.9.9.1-1505-gcf577e9
This one may also fix.
But need someone testing

[i30817]

Yes, it seems fixed. If you want to test it yourself, load the savestate linked in my google drive above and save with the ingame save option from the select menu, then quit, reload the game and 'continue'.
Fast memory and JIT enabled, no crash.
I didn't perform a regression test to see what fixed it.

[unknownbrackets]

Glad it's fixed either way. I'll mark this as closed.

-[Unknown]