[bug] DohnaDohna failed to start with wine8 but fine on x86 about box86 HOT 18 CLOSED

So, after some analysis, it's a new case of Obfuscation and Anti debug methods.

The Anti-debug use some Divide by 0 and classic Int 3. The Obfuscation as at least some new imul flag manipulation that I found a bit puzzling. Probably some other things too. I'll work on this one on both box86 and box64 (using Wine/Wow64).

from box86.

I bought another device for box86 test. It's not kunpeng

                                 shenmo128@armbian 
                                 ----------------- 
      █ █ █ █ █ █ █ █ █ █ █      OS: Armbian (23.02.0-trunk) aarch64 
     ███████████████████████     Host: M401A 
   ▄▄██                   ██▄▄   Kernel: 5.15.94-flippy-82+o 
   ▄▄██    ███████████    ██▄▄   Uptime: 1 hour, 20 mins 
   ▄▄██   ██         ██   ██▄▄   Packages: 1619 (dpkg) 
   ▄▄██   ██         ██   ██▄▄   Shell: bash 5.1.16 
   ▄▄██   ██         ██   ██▄▄   Resolution: 1920x969 
   ▄▄██   █████████████   ██▄▄   DE: MATE 1.26.0 
   ▄▄██   ██         ██   ██▄▄   WM: Metacity (Marco) 
   ▄▄██   ██         ██   ██▄▄   Theme: Adwaita [GTK2/3] 
   ▄▄██   ██         ██   ██▄▄   Icons: hazy-color-blossom [GTK2/3] 
   ▄▄██                   ██▄▄   Terminal: mate-terminal 
     ███████████████████████     Terminal Font: Noto Sans Mono CJK SC 13 
      █ █ █ █ █ █ █ █ █ █ █      CPU: ARMv8 rev 4 (v8l) (4) @ 1.704GHz 
                                 Memory: 578MiB / 1953MiB 

from box86.

I bought another device for box86 test. It's not kunpeng

                                 shenmo128@armbian 
                                 ----------------- 
      █ █ █ █ █ █ █ █ █ █ █      OS: Armbian (23.02.0-trunk) aarch64 
     ███████████████████████     Host: M401A 
   ▄▄██                   ██▄▄   Kernel: 5.15.94-flippy-82+o 
   ▄▄██    ███████████    ██▄▄   Uptime: 1 hour, 20 mins 
   ▄▄██   ██         ██   ██▄▄   Packages: 1619 (dpkg) 
   ▄▄██   ██         ██   ██▄▄   Shell: bash 5.1.16 
   ▄▄██   ██         ██   ██▄▄   Resolution: 1920x969 
   ▄▄██   █████████████   ██▄▄   DE: MATE 1.26.0 
   ▄▄██   ██         ██   ██▄▄   WM: Metacity (Marco) 
   ▄▄██   ██         ██   ██▄▄   Theme: Adwaita [GTK2/3] 
   ▄▄██   ██         ██   ██▄▄   Icons: hazy-color-blossom [GTK2/3] 
   ▄▄██                   ██▄▄   Terminal: mate-terminal 
     ███████████████████████     Terminal Font: Noto Sans Mono CJK SC 13 
      █ █ █ █ █ █ █ █ █ █ █      CPU: ARMv8 rev 4 (v8l) (4) @ 1.704GHz 
                                 Memory: 578MiB / 1953MiB 

While it runs well on exagear. Just feel this info might be help. Exagear-huawei's graphic performance sucks.

Since exagear runs well, i doubt whether this problem is caused by lack of libs? Don't know.....

from box86.

The memory error are "normal". Box64 warns you with

BOX64: Didn't detect 48bits of address space, considering it's 39bits

And wine expect 48bits of address space. It should workaround the unavable memory tho.

I'll try the game on my side to see what is happening, thanks for the link.

from box86.

Is it better now with latest box86?

from box86.

Is it better now with latest box86?

Roger. Compiling...

from box86.

Is it better now with latest box86?

[100%] Building C object CMakeFiles/box86.dir/src/wrapped/wrappedturbojpeg.c.o
[100%] Linking C executable box86
[100%] Built target box86
shenmo128@armbian:~/Desktop/box86/build$ ./box86
./box86: /lib/arm-linux-gnueabihf/libm.so.6: version GLIBC_2.35' not found (required by ./box86) ./box86: /lib/arm-linux-gnueabihf/libm.so.6: version GLIBC_2.29' not found (required by ./box86)
./box86: /lib/arm-linux-gnueabihf/libc.so.6: version GLIBC_2.33' not found (required by ./box86) ./box86: /lib/arm-linux-gnueabihf/libc.so.6: version GLIBC_2.32' not found (required by ./box86)
./box86: /lib/arm-linux-gnueabihf/libc.so.6: version `GLIBC_2.34' not found (required by ./box86)

I use this command to compile.
for Other ARM64 64bits Linux platform

mkdir build; cd build; cmake .. -DARM64=1 -DCMAKE_BUILD_TYPE=RelWithDebInfo; make -j3

You can also add -DBAD_SIGNAL=ON to the cmake command if you are on Linux Kernel mixed with Android, like on RK3588 or maybe Termux

You'll need an armhf multiarch environment, and an armhf gcc: On debian, install it with sudo apt install gcc-arm-linux-gnueabihf.

Also, on armbian, you may need to install libc6-dev-armhf-cross or you may have an issue with crt1.o and a few other files not included with box86.

Should i use

for Other ARM Linux platforms

mkdir build; cd build; cmake .. -DARM_DYNAREC=ON -DCMAKE_BUILD_TYPE=RelWithDebInfo; make -j3

You can also add -DBAD_SIGNAL=ON to the cmake command if you are on Linux Kernel mixed with Android, like on RK3588 or maybe Termux

from box86.

I tried the Other ARM Linux platforms command and get this

shenmo128@armbian:~/Desktop/box86/build$ make
[  0%] Building C object CMakeFiles/dynarec_arm.dir/src/dynarec/dynablock.c.o
cc: error: unrecognized command-line option ‘-marm’
make[2]: *** [CMakeFiles/dynarec_arm.dir/build.make:76: CMakeFiles/dynarec_arm.dir/src/dynarec/dynablock.c.o] Error 1
make[1]: *** [CMakeFiles/Makefile2:154: CMakeFiles/dynarec_arm.dir/all] Error 2
make: *** [Makefile:166: all] Error 2

from box86.

Fine i will wait until https://github.com/ryanfortner/box86-debs/tree/master/debian goes to the latest commit

from box86.

Successfully launched with the latest box86 commit build. But blank screen.

Testing box64 and wine wow64

from box86.

Yeah, I have that on my side too, but wasn't sure if it was a box86 issue or a driver issue.
I does work fine with box64 and wine/wow64 on my side (I mean, I mas fixes for it to work, it wasn't initialy)

from box86.

Yeah, I have that on my side too, but wasn't sure if it was a box86 issue or a driver issue. I does work fine with box64 and wine/wow64 on my side (I mean, I mas fixes for it to work, it wasn't initialy)

Same. Box86 failed but box64+wow64 fine.

Anyway, thank you a lot for your effort!

from box86.

I'll continue to work on this. It should work with box86, especialy seeing it working with box64!

from box86.

It seems to work now, with latest box86 version. Can you try again @shenmo7192 ?

from box86.

It seems to work now, with latest box86 version. Can you try again @shenmo7192 ?

So sorry about that, I'm very busy until next year Jan. I'm a college student and I have to prepare for final exam.....

Will try it after the exam

from box86.

No problem. Prepare your exams, I'll be there in january, don't worry!

from box86.

No problem. Prepare your exams, I'll be there in january, don't worry!

I tried and saw the window. My arm32 device's performance is very low and lack of audio output, so i think i just cannot launch it on my device. Well on my kunpeng and box64 it runs well. I think you fixed it.

from box86.

Did you start a game? On my side, it crash, but I don't know if that's expected with that copy or if it's something else.

from box86.