https://bugs.winehq.org/show_bug.cgi?id=42172
--- Comment #4 from Dmitriy xomachiner@gmail.com --- Created attachment 57379 --> https://bugs.winehq.org/attachment.cgi?id=57379 Workaround dump for paladins.exe
I've found a workaround to start the paladins. The attached dump from objdump describes which opcodes should be placed in Paladins.exe to avoid the problem. It's also recommended to replace PaladinsEAC.exe with modified copy of Paladins.exe. To modify the binary one may use any hex editor (e.g. Bless).
I've also obtained some additional information about the crash. Debugging session has shown that address in %ecx is being forming by XORing of two values: one of them located in memory at (*FS:0x2C)+4*(*0x3D18240), and the other may be found at address (*0x2273AF8)+0xBE0. At the crash conditions, the first value equals 0 somehow and %ecx contains only the second value. Most intriguing part of this is the first value becomes correct after the exception had been catched by debugger. I don't have any explanation of such behaviour except only maybe some kind of the race condition.