Dirk Hoffmann

Results 81 issues of Dirk Hoffmann

Until now, headless mode was only used to accelerate the emulator in warp mode. This worked well because warp mode is usually enabled only during load or save operations. Combining...

bug

There is a neat summary of the run-ahead technique here: https://bsnes.org/articles/input-run-ahead I think the run-ahead technique can be integrated into VirtualC64 with reasonable effort, and the emulator would benefit significantly...

enhancement

This is a prerequisite for #781. Implementing run-ahead requires the emulator to be aware of all state changes coming from outside (key presses, joystick movements, etc.). Once this is assured,...

enhancement

Hi, there is a new RESID version in VICE 3.8: - reSID: fixed a Noise Init problem - reSID: filter tweaks - reSID: The 6581 saw+pulse waveform was wrong -...

enhancement
v4.8

This code is part of `m68k_op_movec_32_cr` ``` case 0x802: /* CAAR */ if(CPU_TYPE_IS_EC020_PLUS(CPU_TYPE)) { REG_DA[(word2 >> 12) & 15] = REG_CAAR; return; } m68ki_exception_illegal(); break; ``` In the last line,...

This code fragment always truncates variable `outer` to 16 bit: ``` if(outer) { if(comma) strcat(mode, ","); strcat(mode, make_signed_hex_str_16(outer)); } strcat(mode, ")"); ``` I think it should be something like this:...

The disassembler produces movep.w ($ffff,A0), D0 To make it consistent with the other instructions, it should be changed to: movep.w ($-1,A0), D0

Under some circumstances, Musashi uses too large bit-shift values. This results in undefined behaviour. This is what the UndefinedBehaviorSanitizer (which is part of clang) reports: The problem with undefined behaviour...

I’ve started to run Toni Wilen’s 'cputester' in vAmiga. Toni's test `CHK.W` fails due to a wrong value of the N flag if an exception happens. Musashi sets the N...

According to {m68k_op_move16_32 , 0xfff8, 0xf620, { 0, 0, 0, 4}}, `move16_32` is only available on the 68040, but there is no CPU type check inside the execution handler: ```...