Max/SameBoy
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Fix several potential regressions

Browse Source
This commit is contained in:
Lior Halphon 2022-05-04 19:30:29 +03:00
parent 36e2896ec7
commit 69a5ed3396
1 changed files with 3 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
Core/sm83_cpu.c
View File

@ -145,6 +145,7 @@ static void cycle_write(GB_gameboy_t *gb, uint16_t addr, uint8_t value)
/* The DMG STAT-write bug is basically the STAT register being read as FF for a single T-cycle */
case GB_CONFLICT_STAT_DMG:
GB_advance_cycles(gb, gb->pending_cycles);
GB_display_sync(gb);
/* State 7 is the edge between HBlank and OAM mode, and it behaves a bit weird.
The OAM interrupt seems to be blocked by HBlank interrupts in that case, despite
the timing not making much sense for that.
@ -206,7 +207,7 @@ static void cycle_write(GB_gameboy_t *gb, uint16_t addr, uint8_t value)
uint8_t old_value = GB_read_memory(gb, addr);
GB_advance_cycles(gb, gb->pending_cycles - 2);
GB_display_sync(gb);
if (gb->model != GB_MODEL_MGB && gb->position_in_line == 0 && (old_value & 2) && !(value & 2)) {
old_value &= ~2;
}
@ -277,6 +278,7 @@ static void cycle_write(GB_gameboy_t *gb, uint16_t addr, uint8_t value)
GB_advance_cycles(gb, gb->pending_cycles);
if (gb->model <= GB_MODEL_CGB_C) {
// TODO: Double speed mode? This logic is also a bit weird, it needs more tests
GB_apu_run(gb, true);
if (gb->apu.square_sweep_calculate_countdown > 3 && gb->apu.enable_zombie_calculate_stepping) {
gb->apu.square_sweep_calculate_countdown -= 2;
}