 No.18418
Tafoid Administrator
Jan 28, 2021, 19:55
|
As tested, it appears the regression happens here: https://github.com/mamedev/mame/commit/4aa6f2a3429868945510181c1415885d97852ca8 It is to be noted that there were acknowledged issues with this commit as the device was flagged as having "imperfect sound" but they have not been resolved or even looked at since that time. |
|---|---|
 No.21268
hap Developer
Apr 4, 2023, 00:31
edited on: Apr 4, 2023, 00:44
|
After the game over message, the audiocpu jumps to its reset routine. (also happens in 0.194) (...) C10FA6: rts C0B3F8: sf $d526.w C0B3FC: move.w #$2700, D0 C0B400: dc.w $a000; opcode 1010 <-- huh C10D8C: move.w D0, (A7) C10D8E: addq.l #2, ($2,A7) C10D92: rte C0B402: jmp $c1089e.l <-- jump to reset routine C1089E: moveq #$7f, D0 gunlocko doesn't do this, and the bug doesn't happen in that set. The audiocpu roms are the same as the newer sets. |
 No.21270
ywy Tester
Apr 4, 2023, 06:41
edited on: Apr 4, 2023, 06:42
|
maincpu side difference in the new sets is probably this, part of code insert $141a-14a3:001450 move.b #$ff, $c00488.l 001458 clr.w $c00480.l 00145E clr.w $c00482.l 001464 lea $c00000.l, A0 00146A move.w #$ff, D1 00146E moveq #$0, D0 001470 move.l D0, (A0)+ 001472 dbra D1, 70 001476 rts |