06499: xsleena and clones: Some enemies in the second stage has disappeared

ID	Category [?]	Severity [?]	Reproducibility	Date Submitted	Last Update
06499	Gameplay	Major	Always	Feb 9, 2017, 13:12	Mar 31, 2017, 18:04

Tester	StHiryu	View Status	Public	Platform	MAME (Unofficial Binary)
Assigned To	cuavas	Resolution	Fixed	OS	Windows Vista/7/8 (64-bit)
Status [?]	Resolved			Driver	technos/xain.cpp
Version	0.182	Fixed in Version	0.183	Build	64-bit
		Fixed in Git Commit		Github Pull Request #

Summary	06499: xsleena and clones: Some enemies in the second stage has disappeared
Description	According this post in the mameworld forum, there is a regression in Xain'd Sleena. The insects in stage 2 doesn't appear anymore: http://www.mameworld.info/ubbthreads/showflat.php?Cat=&Number=363359&page=0&view=collapsed&sb=5&o=&fpart=1&vc=1&new=1486652757#Post363343
Steps To Reproduce	Start to play, select stage 2 (Lagto Soa), and walk a little: No insects attack us.
Additional Information	Provide a video from previous mame version (179) to compare:
Github Commit

Flags
Regression Version	0.182
Affected Sets / Systems	xsleena and clones

Attached Files

No.13616 cuavas Administrator Feb 10, 2017, 00:37	This is interesting. The game enables the MCU timer interrupt (programmable mode, internal clock source, external input disabled, prescale = 2^7). Before moving to the 68705-with-peripherals, the MCU for this game wasn't receiving timer interrupts at all. Somehow the trigger that spawns these enemies is getting lost now, but I don't know how. I'll investigate it further at some point.
No.13617 cuavas Administrator Feb 10, 2017, 00:45	The timer interrupt handler is pretty simple, it operates on bytes at $21, $22 and $23: 1AB tst 3D 21 1AD beq BA 27 0B 1AF dec 3A 22 1B1 beq B6 27 03 1B3 bclr 7,TCR 1F 09 1B5 rti 80 1B6 inc 3C 23 1B8 bra B3 20 F9 1BA bclr 0,PORTB 11 01 1BC bset 0,PORTB 10 01 1BE bra B3 20 F3 Or in C++: void timer_handler() { if (!(unsigned char )0x21) { PORTB &= 0xfe; PORTB \|= 0x01; } else if (!--(unsigned char )0x22) { ++(unsigned char )0x23; } TCR &= 0x7f }
No.13618 cuavas Administrator Feb 10, 2017, 01:14	https://github.com/mamedev/mame/commit/9ae0bf812b54c2a66a2039389cd3f7cfa40851ee - it was a problem with spurious timer overflow interrupts being generated when TDR=0x00.

No.13616

cuavas

Administrator

Feb 10, 2017, 00:37

This is interesting. The game enables the MCU timer interrupt (programmable mode, internal clock source, external input disabled, prescale = 2^7). Before moving to the 68705-with-peripherals, the MCU for this game wasn't receiving timer interrupts at all. Somehow the trigger that spawns these enemies is getting lost now, but I don't know how. I'll investigate it further at some point.

No.13617

cuavas

Administrator

Feb 10, 2017, 00:45

The timer interrupt handler is pretty simple, it operates on bytes at $21, $22 and $23:

 1AB tst                                             3D 21
 1AD beq   BA                                         27 0B
 1AF dec                                             3A 22
 1B1 beq   B6                                         27 03
 1B3 bclr  7,TCR                                        1F 09
 1B5 rti                                                80
 1B6 inc                                             3C 23
 1B8 bra   B3                                         20 F9
 1BA bclr  0,PORTB                                      11 01
 1BC bset  0,PORTB                                      10 01
 1BE bra   B3                                         20 F3

Or in C++:

void timer_handler()
{
    if (!*(unsigned char *)0x21)
    {
        PORTB &= 0xfe;
        PORTB |= 0x01;
    }
    else  if (!--*(unsigned char *)0x22)
    {
        ++*(unsigned char *)0x23;
    }
    TCR &= 0x7f
}

No.13618

cuavas

Administrator

Feb 10, 2017, 01:14

https://github.com/mamedev/mame/commit/9ae0bf812b54c2a66a2039389cd3f7cfa40851ee - it was a problem with spurious timer overflow interrupts being generated when TDR=0x00.