09313: vcop2: Music missing in beginner stage

ID	Category [?]	Severity [?]	Reproducibility	Date Submitted	Last Update
09313	Sound	Major	Always	1 day ago	17 hours ago

Tester	belegdol	View Status	Public	Platform	MAME (Self-compiled)
Assigned To		Resolution	Open	OS	Linux (64-bit)
Status [?]	Confirmed			Driver	sega/model2.cpp
Version	0.283	Fixed in Version		Build	x64
		Fixed in Git Commit		Github Pull Request #

Summary	09313: vcop2: Music missing in beginner stage
Description	As of 0.283, vcop2 lacks music during gameplay in the beginner stage. It works in 0.280 and is broken in 0.282. I was not able to bisect it any further yet.
Steps To Reproduce	1. Start vcop2 2. Insert coin 3. Start beginner stage
Additional Information
Github Commit

Flags
Regression Version
Affected Sets / Systems	vcop2

Attached Files

No.23857 PL1 Tester 1 day ago	v0.281 has music. Scott
No.23858 gm_matthew Tester 17 hours ago	I have identified the cause of this bug and am currently working on a fix. When the game wants to play a sound effect or music track, it runs a subroutine that writes the appropriate sound command to a buffer in memory, increments the buffer index and unmasks sound interrupts. When a sound interrupt is asserted, it reads the command from the buffer via the index and transmits it to the SCSP. The reason why this is not working properly is that the subroutine unmasks sound interrupts before incrementing the buffer index, but MAME 0.282 and later fires a sound interrupt the instant it is unmasked which results in the wrong command being transmitted because the buffer index has not been incremented yet. (Patching the game program so that it increments the buffer index first results in sound and music working as expected.) It seems that there must be a tiny delay between an interrupt being unmasked and the same interrupt being asserted, at least enough for the i960 to execute an additional write instruction.

No.23857

PL1

Tester

1 day ago

v0.281 has music.

Scott

No.23858

gm_matthew

Tester

17 hours ago

I have identified the cause of this bug and am currently working on a fix.

When the game wants to play a sound effect or music track, it runs a subroutine that writes the appropriate sound command to a buffer in memory, increments the buffer index and unmasks sound interrupts. When a sound interrupt is asserted, it reads the command from the buffer via the index and transmits it to the SCSP.

The reason why this is not working properly is that the subroutine unmasks sound interrupts before incrementing the buffer index, but MAME 0.282 and later fires a sound interrupt the instant it is unmasked which results in the wrong command being transmitted because the buffer index has not been incremented yet. (Patching the game program so that it increments the buffer index first results in sound and music working as expected.)

It seems that there must be a tiny delay between an interrupt being unmasked and the same interrupt being asserted, at least enough for the i960 to execute an additional write instruction.