04902: bladestl and clones: Flip Screen has no sprite flip

ID	Category [?]	Severity [?]	Reproducibility	Date Submitted	Last Update
04902	Flip Screen/Cocktail	Minor	Always	Jun 10, 2012, 23:17	Nov 7, 2014, 18:03

Tester	Tafoid	View Status	Public	Platform
Assigned To		Resolution	Open	OS
Status [?]	Confirmed			Driver	konami/bladestl.cpp
Version	0.146	Fixed in Version		Build
		Fixed in Git Commit		Github Pull Request #

Summary	04902: bladestl and clones: Flip Screen has no sprite flip
Description	Flip Screen has no sprite flip
Steps To Reproduce
Additional Information
Github Commit

Flags
Regression Version	0.135u4
Affected Sets / Systems	bladestl and clones

Attached Files	0000.png (10,432 bytes) Jun 10, 2012, 23:17 Uploaded by Tafoid Sprites misplaced and not flipped correctly

No.11232 Osso Moderator Nov 7, 2014, 08:56 edited on: Nov 7, 2014, 08:57	Before the code was originally device-ified and moved to the now defunct video/konicdev.c, the k007342 and the k007420 shared the same flipscreen related variable. Now that the two video chips are proper devices, only the k007342 reads the flipscreen bit correctly, the k007420 has no way to know it (m_flipscreen is always 0). This affects rockrage.c too.
No.11233 AWJ Developer Nov 7, 2014, 18:03	Based on the bladestl schematics, the way we've split the k007420 and k007342 functionality into two devices is wrong (and I'm in the process of rewriting them to fix this) They aren't a "sprite chip" and a "tilemap chip", they're more like the later Konami paired chipsets with 007420 being the "RAM controller" part (approximately equivalent to 0052109 or 051960) and 007342 being the "ROM controller" part (equivalent to 051962 or 051937) The VRAM address lines and /OE are connected to 007420, while the VRAM data lines, both sets of gfx ROMs (tiles and sprites) and the color PROMs are all connected to 007342. Both chips share the same address range (the same chip select signal goes to both), including the palette (007342 outputs the chip select for the palette RAM)

No.11232

Osso

Moderator

Nov 7, 2014, 08:56

edited on: Nov 7, 2014, 08:57

Before the code was originally device-ified and moved to the now defunct video/konicdev.c, the k007342 and the k007420 shared the same flipscreen related variable.
Now that the two video chips are proper devices, only the k007342 reads the flipscreen bit correctly, the k007420 has no way to know it (m_flipscreen is always 0).

This affects rockrage.c too.

No.11233

AWJ

Developer

Nov 7, 2014, 18:03

Based on the bladestl schematics, the way we've split the k007420 and k007342 functionality into two devices is wrong (and I'm in the process of rewriting them to fix this)

They aren't a "sprite chip" and a "tilemap chip", they're more like the later Konami paired chipsets with 007420 being the "RAM controller" part (approximately equivalent to 0052109 or 051960) and 007342 being the "ROM controller" part (equivalent to 051962 or 051937) The VRAM address lines and /OE are connected to 007420, while the VRAM data lines, both sets of gfx ROMs (tiles and sprites) and the color PROMs are all connected to 007342. Both chips share the same address range (the same chip select signal goes to both), including the palette (007342 outputs the chip select for the palette RAM)