(nw) The at.cpp driver remains a total and utter mess. It seems that at least the EISA and PCI boards should be moved out of this driver in the future.
* Fix issues with remaining GFDM games
* Fix DDR Solo Bass Mix
* Fix regressions with DSEM2, DDR Extreme, DDR Max 2
* Fix MP3 streaming and get rid of workaround to load full files
* Attempt to fix timing issues by emulating a buffer before the song actually plays.
This value was found by choosing something that had a tight but known timing window (the Konami logo sound in DM10) and finding an offset that didn't allow you to hear it loop before the song end command was sent.
* Revert changes for dynamic base address because I still can't figure out a more logical way that Konami could've implemented this.
* Use the last non-0 diff to fix some games
* Some small refactoring
* This should be the same as GF10DM9
* Use a block size that's more in line with what the real FPGA uses
* specpls3_flop.xml: New additions
Added side A of the LC-10 Colour Screen printer driver disk, dumped by Guy Bearman.
Also corrected a couple of entries after asking around and doing some additional research.
* Update specpls3_flop.xml
* Update specpls3_flop.xml
* Fix volume bug
* Implement decryption algorithm for DDR Solo Bass Mix which does not rely
on premade tables/keys.
* This shift is useless
* Make use of bitswap
315_5124.cpp : Use color lookup and correct color levels for 315_5313, Add notes
315_5313.cpp : Use color lookup, Convert bitmap drawing function into bitmap_rgb32, Reduce unused, Fix save pointers, Fix external palette handlers, Implement 3 bit color mode
mega32x.cpp : Use color lookup, Reduce unuseds, Add device_palette_interface
- Emulate strange but documented behavior of unaligned accesses and byte writes to internal peripheral registers
- Use more official register mnemonics
Move palette shadow / hilight toggle bit function into palette handler (excpect hangon; it has external shadow / hilight toggle bit), Simplify some handlers, Fix debugger issues, Fix spacing, Use shorter / correct type values
* Split into Indy (Guinness) and Indigo 2 (Full House) versions.
* Fixed second SCSI controller handling in Full House mode.
* Fixed serial EEPROM hookup. Indigo 2 now plays its boot tune.
-ioc2: Fixed System ID register handling. Indy and Indigo 2 are now properly identified. [Ryan Holtz]
