Replaced drw80pkr with older dump from [Team Europe].
drw80pkr:
- Added various graphics improvements and corrected colors. Game boots much farther and cleaner.
Out of credits, I also added:
* save states to adp.c (because part of the struct was already there), but nothing is working => no flag;
* save states to albazc.c, but I don't want credit, since no variable needed to be saved (i.e. save states were already there)
Side-notes (not worth mention):
* I fear we were missing a local static array in 1942 save states (now it is saved)
* I haven't found a better way to configure 1943 banks than to split the 0x4000 bank into 4 pieces. can anyone come up with a better approach?
Not 100% sure this is the better way to implement this (we pass the eeprom tag as parameter of the PORT_CUSTOM), but I haven't been able to find a better solution.
No driver uses this yet, so I'm open to any suggestion before to use it extensively ;)
I don't like too much this solution, but now the code is self contained: hence, better fixes (e.g. no MDRV_DEVICE_CONFIG_DATAPTR(eeprom_config, default_data, &_data) at all for the NODEFAULT eeprom?) could be added without further modifying the behavior across the drivers.
- fixed bug in vtlb code that caused us to lose track of previously
registered fixed page ranges
- fixed MIPS3 behavior that would not clear out invalid page ranges
from the VTLB under certain circumstances
- added support for TLB sizes less than 48 entries
---------- Forwarded message ----------
From: Oliver Stöneberg <oliverst@online.de>
Date: Sun, Nov 15, 2009 at 3:43 PM
Subject: fixes for cppcheck issues
To: submit@mamedev.org
The following patch fixes some issues reported by cppcheck:
- buffer out-of-bounds accesses in mame/drivers/peplus.c and
mame/drivers/videopkr.c (the latter hopefully fixing the long-
standing crashing issue)
- compilation of mame/machine/irobot.c when DISASSEMBLE_MB_ROM is set
to 1
Switched the SMS and Game gear PSGs to use an XOR instead of XNOR feedback, but this needs further verification. Updated comments to show which chips are currently verified. Cleaned up comments.
- now properly generating TLB fill exceptions under correct circumstances
- TLB exceptions no longer trash low 4 bits of Context
- exceptions with the EXL bit set always go to vector 0x180
future lookup optimizations.
By and large, the only things that used 2-character tags were bulk-
converted sound chips which were generally mapped to "ym" or "ay"
(and a few others). These were changed to "ymsnd" and "aysnd",
mimicking the way CPUs were changed from "main" to "maincpu" a while
back.
